article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the iron - pnictide superconductors have garnered much attention since the first observation @xcite that chemical manipulation of these materials could lead to high - temperature superconductivity @xcite .
particular interest has fallen on the structural `` 122 '' family with the thcr@xmath2si@xmath2 crystal structure , because of their intermetallic nature and relative ease of synthesis .
superconducting transition temperatures ( @xmath9 ) as high as @xmath10 k have been observed in transition metal substituted bafe@xmath2as@xmath2 @xcite , while @xmath9 up to @xmath11 k has been seen when hole doping on the alkaline earth site in ba@xmath5k@xmath6fe@xmath2as@xmath2@xcite .
recently , superconductivity with @xmath9 values reaching as high as @xmath12 k was observed to occur via aliovalent rare earth substitution into the alkaline earth site of cafe@xmath2as@xmath2 @xcite , approaching the highest @xmath9 values of all iron - based superconductors ( @xmath13 k in smo@xmath5f@xmath6feas @xcite ) .
stabilized by the suppression of antiferromagnetic order via a combination of both chemical pressure and electron doping , the superconductivity observed in the ca@xmath5r@xmath6fe@xmath2as@xmath2 series ( r = la , ce , pr , nd ) arises in a highly tunable system where the choice of rare earth species allows for structural tuning toward the collapsed tetragonal phase @xcite , such as shown in the pr- and nd - doped series , or more simply electron doping with minimal change in unit cell size as shown for la substitution @xcite .
the close ionic radius match of the light rare earths with ca in the cafe@xmath2as@xmath2 system makes for an ideal system with which to study the interplay between superconductivity and structural instability .
however , the observations of partial volume - fraction diamagnetic screening in low - concentration rare earth - substituted cafe@xmath2as@xmath2 @xcite , where the solubility limit of la is 30% , poses a challenge to understanding the reasons for a lack of bulk - phase superconductivity .
one promising route is through the inclusion of higher concentrations of rare earth substitution .
previous studies of la substitution into the srfe@xmath2as@xmath2 system @xcite , with @xmath9 values up to @xmath14 k , have shown a significant increase in the superconducting volume fraction when la content reaches @xmath15% la for sr , at which point the volume fraction jumps to nearly @xmath16% .
the introduction of higher concentrations of rare earth in the cafe@xmath2as@xmath2 series is thus a promising route to achieving bulk superconductivity with high @xmath9 values .
while it has been shown that the application of pressure during synthesis can provide the desired results @xcite , a clue to another possible route is provided by the materials themselves ( _ i.e. _ , in the case of the ca@xmath5la@xmath6fe@xmath2as@xmath2 series , where the @xmath17-axis remains constant but the @xmath18-axis increases with increasing rare - earth concentration up to the solubility limit of @xmath19 @xcite . )
it follows that a larger unit cell would likely be able to accomodate a larger concentration of rare earths .
it has also been shown previously that substituting sr in the place of ca leads to a controlled expansion of the unit cell in accordance with vegard s law @xcite ; therefore , such a solid solution may serve as the base formula into which rare earth atoms can be further doped . in this article
, we examine the feasibility of increasing the solubility limit of rare earth substitution by doping la into the ( sr , ca)fe@xmath2as@xmath2 solid solution system .
the synthesis of such single crystals with la substitution shows widely ranging chemical compositions and suggest a competition between sr and la .
we detail the effects of increasing la content using systematic x - ray , electrical transport , and magnetization measurements , and compare these effects with those observed when la is doped into the cafe@xmath2as@xmath2 parent material , specifically tracking the suppression of the afm ordering temperature and signs of superconductivity in the system .
results of energy dispersive spectroscopic ( eds ) analysis of the elements occupying the alkaline earth site .
a ) measured ( actual ) la content vs. nominal ( pre - reaction ) la content .
black close circles represent samples with pre - reaction stoichiometries of ( sr@xmath0ca@xmath1)@xmath5la@xmath6fe@xmath2as@xmath2 .
colored symbols represent samples with pre - reaction stoichiometries of ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 , where @xmath20 is denoted in the legend .
b ) measured la content vs. measured sr content in the same samples shown in a ) . all values are taken from eds analysis .
the black solid line serves as the upper bound for the solubility of la.,width=604 ]
single crystals were grown via a self - flux method using elemental stoichiometries of @xmath21:@xmath22:@xmath23:@xmath24 for sr : ca : la : feas according to the formula ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 with feas flux @xcite .
starting materials were placed inside alumina crucibles and sealed in quartz tubes under partial atmospheric pressure of ar .
the growths were heated and allowed to slow cool , resulting in crystals with typical dimensions of ( @xmath25 ) mm@xmath26 , which were mechanically separated from the frozen flux .
chemical analysis was obtained using both energy - dispersive ( eds ) and wavelength - dispersive ( wds ) x - ray spectroscopy , showing @xmath27:@xmath28:@xmath28 stoichiometry between ( sr , ca , la ) , fe , and as concentrations .
eds was conducted on a large number of samples in order to determine general concentration trends , while wds was used to determine very accurately the concentrations of elements for samples used in x - ray , resistivity , and magnetization measurements .
single crystal x - ray diffraction was performed on a bruker smart apex2 diffractometer equipped with a ccd detector , graphite monochromator , and monocap collimator .
crystal structures were refined ( shelxl-97 package ) using the i4/mmm space group against 113 and 106 independent reflections measured at @xmath29 k and corrected for absorption using the integration method based on face indexing ( sadabs software ) . because three different atoms were occupying the same crystallographic site ,
refinement of chemical compositions was not possible , and refinement instead was focused on obtaining unit cell parameters .
resistivity measurements were performed using the standard four - probe ac method , via gold wire and in / sn solder contacts with typical contact resistance of @xmath30 at room temperature , using up to 1 ma excitation currents at low temperatures .
magnetic susceptibility was measured using a commercial superconducting quantum interference device magnetometer .
unit cell parameters @xmath18 and @xmath17 obtained via single crystal x - ray diffraction of ( sr , ca)fe@xmath2as@xmath2 ( filled points ) @xcite and ( sr , ca , la)fe@xmath2as@xmath2 ( open points ) , presented in ( a ) and ( b ) , respectively .
the @xmath17-axis lattice parameters are in good agreement , but the @xmath18-axis lattice parameters move away from the established trend as the amount of la increases.,width=604 ] eds of the actual concentration of la in the ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 series reveals that at low la values , the actual la content is higher than the nominal content , rising to a limit of @xmath31% for growths with nominal la higher than @xmath32% as shown in fig . 1a . while it is illustrated that increasing the starting ratio of ca : sr leads to higher la concentrations in the final materials , a stronger correlation between the sr , ca , and la concentrations can be found by plotting the measured la content against the measured sr content as shown in fig .
it is evident here that la and sr are inversely correlated in this material and increasing the sr concentration seems to strongly limit the amount of la that is able to dope into the sample .
resistivity of ( sr , ca , la)fe@xmath2as@xmath2 samples as a function of temperature , normalized to @xmath33 k and then vertically shifted for clarity ; the ca - rich samples are presented in ( a ) , while the sr - rich samples are presented in ( b).,width=604 ] single crystal x - ray analysis allows us to analyze the progression of the lattice parameters as a function of the concentrations of sr , ca , and la in each sample . in fig .
2 , we plot the lattice parameters of the samples used in this study alongside the lattice parameters observed for solid solutions of the parent compounds
srfe@xmath2as@xmath2 and cafe@xmath2as@xmath2 @xcite .
previously , it was shown that doping la for ca in cafe@xmath2as@xmath2 results in a @xmath17-axis lattice parameter that does not change , despite expansion of the @xmath18-axis @xcite . taking this into account ,
we have plotted these points as sr@xmath5(ca , la)@xmath6 the ca and la values are taken together in order to determine the composition @xmath23 , which places our points in good agreement with the @xmath17-axis values from the sr@xmath5ca@xmath6fe@xmath2as@xmath2 study ( fig .
2b ; , however , the @xmath18-axis values diverge as la content increases ( fig .
this implies that by selecting the proper sr , ca , and la content , we can tune the @xmath18-axis and @xmath17-axis parameters nearly independently .
this is in striking contrast to most doping studies on these materials , which show a strong coupling between @xmath18 and @xmath17-axes lattice parameters @xcite .
as seen in similar doping studies of iron - pnictides @xcite , it is expected that increasing the la content in these samples will be manifest in resistivity data as a systematic decrease in the nel ordering temperature @xmath34 .
electrical resistivity data of these samples ( shown in fig .
2 ) roughly resembles the expected behavior , as it is obvious that @xmath34 is suppressed upon increased doping of la into the system .
a key difference here lies in the ranging chemical compositions obtained using wds ; subtraction of the la content leaves two classes of samples , _
i.e. _ the ca - rich ( shown in fig . 3a ) and the sr - rich ( shown in fig .
3b ) . in the sr - rich case ,
no sample was found to contain less than @xmath35% la or more than @xmath14% la , whereas in the ca - rich case , a much wider range of la concentrations can be found ( up to @xmath31% la ) . in the sr - rich samples ,
@xmath34 is gradually suppressed down to @xmath36130 k and no superconductivity is found to exist due to la substitution . in the ca - rich case , @xmath34 is suppressed to a slightly lower value of @xmath36100 k , but no trace of a superconducting phase similar to that seen in the ca@xmath5la@xmath6fe@xmath2as@xmath2 case @xcite , where high-@xmath9 values in the range 30 - 47 k are indicative of rare earth doping - induced superconductivity , was found .
note that the ca - rich samples do exhibit traces of a superconductivity onset near @xmath3710 k , which we attribute to the strain - induced phase often observed under non - hydrostatic pressure conditions @xcite and posited to nucleate at afm domain walls @xcite .
it is interesting to highlight the fact that this `` 10 k '' phase appears predominantly in ca - rich samples , suggesting its stability is tied strictly to the cafe@xmath2as@xmath2 magnetic and/or crystallographic structure .
magnetic susceptibility ( @xmath38 ) vs. temperature for several la doped samples .
the data are shifted along the vertical axis for clarity .
the kink in each curve indicates the temperature of the antiferromagnetic transition , @xmath34 .
the inset shows the low field magnetic susceptibility of the @xmath39%-la doped sample at low temperatures , where superconductivity below @xmath9 @xmath35 k is seen ; the estimated superconducting volume fraction is @xmath40%.,width=377 ] temperature dependence of magnetic susceptibility @xmath38(@xmath41 ) data for ca - rich samples ( shown in fig .
4 ) corroborates the picture drawn by electrical transport data .
as expected from previous studies , @xmath34 is revealed as an antiferromagnetic ordering temperature traced by a kink in @xmath38(@xmath41 ) .
the suppression of @xmath34 occurs at the same rate observed in transport data , with ordering at @xmath36100 k still present for samples which show superconductivity at @xmath3710 k ( inset of fig .
4 ) at low field .
the meissner screening fraction of the @xmath3610 k superconductivity of this sample is still seen to be relatively small , of the order of @xmath42% .
a slight curie tail is observed in the highly - doped la samples at low temperatures , similar to that observed in ca@xmath5la@xmath6fe@xmath2as@xmath2 samples @xcite , which may arise from paramagnetism associated with the feas lattice .
( a ) the phase diagram for the ca@xmath5la@xmath6fe@xmath2as@xmath2 system @xcite , where high @xmath9 superconductivity is induced on the border of afm order and coexists with the @xmath3710 k superconducting phase .
( b ) suggested phase diagram for the ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 system .
the solid and broken lines are guides to the eye.,width=302 ] figure 5 presents a proposed phase diagram for the ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 system in comparison with that of the ca@xmath5la@xmath6fe@xmath2as@xmath2 system @xcite . as shown in fig .
5b , a key observation in ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 is the absence of the high @xmath9 superconducting phase , which is observed ubiquitously in the rare - earth doped cafe@xmath2as@xmath2 materials upon suppression of the afm phase , despite a similar electron doping scheme .
the lack of a high-@xmath9 superconducting phase in ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 samples with la concentrations more than sufficient to induce superconductivity in ca@xmath5la@xmath6fe@xmath2as@xmath2 ( fig .
5a ) suggests that a scenario where the superconducting phase arises solely from the presence of sufficient rare earth atoms that presumably cluster or percolate in some manner is improbable .
the persistence of afm order up to high concentrations of la may play a role here , as it seems as though high @xmath9 superconductivity competes with afm order and does not emerge until the complete suppression of magnetic ordering .
indeed , in every other rare - earth substituted @xmath43 system , @xmath44 + k superconductivity and antiferromagnetism are never found to coexist @xcite .
this agrees with the occurrence of the highest @xmath9 in the 1111 iron - pnictide family @xcite , indicating that the highest @xmath9 superconducting phase and magnetic ordering may be mutually exclusive .
of course , further investigation will be necessary to bear out such a result .
however , the conspicuous absence here of the high @xmath9 phase , which has been thought to be an impurity phase of reofeas , despite similar growth techniques and materials , lends credence to the idea that it is in fact intrinsic to the rare - earth substituted cafe@xmath2as@xmath2 system .
in summary , we have studied the effect of electron doping by la substitution on ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 solid solutions by growing single crystals .
we have constructed a phase diagram based on transport , magnetic susceptibility and structural characterization .
chemical analysis indicates a strong inverse correlation between sr and la .
nonetheless , independent tunability of the @xmath18- and @xmath17-axis lattice parameters can be achieved .
the sr - rich and ca - rich regions show differing behavior ; in ca - rich samples , antiferromagnetic ordering is found to coexist with superconductivity with @xmath3710 k with a volume fraction @xmath35% .
but , in contrast to cafe@xmath2as@xmath2 , in neither case is @xmath34 fully suppressed and no high-@xmath9 superconducting phase is observed , placing the constraint that complete suppression of afm order is a necessary condition for the latter phase , which may provide an important clue for the superconducting pairing in the new iron - superconductors .
0.5 cm acknowledgements this work was supported by afosr - muri grant fa9550 - 09 - 1 - 0603 .
11 kreyssig a , green m a , lee y , samolyuk g d , zajdel p , lynn j w , budko s l , torikachvili m s , ni n , nandi s , j b leao , poulton s j , argyriou d n , harmon b n , mcqueeney r j , canfield p c , and goldman a i , 2008 _ phys .
b _ * 78 * 184517 . | the effects of aliovalent rare earth substitution on the physical properties of sr@xmath0ca@xmath1fe@xmath2as@xmath2 solid solutions are explored . electrical transport , magnetic susceptibility and structural characterization data as a function of la substitution into ( sr@xmath3ca@xmath4)@xmath5la@xmath6fe@xmath2as@xmath2 single crystals confirm the ability to suppress the antiferromagnetic ordering temperature from 200 k in the undoped compound down to @xmath7 k approaching the solubility limit of la . despite up to @xmath8 la substitution ,
the persistence of magnetic order and lack of any signature of superconductivity above 10 k present a contrasting phase diagram to that of ca@xmath5la@xmath6fe@xmath2as@xmath2 , indicating that the suppression of magnetic order is necessary to induce the high - temperature superconducting phase observed in ca@xmath5la@xmath6fe@xmath2as@xmath2 . |
despite the successes of the ads / cft correspondence , a full understanding of the duality is still lacking . since the original conjecture @xcite , much of the research effort has been aimed at developing a precise dual dictionary connecting geometric quantities in the bulk to observables in the boundary field theory .
however , this dictionary remains poorly understood , hampering our ability to reconstruct the bulk geometry from the field theory , compute field theory quantities from the classical bulk theory , and assess whether or not the correspondence is complete .
this last effort has been placed under scrutiny by the recent controversy over the black hole interior ( see e.g.@xcite ) , and direct efforts to construct or demonstrate an inability to construct a full dual dictionary , may serve to address it . to resolve such questions , a good starting point is the identification of bulk probes that depend primarily on the bulk geometry and are dual by the ads / cft dictionary to known field theory observables .
the existence of such bulk probes is not guaranteed ; however , spacelike extremal surfaces , which are covariantly - defined and depend exclusively on the bulk geometry , are dual to well - understood field theory observables .
in fact , spacelike extremal surfaces constitute most of the probes used in ads / cft .
the dual observables include correlators @xcite , entanglement entropy @xcite , and wilson loops @xcite .
the length of spacelike geodesics with boundary endpoints allows one to compute , in the wkb approximation , the two - point correlator of a high conformal dimension field operator at the endpoints .
the area of a codimension 2 spacelike extremal surface anchored on some region @xmath0 on the boundary of an asymptotically locally anti - de sitter ( alads ) spacetime is associated with the entanglement entropy of @xmath0 within the boundary field theory . a field theory observable which is dual to an extremal surface must in some way
encode information about the bulk geometry at the location of the extremal surface .
if there is a limit on how far into the bulk such surfaces can reach , this also limits our ability to reconstruct the geometry from the corresponding dual observables .
the natural question that arises is how much of the bulk can be recovered from extremal surface probes .
extremal surface probes and their maximal reach were studied in @xcite for the case of static and translationally or spherically symmetric asymptotically ads spacetimes , and in a large amount of literature ( e.g. @xcite and @xcite ) for eternal ads black holes .
studies have also been done for extremal surfaces in time - dependent geometries ( e.g. @xcite ) . in this paper
, we establish a general constraint on the reach of probe spacelike extremal surfaces in spacetimes which are not necessarily alads .
we find that many generic geometries admit a certain kind of surface which acts as a barrier for probe surfaces anchored to one boundary .
we furthermore show that trapped surfaces with nonpositive null extrinsic curvature give rise to extremal surface barriers , and that on spacetimes which admit a totally geodesic spatial slice
the existence of barriers implies the existence of either a singularity or a marginally trapped surface .
more explicitly , we consider a codimension 1 `` splitting surface '' @xmath1 which divides the spacetime into two regions .
@xmath1 may be spacelike , timelike , or null .
we take some class @xmath2 of spacelike extremal surfaces of codimension 1 or greater , which are anchored to some boundary @xmath3 on one side of @xmath1 .
then , given two simple assumptions , we can show that no members of this class of extremal surfaces cross ( or even touch ) @xmath1 .
in other words , the splitting surface acts as an extremal surface barrier .
the two assumptions are as follows : 1 .
all of the extremal surfaces in the class @xmath2 can be continuously deformed ( while remaining extremal ) to surfaces which lie on the exterior of @xmath1 .
@xmath1 has nonpositive extrinsic curvature ( as measured by normal vectors pointing outward , towards the exterior of @xmath1 ) . in other words
, @xmath1 can only bend outward relative to its tangent plane .
given these assumptions , @xmath1 acts as a barrier to any of the extremal surfaces lying in the class @xmath2 .
this generalizes some previously known result in riemannian manifolds , that surfaces with an inward - pointing mean curvature vector can sometimes act as local barriers to minimal surfaces , and also act as global barriers in certain circumstances ( see e.g. @xcite ) .
our proof uses elements from the approach of @xcite and @xcite , and shows that spacelike extremal surfaces can not approach surfaces with negative extrinsic curvature .
they can , however , approach arbitrarily close to totally geodesic surfaces .
we do not need to make any assumptions about causality or an energy condition ; the restriction is purely geometrical in nature .
the first criterion is often satisfied : it simply requires that there be no obstruction to deforming the extremal surface ( while maintaining extremality ) away from @xmath1 .
( e.g. geodesics in the btz black hole geometry @xcite ) .
this is not an issue for our proof , since we do not require that the deformed extremal surface be minimal .
] in many cases , this criterion holds for all extremal surfaces due to the topology and geometry of the spacetime , so the class @xmath2 consists of all extremal surfaces . for example , this is true in vacuum ads spacetimes . more generally , if the extremal surfaces are anchored to just one connected component of the boundary @xmath3 , and the bulk spacetime is homotopically trivial , then all surfaces may be continuously deformed towards @xmath3 .
however , it may be an additional constraint on the geometry that the surfaces can remain extremal while this happens .
for example , if there is another region @xmath4 which acts as a barrier to extremal surfaces anchored anywhere on the boundary of the spacetime @xmath5 , then @xmath4 may be an obstruction to deforming extremal surfaces past @xmath1 .
the second criterion may be satisfied by a totally geodesic surface , i.e. a surface with vanishing extrinsic curvature . can also be equivalently defined as a surface @xmath1 in @xmath6 such that every geodesic on @xmath1 ( with respect to the induced connection on @xmath1 ) is also a geodesic on @xmath7 ( with respect to the connection on @xmath7 . ]
an example of such a surface is a non expanding black hole horizon ( which , given the area theorem @xcite , is a stationary black horizon for spacetimes obeying the null energy condition ) . however , totally geodesic surfaces do not appear in generic spacetimes .
fortunately , the result also applies to surfaces with negative extrinsic curvature , which can appear in generic spacetimes .
examples will be given in section [ examples ] .
in fact , if there exists a surface @xmath1 whose extrinsic curvature is negative everywhere , we will show that it is not actually the tightest bound on extremal surfaces .
there will be some other `` outermost '' barrier @xmath4 some distance away , which has at least partly nonnegative extrinsic curvature ( in section [ singularities ] , we will prove a theorem relating compact outermost barriers to trapped surfaces and singularities ) .
the existence of such extremal surface barriers and their link to singularities is a curiosity which is of relevance to the firewall controversy .
the fact that many bulk probes do not extend beyond the barrier may suggest that the firewall , if it exists , may be at or behind the outermost barrier . at the very least
, it shows that field theory observables that are dual to extremal surfaces anchored at one boundary do not directly reveal any information about the interior .
any information about this region in the bulk must come from probes that are not extremal surfaces . in some cases , the presence of a barrier
might even indicate that the boundary field theory does not have sufficient information to reconstruct the geometry behind the barrier @xcite . in that case ,
additional factors of the hilbert space may be needed to describe that region .
an interpretation involving a loss of determinism may at first seem implausible in the case where @xmath1 is spacelike : one could use the bulk equations of motion to evolve the information in @xmath1 forwards or backwards in time and thus reconstruct some or all of the interior of @xmath1 .
however , the presence of a firewall at @xmath1 would result in a breakdown of the classical bulk equations of motion . in this case , our hope of reconstructing the interior of @xmath1 relies entirely on the ads / cft correspondence dictionary .
this dictionary , however , is primarily composed of extremal surface probes , which we prove below can not probe past @xmath1 . this paper is structured as follows . in section
[ theorem ] , we prove our sufficient condition for a spacetime to have a `` barrier surface '' past which extremal surfaces can not reach ; we further prove that trapped surfaces form extremal surface barriers .
we will also show that the outermost barrier surface must have partly nonnegative extrinsic curvature .
section [ examples ] provides examples of bulk spacetimes with extremal surface barriers . in section [ singularities ]
, we prove that in a certain a class of geometries , barriers occur only in the presence of singularities or trapped surfaces .
we first prove in two parts that extremal surfaces do not probe past surfaces with a nonpositive extrinsic curvature tensor , and further , that barriers with everywhere negative extrinsic curvature are not the outermost barrier surfaces .
@xmath6 will be a lorentzian manifold with at least one boundary @xmath3 for the rest of this paper .
we start with two definitions .
define a codimension 1 surface @xmath1 in @xmath7 to be a _ splitting surface _ if it separates @xmath7 into two open regions , ext@xmath8 and int@xmath8 , where we define the interior and exterior relative to the normal which defines the extrinsic curvature , and @xmath9 .
this normal is taken to point towards ext@xmath8 .
let @xmath1 now be a splitting surface .
let @xmath10 be a family of spacelike extremal surfaces in @xmath7 of codimension @xmath11 such that all the @xmath12 can all be continuously deformed from some initial surface @xmath13 anchored at ext@xmath14 , and all the @xmath12 are anchored at ext@xmath15 ( see , e.g. fig .
[ def ] ) .
, so that the extrinsic curvature @xmath16 is well - defined
. presumably this assumption can be weakened if care is taken in dealing with distributional extrinsic curvatures . ]
then any surface in the family @xmath12 shall be called _ @xmath1-deformable_.
* theorem 2.1 : * let @xmath1 be a codimension 1 splitting surface in @xmath7 such that for any vector field @xmath18 on @xmath1 , @xmath19 , where @xmath20 is the extrinsic curvature of @xmath1 .
, the induced metric on @xmath1 , the extrinsic curvature on @xmath1 is given by @xmath21 .
for a surface @xmath22 of codimension @xmath23 , the extrinsic curvature carries an additional index , and is defined in terms of @xmath24 , the induced metric on @xmath22 : @xmath25 .
] any @xmath1-deformable spacelike extremal surface which is anchored within @xmath26 remains in ext@xmath8 .
moreover , no such spacelike extremal surface ever touches @xmath1 ( or even comes arbitrarily close to touching @xmath1 ) . , we could prove , using precisely the same formalism used in the proof above , that surfaces anchored in int@xmath8 remain in int@xmath8 . ]
the proof closely follows that of lemma b in @xcite .
let @xmath12 be the family of surfaces all deformable from @xmath27 as defined above .
if all surfaces @xmath12 are in ext@xmath8 , we are done , since that implies that surfaces which are anchored at @xmath28 can never be deformed past @xmath1 .
suppose there exists a surface @xmath29 in @xmath10 such that @xmath30 .
then , since @xmath29 is linked to @xmath27 via a series of smooth deformations , there exists a `` midway '' surface @xmath22 in the family of surfaces which coincides with @xmath1 at a set of points @xmath31 and is tangent to @xmath1 at those points .
[ def ] illustrates the deformation family .
we focus on one coincident point @xmath32 of @xmath1 and @xmath22 .
as explained in @xcite , @xmath33 is a measure of how much @xmath1 curves away from the tangent plane normal to @xmath34 with motion away from @xmath32 in the @xmath18 direction , where @xmath34 is the normal to @xmath1 when @xmath1 is not null , and we take @xmath35 to be the null generator for null @xmath1 .
similarly , @xmath36 measures how much @xmath22 curves away from the plane normal to @xmath1 at @xmath32 as one moves in the @xmath18 direction away from @xmath32 ( restricted to motion tangent to @xmath22 ) .
@xmath22 is outside of @xmath1 , so @xmath1 curves away from its tangent plane at least as much as @xmath22 does : @xmath37 in particular : @xmath38 where @xmath24 is the induced metric on @xmath22 . the equality in eq .
[ firstcond ] follows from the fact that @xmath22 is extremal .
@xmath22 is spacelike , so @xmath24 is riemannian and has no temporal components , which implies that @xmath39 , since @xmath40 .
we therefore have a contradiction with eq . [ negcond ]
, so the surfaces in the family @xmath10 never reach @xmath1 , which in turn implies that @xmath29 does not exist .
moreover , it is impossible for @xmath22 to approach arbitrarily close to @xmath1 , because in the limit eq .
[ negcond ] would still hold , and we would again arrive at a contradiction . may allow us to prove the existence of a barrier with a weaker condition than in theorem 2.1 .
for example , to prove that a spacelike codimension 1 surface is a barrier for @xmath41-dimensional extremal surfaces , it is sufficient if the sum of the largest @xmath41 eigenvalues of @xmath20 are negative .
however , the analysis is more complicated for timelike or null barriers . ] * theorem 2.2 : * let @xmath1 be a splitting surface as above , but we now take it to be totally geodesic , i.e. @xmath43 . any @xmath1-deformable spacelike extremal surface which is anchored in the intersection of the boundary @xmath3 with one component of @xmath7 ( as divided by @xmath1 ) remains in that component has some nontrivial topology or geometrical defect that creates an obstruction for @xmath1-deformability . if we can find another spacetime @xmath44 which agrees with @xmath7 everywhere inside @xmath1 and also obeys the following conditions : ( a ) the extremal surface @xmath22 is the same in @xmath7 and @xmath44 , ( b ) the splitting surface @xmath1 is the same in @xmath7 and @xmath44 , and ( c ) @xmath22 is @xmath1-deformable in @xmath44 , then @xmath1 is still a barrier in @xmath7 , by the proof of theorem 2.2 . ] .
note that in the case where @xmath1 is totally geodesic , it provides a barrier both for surfaces in int@xmath8 and ext@xmath8 .
the proof is identical for @xmath45 and @xmath46 . for concreteness
we assume @xmath47 .
we take @xmath22 to be the midway surface as above , and we show that if @xmath22 agrees with @xmath1 at @xmath32 and is tangent to it , then @xmath48 .
this would imply that @xmath22 is no longer anchored at @xmath28 , and therefore it is impossible to deform a surface to reach and cross @xmath1 while maintaining boundary conditions in ext@xmath8 .
the saturated equation is : @xmath49 this shows that the two surfaces curve away in the normal direction equally , and is an indication that the two surfaces must agree everywhere on @xmath22 if they coincide at a point .
we now show this in more detail .
suppose @xmath1 is null , and let @xmath34 be the null generator of @xmath1 directed towards @xmath22 .
the corresponding covector is defined : @xmath50 , as illustrated in fig .
consider a small neighborhood @xmath51 of @xmath32 of length scale @xmath52 .
let @xmath53 and let @xmath54 .
define the coordinate @xmath55 : @xmath56 where @xmath57 is obtained by parallel transporting @xmath58 along @xmath1 , and @xmath59 is a vector pointing from @xmath60 to @xmath32 in the small neighborhood around @xmath32 .
note that because @xmath1 is totally geodesic , @xmath61 is null . within @xmath51
, we can always find an @xmath62 such that @xmath63 .
for such a choice of @xmath60 , we therefore obtain : @xmath64 where @xmath65 is obtained by parallel transporting @xmath66 along @xmath22 .
the null extrinsic curvatures therefore obey : @xmath67 let @xmath68 , defined on @xmath1 , be the proper distance from @xmath32 .
we can use @xmath68 as a coordinate labeling points on @xmath1 .
let @xmath69 be the volume element of a constant @xmath68 slice .
define @xmath70 to be a green s function on the ball of points @xmath71 with @xmath72 , where @xmath73 where @xmath74 is defined on @xmath1 . for sufficiently small @xmath75 ,
the metric on @xmath1 is approximately a flat euclidean - signature metric , so @xmath76 or @xmath77 in @xmath78 . since @xmath79 (
regardless of what @xmath80 is ) , we find @xmath81 .
when @xmath75 is small , these inequalities must continue to hold if the metric is slightly deformed by nonzero curvature .
integrating over a ball @xmath82 of radius @xmath83 : @xmath84 since @xmath85 is never negative ( that would contradict the assumption that @xmath22 does not cross @xmath1 ) , we conclude that @xmath86 everywhere on a neighborhood of @xmath32 . but
this implies that @xmath86 everywhere , since it solves an elliptic equation , and thus @xmath48 .
this completes the proof for null @xmath1 .
this reasoning applies directly to the cases where @xmath1 is timelike or spacelike by simply substituting the @xmath22-pointing spacelike or timelike normal of @xmath1 , respectively , for the null generator in the @xmath22 direction .
this proof allows us to generalize a previously - known result for totally geodesic surfaces : + * corollary 2.3 : * let @xmath1 be a totally geodesic codimension @xmath87 surface in @xmath7 .
let @xmath22 be a codimension @xmath88 spacelike extremal surface satisfying the following conditions : 1 .
@xmath22 coincides with @xmath1 at some point(s ) @xmath32 , 2 .
@xmath22 does not cross @xmath1 anywhere ( and therefore @xmath22 is tangent to @xmath1 at @xmath32 ) , then @xmath48 .
+ let @xmath89 be the outwards - pointing normals to @xmath1 as above ( null generators in the @xmath22 direction , if @xmath1 is null ) , where @xmath90 runs from @xmath91 .
let @xmath92 .
let @xmath51 be a neighborhood of @xmath32 .
we define @xmath55 for @xmath53 as above : @xmath93 where @xmath54 is chosen so that @xmath94 . then : @xmath95 where for the case of null @xmath1 , @xmath96 , and for timelike and spacelike cases , we just require them to be equal . as before ,
we obtain @xmath97 by parallel transporting along @xmath1 and @xmath98 by parallel transporting along @xmath22 . the remaining reasoning in the proof of theorem 2.2 above yields @xmath99 , and we are done .
note that theorems 2.1 and 2.2 , and corollary 2.3 together also give a complete description of the behavior of extremal surfaces in the presence of barriers , provided that all disconnected components are @xmath1-deformable .
corollary 2.3 also leads to a natural corollary of theorems 2.1 and 2.2 , which combines the assumptions made in those theorems : * corollary 2.4 : * let @xmath1 be a splitting surface in @xmath7 such that @xmath100 for some vector fields @xmath101 on @xmath1 and @xmath102 for all other vector fields @xmath103 on @xmath1 .
then any @xmath1-deformable spacelike extremal surface which is anchored in ext@xmath8 remains in ext@xmath8 .
moreover , although such extremal surfaces might conceivably come arbitrarily close to @xmath1 if they only propagate along the @xmath104 directions , they can not approach @xmath1 if they propagate along the @xmath105 directions .
as above , let @xmath10 be a family of @xmath1-deformable surfaces , all deformable from @xmath13 , and let @xmath22 be the midway surface between @xmath29 and @xmath27 .
we again focus on a coincident point @xmath106 .
[ negcond ] , we have @xmath107 , where @xmath24 is the induced metric on @xmath22 . we can decompose the metric into the components that lie along the @xmath108 directions and the @xmath109 directions : @xmath110 where @xmath111 and @xmath112 for some constants @xmath113 and @xmath114 .
. [ negcond ] can be decomposed as follows : @xmath115 if @xmath22 has no components in the @xmath105 directions , then @xmath116 , and eq .
[ decomposition ] reduces to @xmath117 .
this is simply the case of theorem 2.2 , so @xmath22 could potentially get arbitrarily close to @xmath1 while remaining on ext@xmath8
. if @xmath22 has at least one component in @xmath105 directions , then eq . [ decomposition ] yields @xmath118 .
this is simply the case of theorem 2.1 , so @xmath22 can not approach @xmath1 , and must always remain in ext@xmath8 .
we conclude that if @xmath22 propagates in the directions in which @xmath1 has vanishing extrinsic curvature , @xmath22 can approach @xmath1 , but if @xmath22 propagates in the directions in which @xmath1 has negative extrinsic curvature , @xmath22 can not approach @xmath1 .
one consequence of corollary 2.4 is that any null splitting surface @xmath1 which is foliated by surfaces where _ all _ components of the null extrinsic curvature are nonpositive is an extremal surface barrier .
but if we are only interested in whether or not @xmath1 is a barrier to codimension 2 extremal surfaces , it turns out the we can do better : it is sufficient if the expansion @xmath119 , i.e. the _ trace _ of the null extrinsic curvature is nonpositive .
this is because when a codimension 2 extremal surface @xmath22 touches ( but does not cross ) @xmath1 , we can take the trace of eq . [
2.1 ] over all @xmath120 spacelike directions to obtain : @xmath121[thetas ] as in the case of corollary 2.3 , the above equation can only be saturated if @xmath22 lies on @xmath1 .
this means that @xmath1 is a barrier to @xmath1-deformable spacelike extremal codimension 2 surfaces .
if we assume the null curvature condition @xmath122 , we can additionally show how to use a codimension 2 extremal surface to construct a barrier to other codimension 2 surfaces .
* theorem 2.5 : * let @xmath123 be a codimension 2 spacelike extremal surface .
let @xmath1 be the union of null congruences shot outwards from @xmath123 towards both the future and the past directions .
assuming the null curvature condition , @xmath1 is a barrier to @xmath1-deformable codimension 2 surfaces anchored to ext(@xmath1 ) . by the standard construction
, the raychaudhuri equation implies that the two null congruences shot out from @xmath123 converge , so that @xmath124 , where @xmath125 is defined as the expansion moving outwards away from @xmath123 . as in the proof of theorems 2.1 and 2.2 , we assume for contradiction that @xmath22 is a midway surface touching @xmath1 but not crossing it .
( it does not matter which of the two null congruences @xmath22 touches first , but whichever one it touches first , we are only interested in the @xmath125 of that congruence . ) at the point of coincidence , eq . [ thetas ] implies that @xmath126 and @xmath127 both vanish .
therefore , by the same reasoning as in corollary 2.3 , @xmath22 must lie entirely on @xmath1 .
but then it can not be anchored at ext(@xmath1 ) . even without assuming the null curvature condition
, the following corollary follows immediately : * corollary 2.6 : * if @xmath1 is a null surface foliated by ( marginally ) outer trapped surfaces ( i.e. @xmath119 ) , then @xmath1 is a barrier to @xmath1-deformable codimension 2 surfaces anchored to ext(@xmath1 ) .
thus , if we are only interested in codimension 2 extremal surfaces ( e.g. for purposes of holographic entanglement entropy @xcite ) , we can prove the existence of barriers in more circumstances .
we proved that the presence of the surface @xmath1 with nonpositive spatial extrinsic curvature is a sufficient condition for an extremal surface barrier .
this raises the question of whether this barrier is the closest one to the extremal surfaces .
consider some region on the boundary of the spacetime , @xmath129 .
let @xmath130 denote the spacelike extremal surfaces anchored at @xmath0 .
we define the outermost barrier @xmath1 for surfaces anchored at @xmath0 to be given by : @xmath131 ( for some ( non - alads ) spacetimes , @xmath0 may be a proper subset of ext@xmath8 , as shown in fig .
[ outermost ] . )
the above definition is partly motivated by @xcite , although the case under discussion there involved only minimal area surfaces , whereas we consider all extremal surfaces .
note that we can similarly define the outermost barrier for a subset @xmath132 of all spacelike extremal surfaces anchored at @xmath0 by defining @xmath133 .
we now prove some properties of the outermost barrier . + * theorem 2.7 : * let @xmath1 be an outermost barrier in @xmath7 , and let @xmath10 be a family of extremal surfaces that can all be smoothly deformed from some initial surface @xmath13 such that some surface @xmath134 is arbitrarily close to @xmath1 .
then @xmath107 , where @xmath135 is the induced metric on @xmath22 .
by definition of outermost barrier , for any point @xmath32 , spacelike extremal surfaces in ext@xmath8 can be deformed to either come arbitrarily close to @xmath1 or coincide with and be tangent to @xmath1 at @xmath32 . in the latter case
, we can simply take the limiting surface that touches @xmath1 . in any small neighborhood where @xmath1 and one of these spacelike extremal surfaces touch or nearly touch , eq .
[ negcond ] yields @xmath107 , where @xmath24 is the induced metric on the extremal surface , and we are done . in particular
, the extremal surface is spacelike and @xmath24 is riemannian , so in order for all of the components of @xmath20 to add up to a nonnegative number , at least one eigenvalue of @xmath136 must be nonnegative . whenever we can find a barrier of entirely negative extrinsic curvature
, we can therefore find a barrier of partly nonnegative curvature .
this is in agreement with theorem 2.1 , which implies that extremal surfaces can not come arbitrarily close to a surface with strictly negative curvature .
the existence of a barrier is therefore quite generic , since we expect it to occur for any surface @xmath1 which splits the spacetime into 2 regions and has negative or vanishing extrinsic curvature components .
many of the known analytic solutions that are alads admit a totally geodesic surface that acts an as extremal surface barrier , as well as multiple surfaces with negative extrinsic curvature .
totally geodesic surfaces are quite special and therefore spacetimes containing them are not representative of general alads spacetimes .
we expect , however , that small perturbations of spacetimes that admit splitting surfaces with negative extrinsic curvature do again result in spacetimes with splitting surfaces of nonpositive extrinsic curvature .
the existence of such surfaces is thus stable under small perturbations of the metric .
we therefore conclude that _ many generic alads spacetimes admit an extremal surface barrier_. moreover , we show in the examples below that the barrier often separates a region of proximity to a singularity from the boundary of the spacetime . in section [ singularities ]
, we further prove that for certain spacetimes , the existence of a compact outermost barrier implies the existence of singularities . the implications of this barrier in the context of ads / cft are disturbing : as rule of thumb , the best - understood aspect of the duality dictionary is generically limited in the scope of information it contains about the bulk .
this at best suggests that we must change our approach towards extracting bulk information by using probes that are not extremal surfaces . at worst indicates that complete information of a large class of bulk geometries
may simply not be contained in the dual field theory @xcite .
in this section , we provide some examples of spacetimes with barriers .
some of these barriers are totally geodesic , others have negative extrinsic curvature .
the latter condition is stable under small perturbations , although the former is not .
the simplest example is pure ads itself .
if we consider the metric on the poincar patch : @xmath137 surfaces of constant @xmath138 or surfaces of constant @xmath139 are totally geodesic .
since extremal surfaces in pure ads remain on constant time slices , it is clear that they do not cross constant @xmath138 totally geodesic surfaces .
similarly , a surface which is anchored between @xmath140 and @xmath141 will not propagate to @xmath142 or @xmath143 .
the next example we consider is an isotropic ads cosmology .
the metric on global ads can be written , via a coordinate transformation ( see e.g. @xcite ) , as an open friedmann - robertson - walker ( frw ) universe in the interior of the lightcone , and foliated into de sitter slices in the exterior of the lightcone : @xmath144 where the two are related via an analytic continuation ( see fig . [ frw ] ) .
any small perturbation to this metric results in a curvature singularity replacing the coordinate singularity .
if we couple the metric to , say , a scalar field , as in @xcite , the resulting geometry features a curvature singularity ( the first indication of some sort of an accumulation surface for codimension 2 surfaces in this geometry was found in @xcite ) : @xmath145 where @xmath146 and @xmath147 at the lightcone and @xmath148 is a big - crunch collapse .
it was found in @xcite that this spacetime contains a totally geodesic barrier : there is a point @xmath149 at which @xmath150 reaches a maximum .
the extrinsic curvature of any constant @xmath138 surface with respect to the past - pointing timelike normal is given by : @xmath151 where @xmath152 is the frw metric inside the lightcone .
the constant @xmath153 surface @xmath1 is therefore totally geodesic , and surfaces with @xmath154 have negative extrinsic curvature .
we demonstrate the barrier at work here using spacelike radial geodesics as an example of extremal surface probes .
this is easily generalized to extremal surfaces with fewer symmetries , but for brevity we limit ourselves to the simplest case . consider a geodesic with endpoints at @xmath155 and @xmath156 , where @xmath125 is one of the suppressed angles in eq .
the length functional for this geodesic within the frw region is : @xmath157 where we have chosen @xmath158 to parametrize the geodesic .
the boundary conditions @xmath155 and @xmath156 imply @xmath159 .
we solve for the geodesic from the lightcone interior outwards , starting from @xmath160 to the lightcone and the exterior region , using the reflection symmetry in @xmath125 along the geodesic path . since @xmath158 is a cyclic coordinate ,
the hamiltonian is conserved : @xmath161 solving for @xmath162 yields : @xmath163 since the scale factor vanishes at the lightcone and at the singularity , and reaches its maximum at some intermediate time @xmath153 , it follows from eq .
[ tprime ] that any geodesic with @xmath164 must propagate towards progressively larger @xmath138 .
geodesics with @xmath165 stay on @xmath166 for all @xmath158 .
this is precisely what theorem 2.2 states : extremal surfaces in the interior of @xmath1 ( i.e. @xmath154 ) remain in the interior , extremal surfaces in the exterior of @xmath1 ( i.e. @xmath167 ) remain in the exterior , and extremal surfaces that coincide with @xmath1 ( i.e. @xmath153 ) lie on @xmath1 , where @xmath1 is the totally geodesic surface given by @xmath153 .
one implication of this result is that , in the isotropic ads cosmology , only geodesics on the exterior of @xmath1 make it to the lightcone at @xmath168 and from there to the boundary .
the totally geodesic surface prevents boundary - anchored probes from getting arbitrarily close to the singularity .
another prominent example of a spacetime with a totally geodesic barrier surface as prescribed in section [ theorem ] includes any black hole with a stationary horizon @xcite .
this category includes alads spacetimes such as schwarzschild - ads , rn - ads , kerr - ads , the btz black hole , the planar ads black hole ( the horizon in this case was separately proven to be a barrier surface in @xcite ) , and the hyperbolic ads black hole . in particular , in ads - schwarzschild , there are two totally geodesic surfaces : the past- and future - directed horizons , denoted @xmath169 and @xmath170 in fig .
spacelike extremal surfaces anchored at @xmath171 must therefore always remain in region 1 , while those anchored at @xmath172 must remain in region 3 of the spacetime .
extremal surfaces anchored at both @xmath171 and @xmath172 are not constrained by our theorems to remain in any part of the spacetime .
in fact , @xcite found that spacelike geodesics with one endpoint at @xmath171 and one endpoint at @xmath172 can probe arbitrarily close to the singularity , and in particular , can cross both @xmath169 and @xmath170 .
we emphasize that this does not contradict our results in the previous section because these geodesics are anchored to two boundaries . as a point of interest
, @xcite found a barrier surface , or accumulation surface , for codimension 2 extremal surfaces anchored to both boundaries , while , as noted above , radial geodesics observe no such barrier .
, has just one boundary , does not have a barrier for all extremal surfaces because spacelike geodesics which enter at early times can leave at arbitrarily late times .
however , there is an effective barrier for surfaces that are anchored to the boundary at late times , and there may be a barrier for codimension 2 surfaces . following the first version of this paper , an analysis of the behavior of extremal surfaces in some vaidya spacetimes was done in @xcite . ]
it is clear at this point that spacetimes with singularities seem to have a particular proclivity for admitting barriers .
this is not a coincidence : corollary 2.6 states that null surfaces foliated by ( marginally ) trapped surfaces act as barriers at least for codimension 2 extremal surfaces .
if we additionally assume the stronger statement that _ all _ components of the null extrinsic curvature are nonpositive , we found in corollary 2.4 . that it is a barrier to extremal surfaces of every dimension .
in particular , when the null barrier @xmath1 is spatially compact , i.e. the entire boundary is in ext@xmath8 , then @xmath1 is ruled by trapped surfaces .
if the spacetime is furthermore globally ( or ads ) hyperbolic , spatially noncompact , and obeys the null curvature condition , the existence of trapped surfaces guarantees the existence of singularities and horizons @xcite . in the next section
, we will prove a partial converse of our conclusions from corollary 2.4 : barriers imply the existence of trapped surfaces or singularities , at least for spacetimes which admit a totally geodesic spacelike slice .
it is worthwhile here to comment on any bearing this might have on the recent controversy over the completeness of the ads / cft correspondence .
the question of whether the black interior is fully described by the boundary field theory is particularly relevant to this discussion . in @xcite , it was shown that there exist operators in the static black hole interior that can not exist in the dual field theory .
the fact that a well - used probe of the bulk geometry can not reach into a bulk region which is not completely described by the boundary field theory may not be coincidental . since the limited reach of extremal surfaces in the static black hole geometry is simply a special case of the barrier theorems presented in section [ theorem ] , we may expect that other black hole alads geometries admitting barriers may manifest similar incompleteness on the dual field theory side .
one may also speculate that field theories dual to other geometries , not necessarily black holes , may exhibit the same behavior near a singularity .
the examples in the previous section suggest that there is some connection between singularities ( particularly singularities masked by horizons ) , and extremal surface barriers . in this section
, we will prove this result in some special cases .
consider a spacetime manifold @xmath7 containing a totally geodesic spacelike slice @xmath173 .
such slices can be found , for instance , in static spacetimes , or more generally , spacetimes with a moment of time reflection symmetry .
since we intend to drop in extremal surfaces from the boundary of @xmath173 , we must assume that @xmath173 is noncompact . because @xmath173 is totally geodesic , these extremal surfaces remain on @xmath173 .
suppose further that there is an outermost barrier @xmath1 such that the intersection between @xmath1 and @xmath173 is nonempty and compact .
then we prove below that @xmath173 either admits a singularity , or else a marginally trapped surface . if , in addition , we assume global hyperbolicity ( including alads spacetimes with the appropriate generalization of global hyperbolicity ) , the null curvature condition ( @xmath174 for all null vectors @xmath175 ) and the generic condition ( roughly , that each null ray encounters at least a little bit of null curvature or shear ) then the existence of a marginally trapped surface itself implies the existence of a singularity somewhere on @xmath7 @xcite . we first prove a lemma : + * theorem 4.1 : * no compact extremal surface barrier @xmath1 in @xmath7 ever touches extremal surfaces anchored to @xmath3 .
we prove this by contradiction .
let @xmath1 be a barrier , and suppose there exists a spacelike extremal surface @xmath22 which is anchored to @xmath3 . by assumption , @xmath176 everywhere except where it coincides . in any neighborhood of such a coincident point @xmath32
, there exists a point such that : @xmath177 where @xmath178 is the null generator of @xmath1 for null @xmath1 , and the @xmath22-directed normal to @xmath1 otherwise .
@xmath22 can not lie on @xmath1 for a continuous neighborhood ( otherwise it would lie on @xmath1 everywhere , which would violate the boundary conditions on @xmath22 ) .
@xmath22 can therefore touch @xmath1 while remaining anchored on the boundary .
but then we can slightly deform @xmath22 in a neighborhood of @xmath32 , to make a new surface @xmath179 which crosses @xmath1 . using the elliptical equation of extremality
, we can solve for @xmath179 outside the neighborhood as well . because the perturbation is small , @xmath179
must still be anchored to the boundary .
this shows that @xmath1 is not a barrier , and we have arrived at a contradiction .
theorem 4.1 may at first seem to indicate that compact outermost barriers can not exist , since we have defined outermost barriers as the boundary of the union of some set of extremal surfaces : one might naturally ask how the outermost barrier is constructed , if extremal surfaces can not touch it .
the barrier must be constructed as a limit of extremal surfaces that come arbitrarily close to touching it .
we assume therefore , that one can find sequences of extremal surfaces approaching any point @xmath32 , whose limit is an extremal surface tangent to @xmath1 at @xmath32 .
however , the limiting extremal surface must not be anchored anymore to the boundary , or it would contradict theorem 4.1 .
we will make use of this limit construction of the outermost barrier in the proof of theorem 4.2 below , where we prove a condition relating the existence of barriers to the presence of singularities and trapped surfaces . because we are taking a limit of extremal surfaces
, one may worry that this limit will be related to the boundary in a bizarre way , perhaps by spiraling around , getting arbitrarily close to the boundary without actually being anchored to it .
we will deal with these pathological cases by including them in the following definition : a surface @xmath22 is _ weakly anchored _ if there exists a @xmath180 such that all points @xmath181 are less than distance @xmath182 away from the boundary @xmath3 along @xmath22 ( after compactification of @xmath3 ) . with a timelike barrier @xmath1 , projected onto a spacelike codimension 1 totally geodesic
slice @xmath173 yields a compact barrier @xmath183 for extremal surfaces on @xmath173.,width=188 ] * theorem 4.2 : * let @xmath7 admit a totally geodesic codimension 1 spacelike slice @xmath173
. are a finite distance away from the boundary after compactification . ] given a compact , nonempty , outermost barrier @xmath183 for weakly - anchored , codimension 2 , @xmath183-deformable ( i.e. where the deformations are restricted to @xmath173 ) spacelike extremal surfaces on @xmath173 ( e.g. fig . [ full ] ) , then one of the following is true :
1 . there exists a singularity outside or on @xmath183 , or 2 .
@xmath183 is a marginally trapped surface ( in either time direction ) .
since @xmath183 is an outermost barrier , there exist extremal surfaces that either coincide with it and are tangent to it at coincident points , or come arbitrarily close to coinciding with it .
the former case is ruled out by theorem 4.1 .
let @xmath10 be a family of spacelike extremal surfaces that approach arbitrarily close to some point @xmath184 while remaining weakly anchored on @xmath185 . in a neighborhood of @xmath186
, we take the limit of these surfaces @xmath179 of @xmath187 that reaches @xmath1 at @xmath186 .
since we are operating at the `` physics '' standard of rigor , we assume without proof that ( a ) this limit must exist for some sequence of extremal surfaces , and that ( b ) the resulting extremal surface may be extended outside the neighborhood of @xmath186 by solving the elliptical equation of extremality .
when we solve this equation , by theorem 4.1 , @xmath179 must no longer be weakly anchored on @xmath188 .
since @xmath179 can not be weakly anchored to @xmath188 , part of it must terminate somewhere in the interior of the spacetime .
it can do it in one of the following two ways ( fig .
[ thm3 ] ) : ( a ) : : @xmath179 terminates in a singularity on or outside @xmath183 , or ( b ) : : @xmath179 spirals outside @xmath183 forever , without ever coming to an end . in that case , we can define @xmath189 to be the limit set of the extremal surface as it gets arbitrarily far away from the points where it is anchored to the boundary . is not ( weakly ) anchored to the boundary anywhere , then @xmath190 . ]
@xmath189 must be bounded by some compact innermost surface @xmath191 , where @xmath192 .
can not come arbitrarily close to the boundary , since then it would be boundary - anchored by the generalized definition above . ]
if ( a ) is true , we get the first part of the claim .
we now show that ( b ) results in the second part of the claim .
to do so , we need to show that @xmath193 .
we do this by contradiction .
suppose @xmath194 .
then @xmath195 .
this means that it is _ not _ a barrier for extremal surfaces weakly anchored at @xmath3 , and therefore by theorem 2.1 there must exist a point @xmath196 such that @xmath197 on the other hand , by construction , @xmath198 can come arbitrarily close to @xmath191 at any point @xmath199 .
this implies that the following holds everywhere : @xmath200 ( the proof is essentially the same as theorem 2.7 , but reversing the roles of the interior and exterior . ) we have arrived at a contradiction , so instead we must assume that @xmath201 .
this tells us that @xmath179 must in fact coincide with @xmath183 .
since @xmath179 is extremal , it follows that @xmath183 satisfies @xmath202 this is the same as saying that @xmath183 is a marginally trapped surface , which proves the second part of the claim .
the assumption that there exists an outermost barrier @xmath183 for codimension 2 extremal surfaces living on @xmath173 is reasonable : if we take @xmath7 to have a totally geodesic slice @xmath173 and a compact outermost barrier @xmath1 such that @xmath203 , then in general , we expect that @xmath204 will be compact , and be a barrier for extremal surfaces on @xmath173 . there should then exist some outermost @xmath183 which , at least for alads spacetimes , will still be compact .
( if there is a barrier for _ all _ extremal surfaces , then obviously there must be a barrier for those which are codimension 2 and restricted to @xmath173 . ) if we further assume global hyperbolicity , the null curvature condition , and the generic condition , we are able to say that marginally trapped surfaces only occur in geodesically incomplete spacetimes , so that barriers in these spacetimes exist only in the presence of singularities @xcite .
singularities are therefore intrinsically linked to extremal surface barriers . assuming some form of cosmic censorship , these singularities must be hidden behind horizons , thus outermost barriers are also linked to horizons .
in particular , the direct application of theorem 4.2 to the ads / cft correspondence suggests that in general , extremal surface probes are not good probes for spacetimes with singularities .
we expect that theorem 4.2 can be further generalized to both the noncompact barrier case , as section [ cosmology ] suggests , and to any time - dependent geometry , although this may require assuming the null curvature condition ( implied by the null energy condition and einstein s equations ) , even to prove the existence of the trapped surface .
we proved in theorem 2.1 , theorem 2.2 , and corollary 2.4 that a splitting surface with nonpositive extrinsic curvature acts as a barrier to extremal surfaces ( at least those which are @xmath1-deformable ) .
this can include spacelike , timelike , or null barrier surfaces .
if we are only interested in barriers to codimension 2 surfaces ( as used in the holographic entanglement entropy conjecture @xcite ) , then we have shown in theorem 2.5 and corollary 2.6 that null surfaces foliated by ( marginally ) trapped surfaces are barriers .
conversely , for spacetimes with a totally geodesic slice , the existence of a barrier guarantees the existence of either singularities or else trapped surfaces for spacetimes with a totally geodesic slice @xmath173 , so long as the intersection of the barrier and @xmath173 is compact .
we also showed that outermost barriers have at least one nonnegative extrinsic curvature component ( theorem 2.7 ) , and we argued that extremal surface barriers occur somewhat generically in alads spacetimes .
this presents a setback to attempts to reconstruct the bulk geometries from field theory operators using extremal surface probes .
these results are particularly troubling in light of how pervasive extremal surface probes are in applications of the ads / cft correspondence .
if there are regions which can not be reconstructed from the entire boundary , than either those regions do not really exist as semiclassical regions ( e.g. because there is a firewall at their boundary ) , or else if they do exist , the information inside of them must be contained in a new factor of the hilbert space which is in addition to the boundary @xcite .
we have also studied cases in which there is a barrier to extremal surfaces located on only _
part _ of the boundary . in this case
it is not surprising that the entire bulk can not be reconstructed .
thus the location of the barrier might give clues about how much of the bulk can be reconstructed from a cft region .
see @xcite for discussion of this question .
one should bear in mind that extremal surfaces are not the only probe used in ads / cft .
if there are other bulk probes besides extremal surfaces , it is possible that a firewall ( if it exists at all ) might be located somewhere behind the outermost barrier .
conversely , if there are some extremal surfaces which do not actually correspond to dual observables ( for example , because they are not the _ minimum _ area extremal surface ) , then it could conceivably be that the firewall is actually outside of the barrier
. it may also be that there are regions of space where _ some _ kinds of extremal surfaces can probe , but not enough to fully reconstruct the geometry .
it is unclear what the status of these regions would be . in the case of a vaidya spacetime
where a black hole forms from the collapse of a shell , there is no extremal surface barrier . in particular
, there are spacelike geodesics which pass through any point of the black hole interior .
the amps argument for firewalls @xcite applies to sufficiently old black holes that form from collapse , yet the absence of a barrier might seem to suggest that there can be no firewall
. however , the spacelike geodesics in question are anchored at very large time separation on the boundary .
in fact , @xcite argued that the wkb stationary phase approximation is not dominated by spacelike geodesics passing through the interior of the black hole , but rather by complex geodesics .
it is unclear how to use the barrier results proven here in the case where the extremal surfaces are complex , as was found in @xcite for two - point correlators in the case of an ads - schwarzschild bulk .
however , even for the vaidya spacetime , we still find that there is a partial barrier , for spacelike surfaces anchored sufficiently far to the future of the collapsing shell .
this is because the spacetime is just schwarzschild at late times .
thus the extremal surface barrier might still give clues about the location of the firewall , although it does not necessarily tell us when the firewall would first appear .
in general , our knowledge of the ads / cft dictionary comes from a bootstrapping procedure where ( a ) proposed new ingredients to the dictionary can be checked using our knowledge of bulk physics , while ( b ) the behavior of the bulk can also be predicted using the dictionary .
but if at step ( b ) , our current understanding of the dictionary predicts a firewall resulting in a breakdown of the bulk equations of motion used at step ( a)then it is unclear how to proceed . should we modify the bulk equations of motion or the dictionary ?
perhaps it is the local bulk equations of motion which should be kept sacrosanct , and other things we think we know about ads / cft should be modified in order to preserve them .
ultimately , the question is what choice of dictionary leads to the most consistent , complete , and interesting form of the correspondence .
there are several remaining questions worth noting .
first , we have yet to find a complete description of the most general barrier , in particular for the outermost barrier surface .
such a description would facilitate an understanding of precisely how generic extremal surface barriers are .
we expect that , assuming the null curvature condition , it should be possible to show that compact outermost barriers can only occur in spacetimes with singularities .
we note that ( aside from theorem 2.5 ) the main results of this paper are purely geometrical in nature
they apply to any lorentzian spacetime without needing to use of any energy condition restricting the sign of the curvature .
( although given the existence of trapped surfaces as shown in theorem 4.2 , the null curvature condition would say that there must be singularities . )
this suggests that our results are still valid even in semiclassical spacetimes ( e.g. evaporating black holes ) where local energy conditions are violated .
however , the significance of these geometrical results for ads / cft is not the same , since in the semiclassical ( finite @xmath22 ) regime , there are quantum corrections to the dictionary relating extremal surfaces to boundary physics .
for example , the entanglement entropy of a cft region is no longer given just by the area of an extremal surface .
instead it receives corrections due to the bulk entanglement entropy @xcite .
thus we should really be interested in surfaces which extremize the area _ plus _ the bulk entanglement entropy . to find barriers to these surfaces
, it is natural to look for null surfaces foliated by quantum trapped surfaces , where the sum of the area and the bulk entropy is decreasing .
one can then prove quantum generalizations of the types of classical theorems proven here @xcite .
+ we thank dalit engelhardt , sebastian fischetti , ahmed almheiri , gary horowitz , don marolf , william kelly , and mudassir moosa for helpful discussions .
this work is supported in part by the national science foundation under grant no .
phy12 - 05500 .
ne is also supported by the national science foundation graduate research fellowship program under grant no .
aw is also supported by the simons foundation .
a. almheiri , d. marolf , j. polchinski and j. sully , `` black holes : complementarity or firewalls ? , '' jhep * 1302 * , 062 ( 2013 ) [ arxiv:1207.3123 [ hep - th ] ] .
a. almheiri , d. marolf , j. polchinski , d. stanford and j. sully , `` an apologia for firewalls , '' jhep * 1309 * ( 2013 ) 018 [ arxiv:1304.6483 [ hep - th ] ] .
d. marolf and j. polchinski , `` gauge / gravity duality and the black hole interior , '' phys .
lett . * 111 * , 171301 ( 2013 ) [ arxiv:1307.4706 [ hep - th ] ] .
k. papadodimas and s. raju , `` state - dependent bulk - boundary maps and black hole complementarity , '' arxiv:1310.6335 [ hep - th ] .
s. b. giddings , `` black holes , quantum information , and unitary evolution , '' phys . rev .
d * 85 * , 124063 ( 2012 ) [ arxiv:1201.1037 [ hep - th ] ] .
j. maldacena and l. susskind , `` cool horizons for entangled black holes , '' arxiv:1306.0533 [ hep - th ] .
r. bousso , `` complementarity is not enough , '' phys .
d * 87 * , 124023 ( 2013 ) [ arxiv:1207.5192 [ hep - th ] ] .
k. papadodimas and s. raju , `` an infalling observer in ads / cft , '' jhep * 1310 * , 212 ( 2013 ) [ arxiv:1211.6767 [ hep - th ] ] . s. g. avery , b. d. chowdhury and a. puhm , `` unitarity and fuzzball complementarity : alice fuzzes but may not even know it! , '' jhep * 1309 * , 012 ( 2013 ) [ arxiv:1210.6996 [ hep - th ] ] .
p. kraus , h. ooguri and s. shenker , `` inside the horizon with ads / cft , '' phys .
d * 67 * , 124022 ( 2003 ) [ hep - th/0212277 ] . s. ryu and t. takayanagi , `` aspects of holographic entanglement entropy , '' jhep * 0608 * , 045 ( 2006 ) [ hep - th/0605073 ] .
v. e. hubeny , m. rangamani and t. takayanagi , `` a covariant holographic entanglement entropy proposal , '' jhep * 0707 * , 062 ( 2007 ) [ arxiv:0705.0016 [ hep - th ] ] .
j. m. maldacena , `` wilson loops in large n field theories , '' phys .
* 80 * , 4859 ( 1998 ) [ hep - th/9803002 ] .
v. e. hubeny , `` extremal surfaces as bulk probes in ads
/ cft , '' jhep * 1207 * , 093 ( 2012 ) [ arxiv:1203.1044 [ hep - th ] ] . s. s. pal , `` extremal surfaces and entanglement entropy , '' arxiv:1312.0088 [ hep - th ] .
l. fidkowski , v. hubeny , m. kleban and s. shenker , `` the black hole singularity in ads / cft , '' jhep * 0402 * , 014 ( 2004 ) [ hep - th/0306170 ] . t. hartman and j. maldacena , `` time evolution of entanglement entropy from black hole interiors , '' jhep * 1305 * , 014 ( 2013 ) [ arxiv:1303.1080 [ hep - th ] ] .
j. maldacena and g. l. pimentel , `` entanglement entropy in de sitter space , '' jhep * 1302 * , 038 ( 2013 ) [ arxiv:1210.7244 [ hep - th ] ] .
n. engelhardt and g. t. horowitz , `` entanglement entropy near cosmological singularities , '' jhep * 1306 * , 041 ( 2013 ) [ arxiv:1303.4442 [ hep - th ] ] .
w. fischler , s. kundu and j. f. pedraza , `` entanglement and out - of - equilibrium dynamics in holographic models of de sitter qfts , '' arxiv:1311.5519 [ hep - th ] .
w. w. meeks , iii and s. t. yau , `` the existence of embedded minimal surfaces and the problem of uniqueness , '' _ math .
179(2 ) : 151 - 168 , 1982 .
d. marolf and a. c. wall , `` eternal black holes and superselection in ads / cft , '' class .
grav . *
30 * , 025001 ( 2013 ) [ arxiv:1210.3590 [ hep - th ] ] .
b. czech , j. l. karczmarek , f. nogueira and m. van raamsdonk , `` the gravity dual of a density matrix , '' class .
* 29 * , 155009 ( 2012 ) [ arxiv:1204.1330 [ hep - th ] ] . t. hertog and g. t. horowitz , jhep * 0407 * , 073 ( 2004 ) [ hep - th/0406134 ] .
unpublished manuscript , n. engelhardt , t. hertog , and g. horowitz .
m. mars , `` stability of mots in totally geodesic null horizons , '' class .
* 29 * , 145019 ( 2012 ) [ arxiv:1205.1724 [ gr - qc ] ] .
p. hjicek , `` exact models of charged black holes
. i. geometry of totally geodesic null hypersurface '' , commun .
phys . * 34 * , 37 - 52 ( 1973 ) .
v. e. hubeny and h. maxfield , `` holographic probes of collapsing black holes , '' arxiv:1312.6887 [ hep - th ] .
s. w. hawking and g. f. r. ellis , `` the large scale structure of spacetime , '' cambridge university press , cambridge , 1973 .
m. headrick and t. takayanagi , `` a holographic proof of the strong subadditivity of entanglement entropy , '' phys .
d * 76 * , 106013 ( 2007 ) [ arxiv:0704.3719 [ hep - th ] ] .
a. c. wall , `` maximin surfaces , and the strong subadditivity of the covariant holographic entanglement entropy , '' arxiv:1211.3494 [ hep - th ] . v. e. hubeny and m. rangamani , `` causal holographic information , '' jhep * 1206 * , 114 ( 2012 ) [ arxiv:1204.1698 [ hep - th ] ] .
r. bousso , s. leichenauer and v. rosenhaus , `` light - sheets and ads / cft , '' phys .
d * 86 * , 046009 ( 2012 ) [ arxiv:1203.6619 [ hep - th ] ] .
j. louko , d. marolf and s. f. ross , phys .
d * 62 * , 044041 ( 2000 ) [ hep - th/0002111 ]
. t. faulkner , a. lewkowycz and j. maldacena , `` quantum corrections to holographic entanglement entropy , '' jhep * 1311 * , 074 ( 2013 ) [ arxiv:1307.2892 [ hep - th ] ] . | we present a generic condition for lorentzian manifolds to have a barrier that limits the reach of boundary - anchored extremal surfaces of arbitrary dimension .
we show that any surface with nonpositive extrinsic curvature is a barrier , in the sense that extremal surfaces can not be continuously deformed past it . furthermore , the outermost barrier surface has nonnegative extrinsic curvature . under certain conditions ,
we show that the existence of trapped surfaces implies a barrier , and conversely . in the context of ads
/ cft , these barriers imply that it is impossible to reconstruct the entire bulk using extremal surfaces .
we comment on the implications for the firewall controversy . |
the peculiar galactic globular cluster ( ggc ) ( ngc5139 ) is currently subject to substantial observational efforts covering the whole wavelength spectrum . this gigantic star cluster ,
the most massive known in our galaxy , has ( at least ) three separate stellar populations with a large undisputed spread in age , metallicity ( fe , ca ) and kinematics ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
according to recent abundance measurements based on 400 medium resolution spectra collected by @xcite , it seems that the spread in metallicity among the stars is of the order of one dex ( @xmath3\lesssim -1 $ ] ) .
the metallicity distribution shows three well - defined peaks around @xmath4=-1.7 $ ] , @xmath5 and @xmath6 together with a few metal - rich stars at @xmath7 . on the basis of high - resolution spectra ,
it has been suggested by @xcite that stars in the anomalous branch @xcite might be more metal - rich , with a mean metallicity @xmath8\sim-0.5 $ ] .
a similar metal - rich tail was also detected by @xcite and by suntzeff & kraft ( 1996 , and references therein ) . in the absence of conformity in the literature for the rgb names ,
we here refer to the metal - poor component as , to the metal - intermediate as , and to the anomalous branch as .
results from new observations seem to pose as many new questions as they answer .
it has been suggested ( e.g. * ? ? ?
* ; * ? ? ?
* ) that the population might be significantly younger than .
however , in a recent investigation based on very large telescope ( vlt ) and hubble space telescope ( hst ) data , @xcite found that the population is at least as old as the population and probably a few gyrs older .
the observational scenario was further complicated by the results brought forward by @xcite on the basis of multi - band hst data .
they not only confirmed a bifurcation along the main sequence , but also found a series of different turn - offs ( tos ) and sub - giant branches ( sgbs ) .
the proposed explanations in the literature for these peculiar stellar populations in are many and include , e.g. , increased he content , or a separate group of stars located at a larger distance @xcite .
the latter hypothesis could be further supported by the occurrence of a tidal tail in but we still lack a firm empirical detection ( see * ? ? ?
* ; * ? ? ?
the formation history and composition of thus form a complex puzzle that is being slowly pieced together by investigations based on the latest generation of telescopes .
a well - known promising technique is to study evolved stars simultaneously in optical and nir bands to limit subtle errors due to the absolute calibration , to crowding in the innermost regions , and to reddening corrections .
the main aim of this letter is to investigate whether different assumptions concerning the spread in age and in chemical composition , or differences in distance and in reddening , account for observed optical ( @xmath9 ) and nir ( @xmath10 ) cmds .
near - infrared @xmath11 images of were collected in 2003 with sofi at the new technology telescope of eso , la silla .
the seeing conditions were good and range from 06 to 11 .
together with additional data from 2001 , available in the eso archive , we end up with a total nir sample of 92 @xmath12- and 135 @xmath13-band images that cover a @xmath14arcmin@xmath15 area centred on the cluster .
these data were reduced with daophotii / allframe following the same technique adopted by @xcite .
the nir catalogue includes @xmath16@xmath17 stars .
with fors1 ( standard - resolution mode ) on the antu / vlt telescope , we also collected optical @xmath18 images in 1999 .
these data are from a @xmath19 pointing mosaic centred on the cluster that covers an area slightly larger than @xmath20arcmin@xmath15 .
exposure times range from 10 ( @xmath21 ) to 30 @xmath22 ( @xmath23 ) and seeing conditions were better than 10 .
these data have also been reduced using daophotii / allframe and the photometric catalogue includes @xmath16@xmath24 stars .
optical @xmath25 and @xmath26 ( hereinafter @xmath27 and @xmath28 bands ) data were retrieved from the hst archive .
these images were collected with the acs instrument in 9 telescope pointings , each of which provides a @xmath29-image pair with exposure times of 12 ( @xmath27 ) and 8 @xmath22 ( @xmath28 ) .
the field covered by these data is @xmath16@xmath30arcmin@xmath15 , centred on the cluster .
the acs data were reduced with romafotwo and provide @xmath29 photometry for @xmath31 stars .
the photometry was kept in the vega system ( see e.g. , http://www.stsci.edu/hst/acs/documents ) .
a detailed description of observations and data reduction will be given in a future paper .
here we only wish to mention that raw frames were prereduced using standard iraf procedures and , in addition , the fors1 images were corrected for amplifier cross - talk by following @xcite . to improve the photometric accuracy , carefully chosen selection criteria
were applied to pinpoint a large number ( @xmath32100 ) of point - spread function stars across the individual frames , and several different reduction strategies were used to perform the photometry over the entire data set .
the observed fields were combined to the same geometrical system using iraf.immatch and daomatch / daomaster .
the absolute photometric calibration of ground - based instrumental magnitudes was performed using standard stars observed during the same nights .
the typical accuracy is [email protected]@xmath33mag for both optical and nir data .
figure 1 shows cmds based on the five different photometric bands .
the plotted stars were selected from individual catalogues by using the ` separation index ' _ sep _ introduced by @xcite , since crowding errors dominate the photometric errors . the adopted _ sep _ ranges from 6.5 ( nir data ) to 8 ( acs data ) , which corresponds to stars having less than 0.3% of their measured light contaminated by neighbour stars ; the higher the _ sep _ for a star , the less the severity of the crowding by neighbours .
error bars plotted at the left side of each panel account for photometric and calibration errors in magnitude and color .
the input physics adopted in our evolutionary code has been discussed in detail in a series of papers @xcite . here , we point out that the stellar models ( partly available at http://astro.df.unipi.it/saa/pel/z0.html ) account for atomic diffusion , including the effects of gravitational settling , and thermal diffusion with diffusion coefficients given by @xcite . the amount of original he is based on a primordial he abundance @xmath34 and a he - to - metals enrichment ratio of @xmath35 @xcite .
we adopt the solar mixture provided by @xcite . for details on the calibration of the mixing length parameter and on model validation
, we refer to @xcite and @xcite . to avoid deceptive uncertainties in the comparison between theory and observations , the predictions were transformed into both the @xmath29 vega system and the @xmath36 johnson - cousins bands by adopting the atmosphere models provided by @xcite and @xcite .
data plotted in fig .
1 display the comparison between observations and a set of four isochrones ( solid lines ) with the same age ( 12 gyr ) and different chemical compositions ( see labels ) .
a mean reddening of @xmath37 ( @xmath38 , @xcite ) was adopted and a true distance modulus of @xmath39 ( @xmath40 ; @xcite , @xmath41 , del principe et al .
2004 , private communication ) .
extinction parameters for both optical and nir bands have been estimated using the extinction model of @xcite .
the metal - poor and the metal - intermediate isochrones supply , within current empirical and theoretical uncertainties , a good fit to the bulk of rgb and hb stars .
close to the rgb tip , the isochrones are slightly brighter than the observed stars , which is caused by the adopted mixing - length parameter ( @xmath42 ) .
the sgb and the lower rgb are only marginally affected by this parameter , since the empirical isochrone calibration is based on these evolutionary phases @xcite
. however , the most metal - rich isochrone appears to be systematically redder than the stars at the base of the branch,@xmath28 plane by ferraro et al .
( 2004 , see their fig .
2 ) . ] ( green dots ) and fainter than the sgb stars . clearly , the shape of the sgb does not support metal - abundances @xmath43 , since to stars for more metal - rich populations become brighter than sgb stars . to further constrain this evidence , fig .
2 shows the comparison between observed hb stars ( the entire sample includes more than 2,300 objects ) and predicted zero age horizontal branches ( zahbs ) with a progenitor age of 12 gyr and different chemical compositions . for the sake of the comparison ,
the objects located between hot hb stars and rgb stars were selected in the @xmath44 , @xmath27 plane and plotted as blue dots .
theoretical predictions plotted in this figure show that the more metal - rich zahb appears to be systematically brighter than the observed hb stars for @xmath45 . moreover , the same zahb crosses the rgb region , thus suggesting that the occurrence of hb stars more metal - rich than z=0.002 should appear as an anomalous bump along the rgb .
data plotted in these figures disclose that _ the stellar populations in can not be explained with a ranking in metal abundance_. in line with @xcite and @xcite , we find that plausible changes in the cluster distance , reddening , age , and chemical composition do not supply a reasonable simultaneous fit of the , , and branches .
to further investigate the nature of the branch , we performed a series of tests by changing the metallicity , the cluster age , and the distance .
a good fit to the anomalous branch is possible by adopting the same reddening as in fig.[fig : apjfig1 ] , a true distance modulus of @xmath46 , and an isochrone of 15.5 gyr with @xmath47 and @xmath48 .
figure[fig : apjfig3 ] shows that these assumptions supply a good fit in both the optical and the nir bands , and indeed the current isochrone properly fits the width in color of the sub - giant branch and the shape of a good fraction of the rgb .
moreover , the zahb for the same chemical composition ( dashed line ) agrees quite well with the faint component of hb stars .
note that more metal - poor zahbs at the canonical distance do not supply a reasonable fit of the yellow spur stars with @xmath49 .
the yellow spur is visible in all planes , and in the nir ( * c*,*d * ) it even splits up in brightness due to a stronger sensitivity to effective temperature .
the identification of the entire sample was checked on individual images and , once confirmed by independent measurements , the data indicate a separate hb sequence for the population .
although the fit may appear good , it implies an increase in distance of @xmath50 and an unreasonable @xmath51 gyr increase in age .
this estimate is at variance with the absolute age estimates of ggcs @xcite and with cmb measurements by w - map @xcite .
the discrepancy becomes even larger if we account for the fact that this isochrone was constructed by adopting a he abundance slightly higher ( @xmath48 , vs 0.238 ) than estimated from a he - to - metals enrichment ratio @xmath52 .
this increase implies a decrease in age of @xmath16 1 gyr .
moreover , we are performing a differential age estimate , and therefore if we account for uncertainties : in the input physics of the evolutionary models ( e.g. equation of state , opacity ) ; in the efficiency of macroscopic mechanisms ( like diffusion ) ; in model atmospheres applied in the transformation of the models into the observational plane ; and in the extinction models , we end up with an uncertainty of @xmath53 gyr in cluster age @xcite . owing to the wide range of chemical compositions and stellar ages adopted in the literature for explaining the morphology of the branch
, we decided to investigate whether different combinations of assumed values of distance , chemical composition , and reddening may also simultaneously account for the branch and the hb stars .
we found that two isochrones of 13 gyr for @xmath54 and @xmath55 bracket the stars ( see fig.4 ) , within empirical and theoretical uncertainties .
the fit was obtained using the same true distance modulus adopted in fig.[fig : apjfig3 ] , together with a mild increase in reddening , @xmath2 .
once again theory agrees reasonably well with observations in all color planes .
moreover , data plotted in fig . 5 show that the predicted zahb with z=0.0015 and z=0.003 , for the adopted distance modulus and reddening correction , account for the yellow spur stars .
the same figure shows a sample of 53 rr lyrae stars selected from the variable - star catalogue by @xcite for which we have a good coverage of @xmath12- and @xmath56-band light curves ( del principe et al .
2004 , private communication ) .
the rr lyrae stars only account for a tiny fraction of the yellow - spur stars .
this finding together with the detection a well - defined sequence in panels * b*,*c * , and * d * indicate that this spur might be the hb associated with the population .
note that a few of these ` -hb ' spur stars have also been detected by rey et al .
( 2004 , see their fig .
7 , for @xmath57 ) and by sollima et al .
( 2004a , see their figs .
6 and 12 ) .
we have presented a new set of multi - band photometric data for the ggc and in agreement with previous findings in the literature we find no acceptable fit to the different stellar populations for a single distance , reddening , and age .
we found two reasonable fits for the stars : ( 1 ) by adopting a 0.2 higher distance modulus ( @xmath58 500 pc ) , a metal - intermediate composition ( @xmath47 , @xmath59 ) , and an unreasonable increase in age of @xmath164 gyr ; or ( 2 ) for the same @xmath60 shift , an increase in the reddening , metal - intermediate chemical compositions ( @xmath61 , @xmath62 ) , and an age that , within current uncertainties , is coeval with the bulk of the stars .
current findings indicate that stars are not significantly more metal - rich than @xmath55 .
this evidence is supported by the shape of the sgb , as already suggested by @xcite and by the fit of hb stars .
we are in favour of the latter solution for the following reasons : @xmath63 the difference in distance between the branch and the bulk of stars is of the order of 10% .
this estimate is 34 times smaller than the estimate by @xcite and in very good agreement with the distance of the density maxima detected by @xcite along the tidal tails of the ggc pal 5 .
moreover , recent n - body simulations ( capuzzo dolcetta et al .
2004 ) indicate that clumps along the tidal tails can approximately include 10% of the cluster mass .
@xmath63 we found that by artificially shifting the -branch stars to account for the assumed difference in distance and reddening , they overlap with the population .
it has been recently suggested by piotto et al .
( 2004 ) , on the basis of low - resolution spectra , that the bluer main sequence detected by bedin et al .
( 2004 ) is more metal - rich than the red main sequence .
unfortunately , current acs photometry is only based on shallow acs exposures , and therefore we can not properly identify in our data these stellar populations located in the lower main sequence .
the same outcome applies to the suspected extremely - hot hb progeny of the bluer main sequence , since they have not been detected in the nir bands .
@xmath63 current preliminary findings support recent n - body simulations by chiba & mizutani ( 2004 ) and by ideta & makino ( 2004 ) .
in particular , the latter authors found , by assuming that the progenitor of omega centauri is a dwarf galaxy , that more than 90% of its stellar content was lost during the first few pericenter passages ( see their fig .
2 ) . @xmath63 reddening estimates in the literature cluster around @xmath64 .
however , the map from @xcite indicates reddening variations of 0.02 across the body of the cluster while , more importantly , 2mass data @xcite show a very clumpy reddening distribution at distances beyond 1@xmath65 ( 100@xmath33pc ) from the cluster centre , with large variations @xmath66 across a 4degrees@xmath15 area .
however , the inference of reddening _ between _ the main body of and the supposed background population is very surprising . lies at the comparatively low galactic latitude of @xmath67 , and the reddening variations seen in the schlegel and 2mass maps likely originate in the foreground interstellar material of the galactic disk .
any interstellar material behind must lie at least 1.4@xmath33kpc from the galactic plane , and therefore would most likely be associated with itself .
smith et al .
( 1990 ) have reported a significant detection of @xmath68 in the direction of , blueshifted by @xmath69 km s@xmath70 with respect to the cluster velocity .
however , their interpretation is that this gas is associated with the northern extension of the magellanic stream far beyond the cluster .
it would be a remarkable coincidence if this interstellar material happened to lie between the cluster and a clump in the tidal tail at a 10% greater distance than the main cluster body , while traveling at the quoted relative speed .
however , it is worth noting that smith et al .
( 1990 ) estimated for this cloud a column density of @xmath71 atmos / cm@xmath15 , and therefore a reddening @xmath73 ( predehl & schmitt 1995 ) that is at least a factor of two larger than required by our best fits .
this point is crucial for the proposed explanation and needs to be further investigated .
finally , we note that the comparison between predicted zahbs and hb stars indicates that the occurrence of an old stellar population with @xmath74 would imply the occurrence of an anomalous clump along the rgb .
in fact , more metal - rich , red hb stars cover the same color range of metal - poor rgb stars .
the detection of such a feature along the rgb can supply robust constraints on the progeny of the stellar population .
these results , when independently confirmed , would suggest that the branch might be a clump of stars located 500 pc beyond the bulk of the cluster . no firm conclusion can be drawn on the basis of current data , although this evidence together with the increase in radial velocity among stars measured by @xcite and numerical simulations recently provided by capuzzo dolcetta et al .
( 2004 ) indicates that it could be a tidal tail .
it is a pleasure to thank m. sirianni and n. panagia for fundamental suggestions on the calibration of acs data .
we wish to tank two anonymous referees for their suggestions and pertinent comments that helped us to improve the content and the readability of the manuscript .
we are also grateful to v. castellani for his critical reading of an early version of this manuscript .
this work was supported by the belgian fund for scientific research ( fwo ) in the framework the project
iap p5/36 of the belgian federal science policy , by miur / prin 2003 in the framework of the project : continuity and discontinuity in the galaxy formation , and by danish nsrc in the framework of the project : structure and evolution of stars : new insight from eclipsing binaries and pulsating stars .
bedin , l.r . ,
piotto , g. , anderson , j. , cassisi , s. et al .
2004 , , 605 , 125 bennett , c.l . ,
halpern , m. , hinshaw , g. , jarosik , n. et al .
2003 , , 148 , 1 capuzzo dolcetta , r. , di matteo , p. , & miocchi , p. 2004
, , accepted , astro - ph/0406313 cardelli , j.a . ,
clayton , g.c . ,
mathis , j.s .
1989 , , 345 , 245 cariulo , p. , deglinnocenti s. , castellani , v. , 2004 , a&a 421 , 1121 cassisi s. , castellani v. , deglinnocenti s. & weiss a. 1998 , a&as , 129 , 267 castellani , v. , & deglinnocenti , s. 1999 , , 344 , 97 castellani , v. , deglinnocenti , s. , marconi , m. , prada moroni , p.g . & sestito , p. 2003
, a&a , 404 , 645 castelli , f. 1999 , a&a , 346 , 564 castelli , f. , gratton , r. g. , & kurucz , r. l. 1997 , , 318 , 841 chiba , m. , & mizutani , a. 2004 , pasa , 21 , 237 dallora , m. , storm , j. , bono , g. , ripepi , v. et al . , 2004 , , 610 , 269 ferraro , f.r .
, bellazzini , m. , pancino , e. 2002 , , 573 , 95 ferraro , f.r . , sollima , a. , pancino , e. , bellazzini , m. , et al .
2004 , , 603 , 81 freyhammer , l.m . , andersen , m.i . ,
arentoft , t. , sterken , c. , nrregaard , p. 2001 , experimental astronomy , 12 , 147 gratton , r.g . , bragaglia , a. , carretta , e. , clementini , g. , desidera , s. , grundahl , f. , lucatello , s. 2003 , a&a , 408 , 529 hilker , m. & richtler , t. , 2000 , , 362 , 895 hilker , m. , kayser , a. , richtler , t. , willemsen , p. 2004
, , 422 , 9 hughes , j. ; wallerstein , g. ; van leeuwen , f. ; & hilker , m. 2004 , aj , 127 , 980 ideta , m. , & makino , j. 2004 , apj , 616 , l107 kaluzny , j. , olech , a. , thompson , i.b . , pych , w. , krzeminski , w. , schwarzenberg - czerny , a. 2004 , , 424 , 1101 krauss , l. m. & chaboyer , b. 2003 , sci , 299 , 65 law , d.r . , majewski , s.r . ,
skrutskie , m.f . ,
carpenter , j.m .
2003 , , 126 , 1871 lee , y .- w . ,
m . , sohn , y .-
j . , rey , s .- c .
1999 , nature , 402 , 55 leon , s. , meylan , g. , combes , f. 2000 , , 359 , 907 lub , j. , 2002 , in asp conf . ser .
265 , omega centauri : a unique window into astrophysics , ed . f. van leeuwen , j. hughes and g. piotto ( s. francisco : asp ) , 95 noels , a. & grevesse , n. 1993 in origin and evolution of the elements , ed .
n. prantzos , e. vangioni - flam , m. casse ` ( cambridge : cambridge univ . press ) , 15 norris , j.e .
, 2004 , , 612 , 25 norris , j.e . ,
freeman , k. c. ; mighell , k. j 1996 , , 462 , 241 odenkirchen , m. et al .
2003 , aj , 126 , 2385 pagel , b.e.j .
& portinari , l. 1998 , mnras , 298 , 747 pancino , e. , 2004 , in carnegie obs . astrophysics ser .
4 , origin and evolution of the elements , ed . a. mcwilliam and m. rauch ( cambridge : cambridge univ .
press ) , 45 predehl , p. , & schmitt , j. h. m. m. 1995 , a&a , 293 , 889 rey , s .- c . , lee , y .- w .
, ree , c.h . ,
m . , sohn , y .-
j . , walker , a.r .
, 2004 , aj , 127 , 958 schlegel , d.j . , finkbeiner , d.p . , davis , m. , 1998 , , 500 , 525 smith , v.v .
2004 , in carnegie obs .
astrophysics ser .
4 , origin and evolution of the elements , ed .
a. mcwilliam and m. rauch ( cambridge : cambridge univ .
press ) , 1 sollima , a. , ferraro , f.r .
, pancino , e. , & bellazzini , m. 2004 , mnras , astro - ph/0411546 sollima , a. , pancino , e. , ferraro , f.r . , & bellazzini , m. 2004 , in chemical abundances and mixing in stars in the milky way and its satellites , ed .
l. pasquini , s. randich ( berlin : springer - verlag ) , in press smith , g. h. , wood , p. r. , faulkner , d. j. , wright , a. e 1990 , apj , 353 , 168 stetson , p.b . , bruntt , h. , & grundahl , f. 2003 , pasp , 115 , 413 suntzeff , n. b. , kraft , r. p. 1996
, aj , 111 , 1913 thompson , i.b . ,
kaluzny , j. , pych , w. , burley , g. et al .
2001 , , 121 , 3089 thoul , a. , bahcall , j. , & loeb , a. 1994 , apj 421 , 828 | we present three different optical and near - infrared ( nir ) data sets for evolved stars in the galactic globular cluster .
the comparison between observations and homogeneous sets of stellar isochrones and zero - age horizontal branches provides two reasonable fits .
both of them suggest that the so - called anomalous branch has a metal - intermediate chemical composition ( @xmath0 \le -0.8 $ ] ) and is located @xmath1 pc beyond the bulk of stars .
these findings are mainly supported by the shape of the subgiant branch in four different color - magnitude diagrams ( cmds ) .
the most plausible fit requires a higher reddening , @xmath2 vs. 0.12 , and suggests that the anomalous branch is coeval , within empirical and theoretical uncertainties , to the bulk of stellar populations .
this result is supported by the identification of a sample of faint horizontal branch stars that might be connected with the anomalous branch .
circumstantial empirical evidence seems to suggest that the stars in this branch form a clump of stars located beyond the cluster . |
the problem of growing trees belongs to larger class of problems of evolving networks a new area with many interdisciplinary applications , from biology and computational science to linguistics @xcite . in statistical mechanics ,
we often investigate the state of thermodynamic equilibrium , which is unique and therefore it can not preserve any information .
however , in other sciences memory on past states is an essential ingredient of the system .
here we are interested in search how the structure of the origin of a tree , i.e. of a graph from which the tree is constructed , influences the overall characteristics of the growing system .
a network containing @xmath0 nodes is fully characterized by its connectivity matrix @xmath1 : @xmath2 if the nodes @xmath3 are linked together , and @xmath4 elsewhere .
more convenient but somewhat redundant is the distance matrix @xmath5 , where the matrix element @xmath6 is the number of links along the shortest path from @xmath7 to @xmath8 .
it is often simpler to describe a network statistically .
a local characteristics of a network includes the degree distribution , i.e. the probability that a node is linked to a given number @xmath9 of neighbors .
a global characteristics includes the node - node distance distribution .
whereas the former can be treated as complete only conditionally @xcite , a few is known on the latter .
recent progress of knowledge on the mean node - node distance @xmath10 $ ] is due to applications of equilibrium statistical mechanics , scaling hypotheses and/or assumptions of lack of correlations between nodes @xcite .
here , @xmath11 denotes an average over @xmath12 non - diagonal matrix elements and @xmath13 $ ] is an average over different matrices , i.e. different graphs . by growing we mean adding subsequent nodes to an already existing graph . when each node is added with one link only @xmath14 , a tree a compact graph without loops and without multiple edges
is formed . in trees ,
a path between each two nodes is unique , and it can not be changed during the growth process . when a node is added , the node - node distance matrix @xmath5 is increased by one column and one row . once the matrix elements are formed , they do not change their values . however , if nodes are added with two or more links @xmath15 , a kind of shortcuts are formed and some node - node distances may be shortened . the main goal of this work is to demonstrate , that the node - node distance distribution of a growing tree preserves an information on the structure of the initial tree , from which it is formed . below we deal with two kinds of growing trees , which differ in the degree distribution .
let us consider the linking of new nodes to randomly selected nodes .
when the selection is made without any preference , we obtain a so - called exponential tree . in this case , the degree distribution @xmath16 , where @xmath9 is the number of links of a node .
nodes can be selected also with some preference with respect to their degree .
if the linking probability is proportional to the degree @xmath9 , we obtain the scale - free or barabsi albert networks @xcite . in this case ,
@xmath17 , with @xmath18 @xcite . to achieve our goal ,
the simplest method is to calculate the mean node - node distance @xmath19 for trees of @xmath0 nodes , the formation of which has started from two different trees with four nodes .
this is done in the next section with iterative equations , which has been derived recently for the exponential trees @xcite . in section [ sec_algorithm ] , the growth algorithms are introduced , basing on an evolution of the distance matrix . in section [ sec_results ] ,
numerical results are presented for the exponential trees and the barabsi albert scale - free trees .
we show also that the memory on the ancestral network is much reduced , if the trees are substituted by graphs with cyclic paths , i.e. with @xmath20 . the last section is devoted to discussion .
consider the probability that a tree of a given structure is grown .
trees are different if there is no one - to - one correspondence between their pairs of linked nodes @xcite .
let us denote the number of different trees with @xmath0 nodes by @xmath21 .
it is easy to check by inspection , that @xmath22 and @xmath23 . as @xmath24 , the probability or weight of the tree of three nodes ( fig . [
fig_trees](a ) ) must be one .
an exponential tree of four nodes can be formed by linking a new ( fourth ) node either to one of two end nodes , or to the central one .
then , the probability of a chain of nodes ( fig . [ fig_trees](b ) ) is @xmath25 , and the probability of a star - like - tree ( fig . [ fig_trees](c ) ) is @xmath26 . from the chain , a longer chain ( fig . [ fig_trees](d ) ) can be produced in two ways , then its weight is @xmath27 . from the star , another star ( fig .
[ fig_trees](f ) ) can appear with the probability @xmath28 .
the remaining tree ( fig . [ fig_trees](e ) ) can be formed from either the chain or the star , then its weight is @xmath29 .
we note that in the case of the scale - free trees , the weights of the trees presented in fig .
[ fig_trees ] are : 1 , 1/2 , 1/2 , 1/6 , 7/12 and 1/4 , respectively .
this is a simple demonstration , that the weights of trees in two different classes are different .
any possible tree can be formed from a tree of three nodes ( fig . [ fig_trees](a ) ) .
the way to form chains and stars is unique and then , their weights are relatively small .
example giving , the weight of an exponential star of @xmath0 nodes is .
we could eliminate stars , if we develop trees from the chain shown in fig .
[ fig_trees](b ) .
seemingly , the weights of other trees should not be changed much , but all of them are influenced by the lack of the stars .
example giving , in this case the tree shown in fig .
[ fig_trees](e ) can be formed in one unique way . as a consequence , the whole distribution of weights is rebuilt . with the iterative equations derived recently @xcite
, we can calculate the mean distance @xmath30 and the mean square of distances @xmath31 $ ] for two `` families '' of trees .
one is formed from the chain - like tree shown in fig . [ fig_trees](b ) and labeled as `` z '' , and another from the star - like tree presented in fig . [ fig_trees](c ) and marked as `` y '' .
then , the first `` family '' does not contain stars , and the second one does not contain chains .
the equations are : [ eq_iter ] @xmath32 and @xmath33 the information on the initial trees is encoded in the initial values of @xmath34 and @xmath35 .
it is easy to check , that for the chain @xmath36 , @xmath37 and for the star @xmath38 , @xmath39 .
similar method has been used in @xcite .
the difference is that here , the eqs .
are exact , but they apply only to the exponential trees .
two initial trees with four nodes ( the chain and the star ) are represented in the computer memory as two distance matrices @xmath40 and @xmath41 . the starting point are two matrices for two trees of four nodes : @xmath42 for the chain and the star , respectively . selecting a node
to link a new node is equivalent to select a number @xmath43 of column / row of the matrix .
then the matrix is supplemented by new column and row , which are copies of the @xmath43-th column / row but with all elements incremented by one @xmath44 and obviously @xmath45 the eq
. served in the derivation of the iterative formulas @xcite .
the same numerical technique is applied also to the case of the barabsi albert scale - free trees .
the only difference is that in this case , the node @xmath43 is selected with preference of the number of its pre - existing links .
namely , @xmath46 where @xmath47 is the number @xmath9 of links of @xmath7-th node .
additional matrix @xmath48 contains the indices of row of the distance matrix @xmath5 where `` 1 '' is encountered .
each case @xmath49 indicates a link between nodes @xmath7 and @xmath8 .
the matrix @xmath48 is useful to select nodes of given degree for the scale - free trees and graphs , according to the so - called kertsz algorithm @xcite .
further , the same technique is applied to simple graphs , where new nodes are attached to previously existing ones by @xmath50 links .
then , cyclic paths are possible and the distance matrix @xmath5 is to be rebuilt when adding each node .
the algorithm is as follows : let us suppose that @xmath51-th node is added to existing nodes @xmath52 and @xmath53 .
then [ eq_graphs ] @xmath54 for new , @xmath51-th , column / row @xmath55 and again for the diagonal element @xmath56 in the case of growing graphs @xmath57 .
the gray sites show randomly chosen columns / rows ( nodes to which new node will be attached ) .
the black sites show matrix elements which are reevaluated from eq . due to newly created shortcuts .
the last columns / rows are constructed according eqs . and . starting with the y - like star new nodes were subsequently added to nodes @xmath58 . ]
one step of construction of the matrix @xmath5 for simple graphs @xmath57 is presented in fig .
[ fig_matrix ] .
an example of the construction @xmath5 for trees @xmath14 is given in @xcite .
in figs . [ fig_met_exp ] and [ fig_met_ab ] the dependences ( a ) @xmath59 and ( b ) @xmath60 obtained from growth simulations are presented , for exponential trees and for scale - free trees , respectively .
the results of simulations are averaged over @xmath61 independent growths . for both kinds of trees
the difference in average node - node distance @xmath62 tends to the constant value during the growth process . for higher moments
the effect is even stronger and @xmath63 increases with the tree size @xmath0 . in fig .
[ fig_met_exp ] we give also the results for @xmath64 and @xmath65 calculated with eq . for exponential trees of @xmath66 nodes .
the fact that @xmath62 and @xmath63 do not decrease with @xmath0 means that growing structures preserve the memory on their initial shapes . in the case of simple graphs ( @xmath50 ) , the distance matrix @xmath5 must be reevaluated , what makes the time of the calculation substantially larger .
the results for graphs are averaged only over one thousand of independent growths .
the curves @xmath19 and @xmath67 for both kind of simple graphs are shown in fig .
[ fig_dn ] .
the linear fits for @xmath68 are @xmath69 and @xmath70 for the exponential graphs and the scale - free graphs , respectively . the functions @xmath64 and @xmath65 for both kind of evolving graphs are shown in fig .
[ fig_meg ] .
for the scale - free graphs , we observe some small memory effect , which manifests as a constant mutual shift of the plots @xmath67 vs. @xmath71 . in this case
it is not clear if the effect vanishes or not , when @xmath0 tends to infinity . and ( b ) @xmath65 for _ exponential trees _ obtained with iterative formula as well as from the direct growth simulations .
the results of simulations are averaged over @xmath61 independent growths.,title="fig:",scaledwidth=45.0% ] and ( b ) @xmath65 for _ exponential trees _ obtained with iterative formula as well as from the direct growth simulations .
the results of simulations are averaged over @xmath61 independent growths.,title="fig:",scaledwidth=45.0% ] and
( b ) @xmath65 for _ scale - free trees
_ obtained from the growth simulations .
the results are averaged over @xmath61 independent growths.,title="fig:",scaledwidth=45.0% ] and ( b ) @xmath65 for _ scale - free trees _ obtained from the growth simulations .
the results are averaged over @xmath61 independent growths.,title="fig:",scaledwidth=45.0% ] and ( b ) @xmath67 for the exponential and scale - free _ graphs _ and different initial configurations obtained from the growth simulations .
the results are averaged over @xmath72 independent growths .
the dependence on the initial configuration is not visible in the scale of the plot.,title="fig:",scaledwidth=45.0% ] and ( b ) @xmath67 for the exponential and scale - free _ graphs _ and different initial configurations obtained from the growth simulations .
the results are averaged over @xmath72 independent growths .
the dependence on the initial configuration is not visible in the scale of the plot.,title="fig:",scaledwidth=45.0% ] and @xmath65 for the exponential and the scale - free _ graphs _ obtained from the growth simulations .
the results are averaged over @xmath72 independent growths.,scaledwidth=45.0% ]
in the case of the exponential trees , the results of the simulations agree well with the curves obtained from the iteration equations .
this fact supports the reliability of the numerical equation for the scale - free trees and the graphs with @xmath50 , where we have no analytical calculations .
main result of this work is , that the node - node distance distribution in a growing tree depends on its initial structure .
our calculations indicate , that both the average distance @xmath30 and its second moment @xmath73 in trees display this kind of memory .
the information is encoded in the constant @xmath74 in the expression @xmath75 .
the constant @xmath74 varies by about 0.109 and 0.164 , when we change the shape of the initial tree of four nodes from the y - like star to the z - like chain for the exponential and scale - free trees , respectively . in the second moment
@xmath76 , it is the constant @xmath77 which depends on the initial shape .
this is true both for the exponential and the scale - free trees .
the memory effect is much reduced or even disappears in the case when new nodes are linked to the network by at least two edges . in this case , the distance matrix @xmath5 is rebuilt by new edges which can shorten distances between initially far nodes by providing new paths between them .
concluding , we have demonstrated that the growing trees carry an information on their initial geometrical structure .
the validity of this result relies not only on pure geometry , but also on a particular application of the graph theory .
as remarked in the introduction , the list of examples of such applications is quite rich @xcite .
if the considered network is due to the citation index , we trace the flow of a new idea from one paper to another .
we see that around some seminal papers , networks of citations are formed , as it happens in the case of ref .
sometimes there are two or more seminal papers , and then the shape of the network depends on their clarity , ease of mathematical formulation and individual preferences of the readership , formed in personal contacts .
if new results spread just by reading papers , it spreads slowly : somebody reads it , tells to a friend , the friend s student is asked to calculate a similar problem .
the tree is ` chain - like ' . on the contrary
, each conference makes the tree of spreading of ideas to be more ` star - like ' , where the possible sources of getting new information are multiple and efficient .
the authors are grateful to prof .
kazimierz raski for valuable help and to mr .
pawe kuakowski for comments on manuscript .
the numerical calculations were carried out in ack cyfronet
the machine time on sgi 2800 is financed by the polish state committee for scientific research ( kbn ) under grant no .
kbn / sgi2800/agh/018/2003 .
99 r.albert and a .- l.barabsi , rev .
* 286 * ( 2002 ) 47 .
s.n.dorogovtsev and j.f.f.mendes , adv .
* 51 * ( 2002 ) 1079 .
m.e.j.newman , siam review * 45 * ( 2003 ) 167 .
s.n.dorogovtsev and j.f.f.mendes , in _ from the genome to the internet _ , eds .
s.bornholdt and h.g.schuster , viley - vch , berlin , 2002 .
z.burda , j.d.correia , and a.krzywicki , phys . rev .
* e56 * ( 2001 ) 46118 .
s.n.dorogovtsev , j.f.f.mendes , and a.n.samukhin , nucl . phys .
* b653 * ( 2003 ) 307 .
a.fronczak , p.fronczak , and j.hoyst , cond - mat/0212230 .
p.biaas , z.burda , j.jurkiewicz , and a.krzywicki , phys . rev . *
e67 * ( 2003 ) 066106 .
l.barabsi and r.albert , science * 286 * ( 1999 ) 509 .
k.malarz , j.czaplicki , b.kawecka-magiera , and k.kuakowski , int . j. mod . phys .
* c14 * ( 2003 ) 1201 .
k.malarz , j.karpiska , a.kardas , and k.kuakowski , task quarterly * 8 * ( 2004 ) 115 .
r.j.wilson , _ introduction to graph theory _ , longman scientific and technical , new york , 1987 .
r.v.kulkarni , e.almaas , and d.stroud , phys . rev .
* e61 * ( 2000 ) 4268 .
g.szab , m.alava , and j.kertsz , phys . rev .
* e66 * ( 2002 ) 026101 .
d.stauffer private communication . | we show that the structure of a growing tree preserves an information on the shape of an initial graph . for the exponential trees , evidence of this kind of memory
is provided by means of the iterative equations , derived for the moments of the node - node distance distribution .
numerical calculations confirm the result and allow to extend the conclusion to the barabsi albert scale - free trees .
the memory effect almost disappears , if subsequent nodes are connected to the network with more than one link .
evolving networks , graphs and trees , small - world effect 82.20.m , 05.50.+q |
clusters of galaxies are potentially powerful tools for precision cosmology . in @xmath0cdm and other models of hierarchical structure formation ,
the mass function of clusters and its evolution are very sensitive to cosmological parameters @xcite .
recently , many authors have used the observed cluster abundance in a variety of wavelength ranges to constrain both the present day mean matter density @xmath1 and the amplitude of linear fluctuations on 8 @xmath2 mpc scales ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the future , the first large - scale x
ray surveys will yield catalogs of many thousands of clusters @xcite that will place very tight constraints on cosmological parameters , including the dark energy equation of state and its evolution @xcite . with such a potentially promising dataset on the horizon , particular attention has been paid recently to addressing potential systematic uncertainties that may bias measurements of cosmology . as @xcite observed , the systematic uncertainty in determining @xmath3 using current cluster measurements is driven by the calibration of the mass - temperature relation . if cosmological results from future cluster surveys are to be robust , accurate theoretical models of cluster formation and evolution , including identification and proper treatment of the dominant heating and cooling mechanisms of the intracluster gas , are required .
the observational properties of the intracluster medium ( icm ) are driven primarily by simple gravitational collapse . if this were the only important physical mechanism , clusters would scale self similarly @xcite .
however , it has been known for some time that the observed bulk properties of galaxy clusters temperature , mass , luminosity do not conform to the self
similar model @xcite .
this effect is also seen in the luminosity temperature scaling of groups @xcite and the faintness of the unresolved @xmath4 kev x ray background @xcite .
these departures from self similarity are affected by non gravitational baryonic physics ; a complex interplay between various thermodynamic processes involving star formation and galaxy evolution ( see * ? ? ?
* ; * ? ? ?
* and references therein ) .
motivated in part by observations suggesting a universal entropy floor for clusters @xcite , many studies have proposed `` preheating '' of the gas in order to explain this departure from self similarity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in this scenario , the effects of energy input into the icm from non gravitational processes such as supernovae , star formation , and galactic winds are approximated by a high redshift entropy modification that particularly affects low mass systems @xcite .
subsequent numerical @xcite and analytic @xcite modeling has been successful in reproducing many of the observed cluster scaling relations , both in the x ray and sunyaev
zeldovich ( sz ) effect @xcite .
in particular , @xcite propose that a simple shift of the entropy profile may be a good approximation to the effects of preheating .
however , since this results was obtained in the smooth accretion limit , and real accretion of preheated gas is likely to be somewhat lumpy , it may not hold generally .
this is because it is not clear _ a priori _ how efficiently universal preheating will smooth the gas distribution prior to cluster collapse . in this work
we compare the predictions of this shift model for preheating to the results of high resolution cosmological simulations .
such a comparison tests this analytic prescription , given a realistic accretion history . and
, it will motivate other model assumptions , such as the choice of boundary conditions , that are more difficult to derive from first principles .
at the same time , recent observations of the entropy profiles of nearby clusters have suggested that preheating scenarios which invoke a universal entropy floor are incorrect in detail @xcite .
the same simulations we use to test analytic prescriptions offer the added benefit of a fully informed comparison of the predictions of the preheating scenario to observations . this work is divided into three sections . in [ sec : sim ] , we summarize the simulation and cluster identification procedure . in [ sec : model ] , we compare the corresponding model predictions to the results of our simulations . finally , in
[ sec : obs ] we compare the results of our simulations to recent observations of nearby clusters .
throughout this work , `` preheating '' should be taken to refer specifically to the shift model of @xcite .
the simulations were performed with the adaptive mesh refinement code enzo @xcite assuming a flat @xmath0cdm cosmology .
the cosmological model is consistent with the wmap year 3 results @xcite with @xmath5 .
we ran two complementary sets of simulations .
the first simulation was performed in a comoving periodic box with @xmath6 @xmath2 mpc on a side with @xmath7 dark matter particles and an equal number of grids , allowing for six levels of additional refinement that yielded a minimum cell size of @xmath8 @xmath2 kpc .
this level of resolution was found ( via a smaller set of higher resolution simulations ) to be sufficient for reproducing the core entropy level of the smallest clusters we will investigate in this paper .
the initial conditions were generated with the fitting form of the linear dark matter power spectrum given by @xcite at an initial redshift of @xmath9 ( for details regarding the initialization of such simulations see * ? ? ? * and references therein ) .
due to its relatively small size , this 100 @xmath2 mpc simulation produced only a limited number of high mass clusters . in order to fill out the higher mass end of the scaling relations
, we performed a second set of simulations using the same cosmology and method of generating initial conditions .
first , a dark matter only run was performed in a comoving periodic box @xmath10 @xmath2 mpc on a side with @xmath11 dark matter particles .
then , the largest dark
matter halos were identified using the hop algorithm @xcite , and the simulation was re
run with two nested grids : the first in the original @xmath10 @xmath2 mpc box with @xmath11 dark matter particles and an equal number of grids , and the second in a high resolution region , centered on one of the haloes , with @xmath12 @xmath2 mpc and the same number of particles and grids , using the same refinement technique as the larger simulations .
this gave our high resolution region equivalent mass and spatial resolution to the original runs .
we repeated this process for each of the four largest dark
matter haloes .
@xcite have suggested that simple preheating in the smooth accretion limit is well approximated by a universal shift of the unmodified cluster entropy profile .
this corresponds , in the context of a full cosmological simulation , to raising the entropy level of all the gas in the universe at high redshift , well in advance of cluster collapse , but within the time frame during which we expect to find galactic and stellar entropy input to the icm
. therefore , a fixed amount of entropy @xmath13 was added at @xmath14 by increasing the thermal energy of each grid point by @xmath15 where @xmath16 is the baryon density at that grid
we assume an ideal gas consisting of a fully ionized h he plasma with the cosmic helium mass fraction @xmath17 and @xmath18 .
this was done for four cases : no preheating ( @xmath19 ) and @xmath20 , 155 , and 311 .
the effects of radiative cooling and other additional sources of non gravitational heating were neglected .
snapshots of the @xmath21 emission weighted 2d projection of the gas temperature in the same cluster from two of the simulations are shown in figure [ fig : snapshot ] .
this figure demonstrates visually that adding entropy smoothes the small - scale distribution , but leaves the larger scale structures largely intact . at each redshift output
, clusters were identified using the hop algorithm of @xcite , yielding clusters at @xmath22 up to @xmath23 in the 100 @xmath2 mpc simulations , and clusters up to @xmath24 in the large box simulations , where @xmath25 is the total mass enclosed within a spherical overdensity of 200 relative to the critical density .
these were then analyzed using a @xcite cooling function assuming a fully ionized h he plasma with abundance @xmath26 .
the cluster center for large clusters ( @xmath27 ) was calculated using an iterative process of re centering on the center of mass using successively smaller radii .
for all others , it was taken as the densest dark
matter particle from the hop output .
we found that the choice of centering technique had no more than a 15% effect on the core properties of the cluster , even at the highest masses . for each cluster , we determined the mass ( @xmath28 ) and emission ( @xmath29 ) weighted gas temperatures , and the bolometric x ray luminosity ( @xmath30 ) .
in addition , we consider observables associated with the sunyaev
zeldovich effect ( sz ; * ? ? ?
* ; * ? ? ?
* ) , including the central decrement ( @xmath31 ) and frequency independent , angular diameter distance scaled ( fiads ) integrated sz luminosity out to @xmath32 ( @xmath33 ) .
these were calculated according to @xmath34 where @xmath35 is the thompson cross section , @xmath36 is the mass of the electron , @xmath37 , is the speed of light , @xmath38 is the electron pressure , and @xmath39 is the line element along the line of sight through the cluster center , and @xmath40 where @xmath41 and @xmath42 are the boltzmann and planck constants respectively , @xmath43 @xcite is the cosmic microwave background ( cmb ) temperature , and the integral is performed over a spherical volume out to @xmath32 .
finally , we calculated the gas weighted entropy profile , where entropy was defined as @xmath44 where @xmath45 is the average number density of electrons in a given shell , again assuming a fully ionized plasma .
our analytic model follows the techniques of @xcite and @xcite . given an initial , unmodified , spherically symmetric model in hydrostatic equilibrium , then entropy profile is `` modified '' in this case with a simple shift at all radii and the equations of hydrostatic equilibrium are re integrated , yielding a final `` modified '' gas and temperature distribution . therefore ,
if @xmath46 represents the shifted entropy profile , both the unmodified and modified models satisfy @xmath47 where , casting the pressure and temperature in terms of @xmath48 and @xmath46 @xmath49^{3/5}.\ ] ] and assuming an ideal gas , @xmath50 keeping in mind @xmath51 for the unmodified case .
furthermore , motivated in part by the simulations of @xcite and @xcite , we include @xmath52 in ( [ eq : hydro ] ) to allow for departures from strict hydrostatic equilibrium , in which the gas is supported by a combination of gas pressure and turbulent motions .
the gravitational potential of the dark matter ( dm ) follows a @xcite ( nfw ) profile , as motivated by n - body simulations , with a fixed concentration parameter @xmath53 @xcite .
finally , both the modified and unmodified models satisfy the same outer boundary condition , which is chosen to match the simulation results , but is related to the pressure required to resist the infall of baryonic matter at the viriol radius @xmath54 , where @xmath55 is the cosmic baryon baryon fraction .
the entropy floor is approximated by a shift in the entropy profile defined by : @xmath56 as discussed in @xcite , this shift is a simple phenomenological approximation of a uniform , highredshift preheating .
in particular it assumes that mass shells remain in the same order after the modification ( this is reasonable due to that fact that entropy must be monotonically increasing for stability reasons ) , and that the effect of preheating is to increase the entropy by the fixed , stated amount even after passing through the shock . @xcite examined this issue and found that this simple shift was a reasonable approximation ( to within 20% ) for smooth accretion .
of course , in a full cosmological simulation , the accretion is not generally smooth ; however , as we will show , the shift approximation works remarkably well . for our comparison to be self
consistent , the most important consideration was our choice of unmodified distribution
. this must match the results of the no preheating simulation as closely as possible ; for if our unmodified model does not match the simulations without preheating , we have no reason to expect that modifications to that model will be consistent either .
we therefore attempted the match the median entropy profile in two mass bins : low ( @xmath57 ) and moderate ( @xmath58 ) mass clusters .
all of the model predictions were made at the median mass in each bin ; @xmath59 and @xmath60 in the low and moderate mass bins respectively . a summary of all the models considered is shown in figure [ fig : unmodified ] .
we first consider a set of _ ab initio _ unmodified models . our first choice was similar to the fiducial model of @xcite and @xcite , in which the unmodified gas distribution is taken to follow the nfw dark matter density profile at the cosmic baryon density in strict hydrostatic equillibrium ( @xmath61 ) .
although it roughly reproduces the profile , in detail this nfw gas entropy profile was both too steep and incorrectly normalized , predicting a higher entropy for @xmath62 than was seen in the simulations .
it furthermore did not reproduce the core entropy seen in simulations , due to the divergent nfw density at small radii . allowing @xmath63 to vary produced somewhat better agreement in the normalization and slope , but still predicted identically zero core entroy .
we then tried a set of alternative models in an attempt to reproduce the core entropy see in the simulations : isothermal and a gas distribution with a flat core .
isothermal gas in hydrostatic equilibrium with @xmath64 ( see * ? ? ?
* ) also had incorrect slope and normalization , with a core entropy that was too low .
we then tried keeping the gas in hydrostatic equilibrium , but modifying the gas distribution such that @xmath65 .
this too , did not fit the simulations , requiring a divergent temperature at the cluster center and a non monotonically increasing entropy profile , both of which were unphysical . since reproducing this median entropy profile from the no preheating simulations with
models yielded no success , we instead chose to use a fit to the median entropy profile in the amr simulations of @xcite .
we parameterized the entropy profile in terms of @xmath66 , the enclosed gas fraction , with the following fit : @xmath67 where @xmath68 and @xmath69 , following the definitions in @xcite .
we chose the outer boundary condition , once again based on the simulations , to be given by @xmath70 .
@xcite showed that the entropy profiles in sph and amr simulations agree very well outside of @xmath71 while inside this radius , the sph simulations show steeper entropy profiles with lower ( but non - zero ) core entropies .
the source of this discrepancy is not clear , although the new entropy conserving version of gadget2 @xcite does show higher entropy cores ( g. yepes , private communication ) . in any case , the size of the difference at @xmath72 is quite small , about 25% and certainly not enough to resolve the discrepancy with observations discussed in section [ sec : obs ] . finally , we note that our simulated clusters were not in strict hydrostatic equilibrium ( see figure [ fig : hydro ] ) , a result which is consistent with previous work @xcite .
the result of this was that even with our fitted entropy profiles , we could not reproduce the simulated density and temperature profiles in detail for @xmath73 .
we found that setting @xmath63 to 0.8 , as suggested by the median profiles in figure [ fig : hydro ] ( and previous work ) generated a good fit to the gas profiles over a wide range of masses .
having found a reasonable entropy profile within our unmodified model , we then modied these profiles according to the prescription in @xcite and @xcite and computed the resulting model predictions for gas density , temperature , and entropy profiles in the same mass bins using matching entropy shifts of @xmath19 ( no preheating ) , 78 , 155 , and 200 . in figure
[ fig : profiles.model ] , we compare the model predictions to the simulation results for the gas density @xmath16 , gas weighted temperature @xmath28 , and gas weighted entropy @xmath74 profiles .
this comparison is made for the median profile in both the low and moderate mass bins .
we find remarkable agreement between our analytic predictions and the simulation results over both mass bins for all three entropy shifts .
the temperature profiles are underpredicted by the model , but never by more than @xmath75 at @xmath76 , and never by more than @xmath77 out to @xmath32 .
the entropy profile in the center is very well reproduced .
there is also a tendency for the model to underpredict the entropy at @xmath78 compared to the simulations , although again never by more than about @xmath79 .
we can therefore conclude , as one can see qualitatively in figure [ fig : snapshot ] that preheating efficiently smoothes the accreted density distribution , making the accretion histories of our simulated clusters more smooth . as a result , shifting the unmodified entropy profile is a good approximation to the effects of preheating in the simulations .
furthermore , we find that the model predictions and simulation results converge as @xmath13 increases .
this occurs because as @xmath13 increases , the entropy input can effectively smooth larger and larger sub haloes , and the simulations converge on the smooth accretion limit , making our modified models a better and better approximation . from the comparison to the median profiles in figure [ fig : profiles.model ] , we find both qualitative and quantitative agreement between the analytic predictions and the simulation results .
this is very encouraging , however comparing the median profiles is only a snapshot of the behavior of the intracluster entropy for the median cluster profile in two mass bins .
the model predictions for the intracluster entropy over a much broader mass range contain a great deal of addition information . in figure
[ fig : entropy.model ] , we show the entropy
mass scalings for two characteristic entropy measures used in observations : the `` core '' entropy , taken at @xmath80 ( lower panel , @xmath81 ) , and the `` outer '' entropy taken at @xmath82 ( upper panel , @xmath83 ) . for the core entropy ,
our analytic predictions are again in remarkable agreement with the simulations .
however at high mass @xmath84 , our analytic predictions are somewhat lower than the simulation results .
this behavior is slightly enhanced for higher values of @xmath13 .
conversely , the models somewhat underpredict the outer entropy at all masses , with the exception of the very highest mass clusters where small numbers of objects make the trend difficult to discern .
the agreement we see for the entropy of the preheated models is of particular interest and somewhat surprising .
the effects of preheating on the entropy of simulated clusters in the cosmological simulation represents contributions from two competing effects .
entropy input into the icm will tend to push out the accretion shock , which will in turn decrease the efficiency of heating at the shock and decrease the core entropy @xcite .
however , as first suggested by @xcite , preheating will also tend to lower the density of accreted haloes .
this may affect a partial transition from lumpy to smooth accretion that will actually increase the efficiency of entropy generation at the accretion shock , and with it the entropy of the system .
our simulations suggest that these competing effects tend to cancel each other out . in order to extract only the effects of entropy input on the simulations
, we match the halos in each simulation , and do a halo by halo comparison . for each cluster , we examine two ratios , or amplification factors : @xmath85 and @xmath86 , where @xmath87 is the gas weighted average entropy @xmath88 at fixed gas fraction @xmath89 , @xmath90 is the entropy at fixed gas fraction @xmath91 , and @xmath92 is the total enclosed gas mass at @xmath32 .
we use @xmath93 to indicate the no preheating case , and compute the ratio for each of the preheated simulations .
these ratios represent the entropy offset between simulated clusters at fixed gas mass , in units of the entropy input @xmath13 , at fixed gas mass in both the core and outskirts .
the results of this comparison are show in figure [ fig : entropyamp ] , with @xmath94 or @xmath95 representing no entropy amplification , and larger values of @xmath96 implying more entropy amplification .
we find only mild evidence of entropy amplification in figure [ fig : entropyamp ] .
the overall mean core and outer amplification factors for @xmath97 are @xmath98 and @xmath99 with associated variances @xmath100 and @xmath101 .
the trend towards more amplification at lower masses for both amplification factors is qualitatively consistent with @xcite and @xcite .
however , the mean amplitude of the amplification is @xmath102 for all values of @xmath13 , and within a wide scatter is consistent with unity .
overall we find that on average , preheating does make entropy generation more efficient , but only by a small fraction .
@xcite found amplification factors of @xmath103 and 1.84 at @xmath104 for a simulated group of mass @xmath105 using entropy floors of @xmath106 and 100 respectively .
as noted above , we find smaller values for a much larger sample .
one possible source of this discrepancy is the nature of the simulations : @xcite include the effects of heating / cooling from a uniform uv background , star formation from a multiphase interstellar medium , and galactic winds powered by supernova feedback , whereas our simulations use a simple implementation of preheating in which all of these processes are subsumed into the high redshift entropy modification .
another possibility is the improved statistical power of our simulations , which model more objects than @xcite .
although there are groups of comparable mass in our sample with amplification factors that are roughly consistent with the findings of @xcite , we also find that the mean amplification is within @xmath107 of unity , even for low mass systems .
however , perhaps the most likely explanation is the timing of the entropy input . in @xcite
, the entropy was added much later , at @xmath108 , when many smaller halos would have already formed .
the gas in these halos is at high density and so , for a fixed amount of entropy , receives much more energy .
this may have lead to large - scale outflows and shocking , which , in turn , may have caused the entropy amplification . in the simulations discussed here , the input is done at much earlier times ( @xmath109 ) as in @xcite , because we want to isolate only the effects of preheating .
we also find some evidence of turbulent mixing in the gas .
this will will tend to equalize the entropy profile , raising the core and lowering the outer entropy .
it has previously been noted that simulated clusters are not in strict hydrostatic equilibrium @xcite , probably due to turbulence @xcite .
figure [ fig : entropyamp ] exhibits two telling trends : the core amplification factor @xmath94 tends towards unity with higher entropy input @xmath13 , and the outer amplification factor @xmath110 tends toward entropy suppression ( @xmath111 ) in systems with @xmath112 .
this is consistent with turbulent mixing , which will be less efficient at amplifying the core entropy at higher @xmath13 as smoothing suppresses the turbulence , and more efficient at suppressing the outer entropy at higher mass as more of the substructure is resolved and turbulence becomes more important .
these two tends are , as we show in figure [ fig : hydro ] , borne out in our simulations ; higher mass clusters with higher @xmath13 tend to have more support against gravity contributed by turbulent motions in the gas . however , though we see trends that suggest turbulent mixing , we concede that this is speculation and warrants a more in depth analysis , which we defer to future work . therefore , our simulations show the combined effect of two processes that are difficult to disentangle .
classic entropy amplification will tend to increase both the core and outer entropy , and will tend towards higher @xmath94 at lower mass .
turbulent mixing will suppress the outer entropy , tending towards lower @xmath110 at higher mass , while simultaneously raising the core entropy more efficiently at higher @xmath13 .
we see all of these trends in our data . as a result , while the magnitude of the amplification favors the more subtle effects of turbulent mixing , we can not firmly argue that either is dominant .
nevertheless , as figure [ fig : entropyamp ] makes clear , the size of either effect is not large .
in addition to the entropy scalings , we compare the model predictions to the simulation results for several observationally motivitated scaling relations . in figure [ fig : scaling.model ] , we present two scalings typically used by x ray observers : the mass temperature ( for @xmath25 ) , and temperature luminosity ( for @xmath30 , the bolometric x ray luminosity ) relations . in figures [ fig : yc.model ] and [ fig : yint.model ] respectively , we present the mass and temperature scalings for two sz observables defined above : the central decrement @xmath31 and the fiads sz luminosity @xmath113 . for the typical x
ray observables show in figure [ fig : scaling.model ] , we again find remarkable agreement between our model and the simulations .
this is due largely to the agreement between the predicted and simulated core density , as the x ray emission scales as @xmath114 and is only weakly dependent on the temperature of the gas .
furthermore , this agreement suggests that the scalings in figure [ fig : scaling.model ] are largely insensitive to asymmetries in the simulations ; we find that our assumption of spherical symmetry is a good approximation to the simulation results . in figures [ fig : yc.model ] and
[ fig : yint.model ] , we present the mass ( left panel ) and temperature ( right panel ) scaling of two typical sz observables .
our analytic predictions are in good agreement with the simulation results for both scalings .
we furthermore find that the central sz decrement , in both the model and simulations , to be a far better indicator of cluster temperature than mass .
we also confirm previous studies , which found the fiads sz integrated luminosity to be a robust , low
scatter mass and temperature indicator with roughly self
similar scaling @xcite .
this mass fiads sz luminosity scaling is remarkably independent of the value of the entropy input @xmath13 , while the temperature scaling exhibits weak @xmath13 dependence .
we have found very good agreement between our analytic model predictions and the simulation results for the median density , temperature , and entropy profiles , the core and outer entropy scalings , and six different observable scalings .
it is important to note , however , that this agreement came in large part by imposing an unmodified entropy profile derived from the simulations .
none of our _ ab initio _ unmodified models were in agreement with the simulation results .
in particular , even when the slope and normalization of the entropy profile were matched closely , as in the case of an nfw gas profile with @xmath115 ( see figure [ fig : unmodified ] ) , there was poor agreement with the simulation results .
this was true even of the scaling of integrated quantities such as the @xmath116 and @xmath117 relations , which were too steep and incorrectly normalized .
at the same time , as the value of the entropy input @xmath13 was increased , we found improved agreement with the simulations .
this convergence is consistent with what we showed explicitly earlier ; preheating efficiently smoothes the icm , and therefore a simple shift of the unmodified entropy profile is a good approximation for the effects of preheating @xcite . at the same time , an unmodified model with no core entropy was quantitatively and qualitatively at odds with simulation results , even when that unmodified entropy profile was a very good fit for @xmath118 . and , though the value of this core entropy can vary depending on the numerical technique used @xcite , it is essential in setting the thermodynamic state and structural properties of the icm .
the importance of this non zero core entropy has been previously downplayed @xcite , but here we find it to be a primary cause of disagreement between analytic model predictions and hydrodynamical simulations , even when the modification procedure is shown to be a good approximation .
there has been some suggestion in the literature , motivated by recent observations , that preheating models which invoke a universal entropy floor are incorrect in detail @xcite . in particular , some authors have suggested that this brand of preheating predicts isentropic cores in low
mass systems that are not observed . in this section
, we use the results of our simulations , which necessarily include all the relevant dynamical and geometric effects , as a test of the preheating scenario itself . towards this end , we compare our results to observations of x ray scalings ( figure [ fig : scaling.obs ] ) and both the core and outer entropy temperature scalings ( figures [ fig : outer.entropy.obs ] , [ fig : outer.entropy.obs2 ] , and [ fig : core.entropy.obs ] ) .
j in figure [ fig : scaling.obs ] , we show the simulation results , along with best fit power laws , for the mass temperature ( upper panel ) and temperature luminosity ( lower panel ) scaling relations as compared to the observations of @xcite . as we mention above , the mass temperature relation from the simulation is roughly independent of the value of the entropy input @xmath13 . and ,
not surprisingly , with the exception of the no preheating simulation all of the simulated mass temperature relations are consistent with the observational data . the temperature luminosity relation , on the other hand , is very sensitive to the entropy input , as the bolometric luminosity scales roughly as @xmath119 and preheating depresses and flattens the core density distribution .
the temperature and luminosity are also both pure observables given a cosmological model as opposed to @xmath25 which @xcite estimate using an isothermal @xmath120model @xcite and therefore provide more robust constraints on @xmath13 .
we find that an entropy input @xmath121 will best fit the observational data .
figures [ fig : outer.entropy.obs ] , [ fig : outer.entropy.obs2 ] , and [ fig : core.entropy.obs ] show the outer ( @xmath83 and @xmath122 ) and core ( @xmath123 ) entropy temperature scaling from the simulations as compared to the observations of @xcite and @xcite .
our simulation results show the nearly isentropic cores in low
mass ( or temperature ) clusters : notice that the entropy is nearly mass independent for low - mass clusters at the highest level of entropy input .
in addition there is a break in the power - law scaling which shifts to higher mass with higher entropy input .
we find that , while observations of the entropy at large radius in rich clusters are consistent our results , fully three dimensional , non radiative cosmological simulations of the preheating scenario of cluster formation are inconsistent with observations in several ways .
first , the entropy input scaling suggested by the temperature luminosity relation is clearly inconsistent with observations of the intracluster entropy ; simulations of the best fit @xmath13 range overproduce the core entropy and underproduce the outer entropy at fixed temperature .
second , the entropy temperature scalings are themselves inconsistent as compared to the observations ; the best
fit entropy input value as determined by the core entropy underproduces the outer entropy , and vice versa .
finally , we show the insentropic low
mass cores predicted by preheating as a function of @xmath13 , which are not observed . as a caveat to our comparison to observations ,
we note that there has been some suggestion , most notably by @xcite that observational studies of the entropy profiles of clusters have been biased towards low core entropy systems .
this may , they argue , explain why some observations appear to be inconsistent with high levels of entropy input @xcite .
however , while this is a possibility , we find dramatic and qualitative disagreement between the scalings in our simulations simulations and observations which would be difficult to reproduce with a low core entropy selection bias .
we furthermore find disagreement at large radius ( see figures [ fig : outer.entropy.obs ] and [ fig : outer.entropy.obs2 ] ) , well outside the cool region .
it is unlikely that a bias related to low entropy in the core would affect the outskirts . as an aside
, we briefly investigate the effects of radiative cooling .
there has been some hope in the literature that the shortcomings of the preheating scenario will be solved by a combination of preheating and radiative cooling @xcite .
unfortunately this is difficult to implement in a simulation because heating and cooling are probably strongly tied together ( e.g. , * ? ? ?
* ; * ? ? ?
nevertheless , we can imagine one simple and concrete extension to the straightforward preheating picture in which we include radiative cooling ( but not star formation or feedback ) at some late epoch , well after preheating has occurred .
we therefore re ran our @xmath124 simulation , including the effects of radiative cooling for @xmath125 ( see * ? ? ?
we find that although the core entropy is modified , the entropy at @xmath126 is relatively insensitive to cooling and so the basic results above do not change .
this is , however , another topic worthy of much more thorough investigation .
we present high resolution hydrodynamic simulations of the preheating scenario of cluster formation , in which the effect of sn and/or agn heating is approximated as a universal increase of the entropy , implemented at high redshift .
we focus on two aspects : ( 1 ) how well analytic models of entropy input reproduce the simulations results , and ( 2 ) how well the preheating simulations match current cluster observations .
we find that analytic models of preheating following those of @xcite and @xcite are remarkably successful at reproducing the results of hydrodynamical simulations in detail .
this agreement extends from scalings of integrated quantities such @xmath116 , @xmath117 , and sz scalings for @xmath31 and the fiads integrated sz luminosity , to core and outer entropy scalings , and finally to the median entropy , temperature , and density profiles of both low and moderate mass clusters .
however , this agreement is dependent on the right choice of an unmodified profile ; in particular one with non
zero core entropy .
when we look in detail at the simulation results , we do not find strong evidence for entropy amplification , the process in which the entropy is amplified even beyond the level injected @xcite .
this differs from the results of @xcite , who looked at a small set of clusters and groups .
the reasons for this are not entirely clear although their preheating model differed from that used here ( in particular the entropy input occured at a lower redshift ) .
our simulations also show that the preheating scenario for cluster formation is in conflict with recent observations of the x ray scalings @xcite and intracluster entropy @xcite . the best
fit value of the entropy input @xmath13 , determined from the x ray scalings ( see figure [ fig : scaling.obs ] ) , underproduces the entropy at @xmath126 and overproduces the entropy at @xmath127 at fixed temperature ( see figures [ fig : outer.entropy.obs ] and [ fig : core.entropy.obs ] ) .
no single value for the entropy input due to preheating can simultaneously match the observed entropy values at both the core _ and _ the outskirts .
this result is in qualitative agreement with previous work @xcite .
the inclusion of radiative cooling at late times ( without any additional feedback ) does not appear to change this result .
therefore , we find this simple preheating scenario to be in disagreement with observations , implying that other , more sophisticated treatments are needed .
we thank zoltn haiman , stefano borgani , mark voit and sheng wang for helpful comments .
in addition , we thank the referee for helpful suggestions . this work was supported in part by nsf grant ast-05 - 07161 , by the u.s .
department of energy under contract no .
de ac0298ch10886 , and by the columbia university initiatives in science and engineering ( ise ) funds .
greg bryan acknowledges support from nsf grants ast-0547823 , and ast-0606959 .
arnaud , m. & evrard , a. e. , 1999 , mnras , 305 , 631 ascasibar , y. , yepes , g. , mueller , v. , & gottloeber , , s. , 2003 , mnras , 346 , 731 babul , a. , balogh , m. l. , lewis , g. f. , & poole , g. b. , 2002 , mnras , 330 , 329 balogh , m. l. , babul , a. , & patton , d. r. , 1999 , mnras , 307 , 463 bahcall , n. a. & bode , p. , 2003
, apj , 588 , l1 bahcall , n. a. & cen , r. , 1992 , apj , 398 , l81 bahcall , n. , et al . , 2003 , apj , 585 , 182 balogh , m. l. , babul , a. , voit , m. g. , mccarthy , i. g. , jones , l. r. , lewis , g. f. , & ebeling , h. , 2006 , mnras , 366 , 624 bertschinger , e. , 1998 , ara&a , 36 , 599 bialek , j. j. , evrard , a. e. , & mohr , j. j. , 2001 , apj , 555 , 597 bode , p. , ostriker , j. p. , weller , j. , & shaw , l. , 2006 , apj , in press , astro - ph/0612663 borgani , s. , rosati , r. , tozzi , p. , stanford , s. a. , eisenhardt , p. r. , lidman , c. , holden , b. , della ceca , r. , norman , c. , & squires , g. , 2001a , apj , 561 , 13 borgani , s. , governato , f. , wadsley , j. , menci , n. , tozzi , p. , lake , g. , quinn , t. , & stadel , c. , 2001b , apjl , 559 , 71 borgani , s. , finoguenov , a. , kay , s. t. , ponman , t. j. , springel , v. , tozzi , p. , & voit , m. g. , 2005 , mnras , 361 , 233 bower , r. g. , 1997 , mnras , 288 , 355 bryan , g. l. 2000 , apj bryan , g. l. & voit , g. m. , 2001 , apj , 556 , 590 bryan , g. l. , 1999 , comput .
, 46 cavaliere , a. & fusco femiano , r. , 1976 , a&a , 49 , 137 cavaliere , a. , menci , n. , & tozzi , p. , 1997
, apj , 484 , l21 david , l. p. , et al . , 1993 , apj , 412 , 479 da silva , a. c. , kay , s. t. , liddle , a. r. , & thomas , p. a. 2004 , mnras , 348 , 1401 dolag , k. , vazza , f. , brunetti , g. , & tormen , g. , 2005 , mnras , 364 , 753 edge , a. c. & stewart , g. c. , 1991 , mnras , 252 , 414 eisenstein , d. j. & hu , w. , 1999 , apj , 511 , 5 eisenstein , d. j. & hut , p. , 1998
, apj , 498 , 137 eke , v. r. , navarro , j. f. , & steinmetz , m. , 2001 , apj , 554 , 114 evrard , a. e. & henry , j. p. , 1991 ,
apj , 383 , 95 fixsen , d. j. et al .
1996 , apj , 473 , 576 gladders , m. d. , yee , h. k. c. , majumdar , s. , barrientos , l. f. , hoekstra , h. , hall , p. b. , & infante , l. , 2006 , astro - ph/0603588 haiman , z. , et al . , 2005 , astro - ph/0507013 haiman , z. , mohr , j. j. , & holder , g. p. , 2001 ,
apj , 553 , 545 helsdon , s. f. , & ponman , t. j. , 2000 , mnras , 315 , 356 henry , j. p. & arnaud , k. a. , 1991 , apj , 372 , 410 henry , j. p. , 2004 ,
apj , 609 , 603 kaiser , n. , 1986 , mnras , 222 , 323 kaiser , n. , 1991 , apj , 383 , 104 kay , s. t. , thomas , p. a. , jenkins , a. j. , & pearce , f. r. , 2004 , mnras , 355 , 1091 lloyd - davies , e. j. , ponman , t. j. , & cannon , d. b. , 2000 , mnras , 315 , 689 makino , n. , sasaki , s. , & suto , y. , 1998 , apj , 497 , 555 majumdar , s. & mohr , j. j. 2004 , apj , 613 , 41 markevich , m. 1998 , apj , 504 , 27 mccarthy , i. g. , babul , a. , holder , g. p. , & balogh , m. l. 2003a , 591 , 515 mccarthy , i. g. , babul , a. , holder , g. p. , & balogh , m. l. 2003b , 591 , 526 mccarthy , i. g. , balogh , m. j. , babul , a. , poole , g. b. , & horner , d. j. , 2004 , apj , 613 , 811 motl , p. m. , hallman , e. j. , burns , j. o. , & norman , m. l. , 2005 , apj , 623 , l63 muangwong , o. , kay , s. t. , & thomas , p. a. , 2006 , apj , 649 , 640 nagai , d. , 2006 , apj , 650 , 538 navarro , j. f. , frenk , c. s. , & white , s. d. m. , 1995 , mnras , 275 , 720 navarro , j. f. , frenk , c. s. & white , s. d. m. , 1997 , apj , 490 , 493 norman m. l. & bryan g. l. , 1999 , in miyama s. m. , tomisaka k. , hanawa k. , eds , astrophys .
library vol .
240 , numerical astrophysics , kluwer , boston , p. 19
oshea , b. w. , bryan , g. l. , bordner , j. , norman , m. l. , abel , t. , harkness , r. , & kritsuk , a. , 2004 , astro - ph/0403044 oshea , b. w. , nagamine , k. , springel , v. , hernquist , l. , & norman , m. l. , 2005 , apjs , 160 , 1 ostriker , j. p. , bode , p. , & babul , a. 2005 , , 634 , 964 pen , u. , 1999 , apj , 510 , l1 pierpaoli , e. , borgani , s. , scott , d. , & white , m. , 2003 , mnras , 342 , 163 ponman , t. j. , cannon , d. b. , & navarro , j. f. 1999 , nature , 397 , 135 ponman , t. j. , sanderson , a. j. r. , & finoguenov , a. , 2003 , mnras , 343 , 331 pratt , g. w. & arnaud , m. , 2003 , a&a , 408 , 1 pratt , g. w. & arnaud , m. , 2005 , a&a , 429 , 791 pratt , g. w. , arnaud , m. , & pointecouteau , e. , 2006 , a&a , 446 , 429 raymond , j. c. & smith , b. w. 1977 , apjs , 35 , 419 reiprich , t. & bhringer , h , 2002 , apj , 567 , 716 rosati , p. , borgani , s. , & norman , c. , 2002 , aar&a , 40 , 539 shuecker , p. , bhringer , h. , collins , c. a. , & guzzo , l. , 2003 , a&a , 298 , 867 spergel , d. n. , bean , r. , dore , o. , nolta , m. r. , bennett , c. l. , hinshaw , g. , jarosik , n. , komatsu , e. , page , l. , peiris , h. v. , verde , l. , barnes , c. , halpern , m. , hill , r. s. , kogut , a. , limon , m. , meyer , s. s. , odegard , n. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. , apj , submitted , astro - ph/0603449 springel , v. , 2005 , mnras , 364 , 1105 sunyaev , r. & zeldovich , y. , 1972 , comments aprophys .
space phys . , 2 , 66 sunyaev , r. & zeldovich , y. , 1980 , mnras , 190 , 413 tozzi , p. & norman , c. , 2001 , apj , 546 , 63 viana , p. p. & liddle , a. r. , 1996 , mnras , 262 , 1023 voit , m. g. & bryan , g. l. , 2001 , apj , 414 , 425 voit , m. g , bryan , g. l. , balogh , m. l. , & bower , r. g. , 2002 , apj , 576 , 601 voit , m. g. , balogh , m. l. , bower , r. g. , lacey , c. g. , & bryan , g. l. , 2003 , apj , 593 , 272 voit , m. g. & ponman , t. j. , 2003 , apjl , 594 , 75 voit , g. m. , 2005 , rev .
77 , 207 voit , m. g. , kay , s. t. , & bryan , g. l. , 2005 , mnras , 364 , 909 wang , s. , khoury , j. , haiman , z. , & may , m. 2004 , phys .
d. , 70 , 7013008 wu , k. k. s. , fabian , a. c. , & nulsen , p. e. j. , 2001 , mnras , 324 , 95 younger , j. d. , bahcall , n. a. , & bode , p. , 2005 ,
apj , 622 , 1 younger , j. d. , haiman , z. , bryan , g. l. , & wang , s. , 2006 , apj , 653 , 27 | we perform a set of non radiative cosmological simulations of a preheated intracluster medium in which the entropy of the gas was uniformly boosted at high redshift .
the results of these simulations are used first to test the current analytic techniques of preheating via entropy input in the smooth accretion limit . when the unmodified profile is taken directly from simulations , we find that this model is in excellent agreement with the results of our simulations .
this suggests that preheated efficiently smoothes the accreted gas , and therefore a shift in the unmodified profile is a good approximation even with a realistic accretion history .
when we examine the simulation results in detail , we do not find strong evidence for entropy amplification , at least for the high - redshift preheating model adopted here . in the second section of the paper ,
we compare the results of the preheating simulations to recent observations .
we show in agreement with previous work that for a reasonable amount of preheating , a satisfactory match can be found to the mass - temperature and luminosity - temperature relations .
however as noted by previous authors we find that the entropy profiles of the simulated groups are much too flat compared to observations . in particular , while rich clusters converge on the adiabatic self similar scaling at large radius , no single value of the entropy input during preheating can simultaneously reproduce both the core and outer entropy levels . as a result , we confirm that the simple preheating scenario for galaxy cluster formation , in which entropy is injected universally at high redshift , is inconsistent with observations . |
evolving from the initial experiments@xcite , the study of atom focusing using standing wave ( sw ) light fields has developed into a broad area of research in atom optics .
a standing wave field acts as a lens as a result of a spatially - dependent light shift , focusing atoms into a periodic set of lines or dots having widths of the order of tens of nanometers and distanced from one another by @xmath7 , where @xmath8 is the wave length of the light . focusing by thick lenses and atomic deposition on a substrate
have been observed for beams of na@xcite , metastable he @xcite , and cr@xcite .
the latest achievements in this area are summarized in a recent review article@xcite .
perhaps even more significantly , lithographic techniques have been developed to create nanosurfaces in semiconductors and metals using metastable atoms as a pattern template for selective etching@xcite .
similarly , cold , trapped atoms can be subjected to sw light pulses@xcite , creating a periodic wave packet which will focus along the field propagation direction at specific times following the atom - field interaction .
previous experiments with atomic beams have been carried out in a thick lens regime for which the atoms focus within the laser beam .
since the substrate surface is typically placed near the plane of peak intensity of the thick sw lens and mechanically fixed to the retroreflecting mirror , in some sense these experiments are easier to set up when compared to thin lens experiments .
the classical and quantum motion through a thick lens have been numerically simulated and compared to experimental data in refs.@xcite .
the classical motion and lens parameters after thick and thin lenses have also been studied in detail within a ray optics formalism@xcite . for that work
the lens characteristics and aberrations beyond the parabolic lens approximation were simulated numerically as were the effects of different atomic velocity distributions .
however , a comprehensive theoretical study of the thin sw lens from a quantum perspective that details corrections to the lens parameters as a result of spherical aberration , chromatic aberration , and angular beam divergence has not appeared previously .
in this article we stress the flexibility and validity of a thin lens approach , where the atoms are detected or used for lithography after the interaction with the field .
we derive exact and computationally expedient results for atomic matter waves by quantizing the center - of - mass motion from the outset , a process that naturally accounts for spherical aberration .
our hope is to stimulate interest in thin lens experiments by introducing a straightforward , yet rigorous theoretical framework for the problem and by deriving analytical results for lens parameters to facilitate lithographic and atom optical configurations .
the thin sw lens creates high contrast periodic structures in free flight and opens up numerous possibilities for fourier atom optics to image and manipulate modulated matter waves .
this approach is highly desirable , especially when coupled with advances in bose - einstein condensate ( bec ) atom laser sources@xcite with large fluxes and narrow velocity distributions . to our knowledge
similar studies have been carried out only numerically using a fresnel diffraction theory in the coordinate representation for a
doughnut-mode optical field@xcite , standing wave field@xcite , and conical lens@xcite . rather than using the diffraction theory simply for numerical propagation of the matter waves , we extend its usefulness by deriving approximate analytical expressions for the sw field lens parameters .
the fresnel - kirchhoff integral , incorporating the lowest - order spherical aberration , gives a relatively simple form for the atomic wave function near the focal plane and sw intensity extrema .
asymptotically for large atom - field pulse areas , this integral leads to a universal atomic density profile , revealing the behavior of the focal distance , peak atomic density , focal spot size [ half - width at half - maximum ( hwhm ) of the atom density at the focus ] , depth of focus , and background density as a function of the pulse area .
the analytical asymptotic results can then be compared to either numerical asymptotic approximations of the fresnel integral or to the exact atomic distributions and lens parameters , most easily evaluated using a fourier approach .
the time evolution of a free particle s wave function , when subject to a spatially periodic initial condition such as that imposed on an atom beam after passing through a sw field or microfabricated grating@xcite , can be expressed in terms of a fourier expansion in the ( plane wave ) spatial harmonics of the modulation period . as such , the results we derive are closely related to the canon of research on periodic light optics@xcite . in the thin lens approximation the fourier components can be calculated exactly for the wave function and density and then superposed to evaluate the atomic spatial distribution : for this approach the calculations effectively take place in a momentum representation , whereas a numerical integration over the free particle propagator gives results in a coordinate representation .
the expressions for the density s fourier components that we use below were obtained recently for far - detuned @xcite and resonant@xcite sw fields . to our knowledge
this exact solution has not been used previously to examine periodic focusing parameters , nor to compare with approximate solutions of focused waves .
however , the fourier components and the degradation of this type of focusing have been proposed recently as a way to study the onset of many - body effects in condensates@xcite .
the fourier approach has several advantages .
first , the time dependence of the fourier coefficients directly reveals the quantum nature of the periodic atomic density as it varies explicitly with the atomic recoil energy of integer pairs of photons .
second , summation of the fourier series leads to fast convergence of the fresnel solutions to arbitrary accuracy for all times when compared to the numerical integration of the full fresnel - kirchhoff integral . with this fourier method for example
, we can reproduce the main fresnel diffraction results of ref.@xcite without having to integrate the partial differential equation or equivalent integral equation on a lattice . furthermore
, the fourier method can be employed for the general case of typical atomic beam experiments , where the incident beam or trapped atoms are most accurately specified by a density matrix in momentum space , such as a thermal velocity distribution .
the implications are that the spatial extent of the atomic distribution is much larger than the wavelength of the standing wave and the wave function is not transform limited ( except perhaps for the case of a condensate ) , so that the focal densities are conveniently found as a function of this initial density matrix . as a direct result , while the failure to account for finite beam size effects with the fourier series limits our calculations to the near - field and fresnel diffraction regimes , it is relatively unimportant for atom focusing immediately following the sw pulse .
we can then calculate the effects of chromatic aberration and angular beam divergence by averaging the exact fourier solution over any longitudinal and transverse atomic velocity distributions .
in addition to lithographic detection schemes , one can detect atomic distributions modulated by sw fields by backscattering a pulsed traveling wave probe off of the density either ( a ) at a certain distance from the grating for the beam case or ( b ) after a certain time for the atoms released from a trap .
the scattered signal is sensitive to the lowest - order fourier component@xcite of the modulated density .
previously , the backscattered signal from a cold vapor subjected to strong sw pulses was observed in rb in a ground state echo experiment@xcite .
the excellent agreement between the fourier theory and experiment that was achieved in that case insures that our theoretical approach can be applied to the atom focusing problem as well .
a new regime of atom focusing arises if the sw field is resonant with the transition between a ground and excited state and the time of interaction is less than the excited state lifetime . after propagating from the sw field for a time just long enough for the excited state to decay ,
the atom spatial distribution consists of two parts@xcite : the stimulated density , caused by direct amplitude modulation of the ground state wave function by the field , and the spontaneous density , produced by excitation to and decay from the excited state . as a result of the periodic talbot revival , or
self - imaging @xcite , of the atomic spatial distribution , free evolution of the stimulated part leads to atom focusing not only at the focal distance @xmath9 but also at the distances @xmath10 , where @xmath11 is a positive integer , @xmath12 is a talbot length , and @xmath13 is the atom de broglie wavelength .
in contrast to the stimulated part , the spontaneous modulation decays at a distance on the order of @xmath14 as a result of the doppler shift associated with spontaneous emission . at a sufficiently large distance , the modulated spontaneous
part disappears , and one can observe and analyze atom focusing by a resonant sw field .
this analysis then reveals a new effect , a rabi - like oscillation of the focal parameters with the atom - field interaction strength that arises from an interference between different components of the matter wave .
this article is arranged as follows . in section
ii we consider the focusing by a far - detuned field acting as a phase grating .
focusing by a resonant field acting as an amplitude grating is studied in section iii .
section iv is devoted to the influence of chromatic aberration ( longitudinal velocity distributions ) and angular beam divergence ( transverse velocity distributions ) .
the quantitative results and illustrative examples are discussed in section v.
atom optical experiments can operate in the spatial or time domain .
for example , in the time domain a vapor of cold atoms interacts with one or more radiation pulses forming spatial gratings in the @xmath15 direction .
correspondingly , in the spatial domain an atom beam traverses one or more optical elements or interaction regions . for a monovelocity beam propagating along the @xmath16 axis with longitudinal velocity @xmath17 and with the optical elements aligned in the @xmath15 direction ,
the spatial domain configurations can be analyzed in the time domain if calculations are performed in the atomic rest frame moving with velocity @xmath17 . in this frame
the optical elements appear as interaction pulses . as a result ,
our calculations are restricted to the time domain without loss of generality ( and are adapted to account for a longitudinal velocity distribution in section iv ) .
we consider the focusing effects of a single pulse .
the atom optical elements couple to the center - of - mass degrees of freedom of an atom with mass @xmath18 and de broglie wavelength @xmath19 .
when the sw laser field , @xmath20 drives the atomic transition between the ground state @xmath21 and the excited state @xmath22 , the schrdinger equation , @xmath23 , governs the motion .
the hamiltonian in the rotating - wave approximation is @xmath24 where @xmath25 is the two - state wave function for the ground ( @xmath26 ) and excited ( @xmath27 ) states , @xmath28 is the center - of - mass momentum operator , @xmath29 is the atomic transition frequency , @xmath30 and @xmath31 are the frequency , propagation wave vector , and pulse envelope function ( centered at @xmath32 , having peak value of unity and duration on the order of @xmath33 respectively , of the field , @xmath34 is the atomic raising ( lowering ) operator associated with the transition @xmath35 @xmath36 and @xmath37 is a rabi frequency for the @xmath38 transition with dipole matrix element @xmath39 .
we have assumed for simplicity that the rabi frequency @xmath40 is real . in the spatial domain
this hamiltonian is written in the atomic rest frame for the slowly - varying wave function @xmath41 in the paraxial optics limit , @xmath42 for @xmath43 the distance from the interaction region in the lab frame is @xmath44 . in the time domain
a paraxial approximation is unnecessary , and @xmath45 is simply the wave function .
experimentally@xcite , for focusing it has proven advantageous to optically pump the atoms into an initial ground state with magnetic quantum number @xmath46 and use circularly polarized fields to avoid multiple rabi frequencies in the state dynamics , justifying our two - level approximation .
pure phase modulation of the atomic ground state during the interaction occurs in the far detuned case @xmath47 where @xmath48 is the atom - field detuning , and @xmath49 is the excited state decay rate .
often , a steady - state light shift potential , derived from the hamiltonian ( [ 1a ] ) , has been used as an effective atom - field interaction@xcite , @xmath50 .
\label{a4'}\ ] ] while this potential is important for smaller detunings and/or larger intensities , it is strictly valid only for @xmath51 . on the other hand , the dressed state potential , @xmath52 neglects spontaneous emission and assumes that the atom in its ground state adiabatically evolves into one of the two dressed states . for our work
we follow the experimental findings of natarajan and coworkers@xcite .
they show that focusing is improved by taking a short interaction time and a large detuning relative to the rabi frequency and decay rate . in our case
these conditions avoid spontaneous emission ( diffusive aberration@xcite ) both during and after the atom - field interaction . in this limit the two potentials , eqs .
( [ a4 ] ) and ( [ a4 ] ) , reduce to the same effective potential , ignoring spatially independent energy terms .
the hamiltonian in a field interaction representation for the ground state wave function @xmath53 after adiabatically eliminating the excited state for @xmath54 and @xmath55 ( @xmath56 ) is @xmath57 if the incident atom wave function is uniform @xmath58 , $ ] then just after a single interaction the wave function is given by [ a6 ] @xmath59 \label{a6 } \\ & = & \sum_{n=-\infty } ^{\infty } i^{n}j_{n}(\theta /2)e^{inqx } , \label{a6'}\end{aligned}\ ] ] where @xmath60 is an effective pulse area for the far - detuned atom - field interaction , and @xmath61 is a bessel function of order @xmath62 .
equation ( [ a6 ] ) is valid in the raman - nath@xcite or thin lens approximation , for which the @xmath63 term is ignored during the interaction so that the field acts as a standing wave phase grating for the atoms .
a standard condition given for the raman - nath approximation for thin lens focusing is@xcite @xmath64 the two - photon recoil frequency is @xmath65 and @xmath66 is the talbot distance for the atom beam .
the momentum eigenstates , @xmath67,$ ] which are coherently superposed to form this wave function , each have the free particle energy @xmath68 .
before we proceed , a clarification is needed to emphasize the differences between the thin lens ( raman - nath ) and thick lens regimes of the sw atom lens .
some of the quantitative differences have been examined by henkel and coworkers@xcite in relation to fraunhofer diffraction of atoms and from a ray optics ( classical ) perspective by mcclelland for focusing@xcite . for focusing considerations
we assume square pulses , @xmath69 for @xmath70 and zero otherwise , in order to obtain quantitative results .
we also require @xmath71 conditions ( [ 6 ] ) are necessary for high contrast , thin lens atom focusing , where we are interested in the atoms after propagating through the lens .
( the pulse shape will affect the coefficients of our results , not the scaling with @xmath0 and @xmath72 . )
the raman - nath regime leading to condition ( [ 3 ] ) is normally defined as an interaction for which the average kinetic energy gained by the atoms remains much smaller than the interaction strength coupling the momentum components , @xmath73 in the square pulse case .
if this condition is violated while ( [ 6 ] ) holds , the lens is thick , and the atoms can focus within the interaction region near the time @xmath74 . if the atoms have not focused completely by the end of the pulse , they will exit the interaction region amplitude modulated .
to lowest order in @xmath75 , the amplitude correction which multiplies the wave function ( [ a6 ] ) immediately after the square pulse can be calculated to be @xmath76 $ ] .
however , the atoms also acquire an additional , spatially - modulated phase shift , @xmath77 to lowest order in @xmath78 .
this is called the wkb correction by henkel and coworkers@xcite and has its semiclassical origin in the harmonic motion of the atoms during the interaction . from a quantum perspective
these amplitude and phase changes are caused simply by the kinetic energy acquired during the interaction .
heuristically , using the wave function which is evolving in the sw field , @xmath79,$ ] the correction which multiplies @xmath80 can be written as @xmath81=\right .
\\ & & \exp [ \theta \omega _ { q}\tau \cos \left ( qx\right ) /4-i\omega _ { q}\tau \theta ^{2}(1-\cos ( 2qx))/24],\end{aligned}\ ] ] giving the rigorously correct result .
if we are interested in far - field diffraction , the momentum state wave function can change significantly from the spectrum of eq .
( [ a6 ] ) if @xmath82 even if condition ( [ 3 ] ) is satisfied . in the fresnel focal region of interest here , while the spatially - dependent phase shift is crucial in determining the thin lens properties of the sw field , we require @xmath83 to assure the dominance of the raman - nath wave function ( [ a6 ] ) near the focus . in other words a corrected raman - nath condition , @xmath84 is sufficient to observe thin lens focusing , but the additional thin lens condition , @xmath85 , may be necessary to ignore corrections to eq .
( [ a6 ] ) for other observables , like the far - field diffraction pattern or the time - dependent behavior of fourier components of the density in echo configurations . putting the results together ,
the corrected wave function to lowest order in @xmath86 @xmath87 @xmath88 \right .
\\ & & \times \exp \left [ i\theta ^{2}\omega _ { q}\tau \cos \left ( 2qx\right )
/24% \right ] \exp \left [ \theta \omega _ { q}\tau \cos \left ( qx\right ) /4\right ] , \end{aligned}\ ] ] could be incorporated into the theoretical work below if necessary .
we have verified these quantitative results using a crank - nicholson technique to integrate the schrdinger equation for the far - detuned sw hamiltonian ( [ a5b ] ) numerically on a lattice@xcite .
returning to eqs .
( [ a6 ] ) as the initial condition for the free motion , since only the wave function phase has been changed during the interaction , the total atom density , @xmath89 is initially uniform .
as different fourier components of the wave function acquire different phase shifts , @xmath90 , during the free evolution , the atom density for @xmath43 becomes spatially modulated .
the period of the spatial modulation is equal to @xmath91 .
transferring to the dimensionless variables , @xmath92 for @xmath43 one finds@xcite [ 5 ] @xmath93 e^{i(nx - n^{2}t ) } , \label{5a } \\
\rho \left ( x , t\right ) & = & \sum_{n=-\infty } ^{\infty } j_{n}\left [ \theta \sin \left ( nt\right ) \right ] e^{inx}. \label{5b}\end{aligned}\ ] ] this result , exact in the thin lens approximation , is used below for the numerical study of atom focusing .
we refer to the exact calculations by this fourier technique as method 1 in the text and figures that follow .
arbitrarily precise values of the focal time @xmath94 ( giving the focal plane position for the beam , @xmath95 ) for a given @xmath0 are defined by the first maximum of the density along @xmath96 , @xmath97 , as a function of @xmath0 using eq .
( [ 5b ] ) . given the focal time
, we can further characterize the lens by the density profile at the focus @xmath98 and its peak @xmath99 , the spot size of the focus @xmath100 , the depth of focus @xmath101 , and the background density @xmath102@xcite .
the spot size @xmath100 ( hwhm of the density profile @xmath103 where @xmath104 is the maximum ) can be defined implicitly as the smallest positive root of the equation @xmath105 similarly , the depth of focus ( confocal parameter ) can be defined using eq .
( [ 5b ] ) as the time window , @xmath106 within which the density along @xmath96 rises from half its peak value at @xmath107 to its peak value at @xmath94 and back again at @xmath108 , given implicitly by @xmath109 this region is not symmetric with respect to the focal time @xmath94 as a result of the spherical aberration of the lens . (
note that the choice of @xmath110 as positive is unimportant even though it implies a red detuning of the field .
the @xmath111 case is identical but shifted in @xmath15 by @xmath112 . )
several of these exact lens parameters , as calculated by method 1 , are compared to the approximate diffraction theory below .
equations ( [ 5b ] ) will be modified to account for finite beam divergence and chromatic aberration in sec .
iv ; by its nature the fourier method includes spherical aberration ( anharmonicity in the sw potential ) to all orders in the lens curvature .
although eqs .
( [ 5 ] ) are exact , they offer no transparent possibilities for obtaining analytical forms for the lens parameters as functions of @xmath113 in contrast , a diffraction theory can be used to find approximate , asymptotic expressions ( @xmath114 for the @xmath0-dependences of the lens parameters . rewriting eq .
( [ 5a ] ) as an integral in the coordinate representation , we can express the wave function as @xmath115 where the fresnel - kirchhoff propagator of the free atom motion is given in the dimensionless variables ( [ a8 ] ) by @xmath116 the integrand wave function is taken as @xmath117 $ ] . in this form
we are first interested in the wave function behavior near the focal points , @xmath118 for integer @xmath119 and times @xmath120 immediately following the interaction .
the relevant time scale for focusing will become apparent shortly .
when @xmath1 , the main contribution to the integral ( [ a7 ] ) near @xmath121 at these times arises from small values of @xmath122 , where the potential in eq .
( [ a5b ] ) is nearly harmonic .
we choose to expand around the point @xmath123 ( @xmath124 ) and downplay the periodicity of the wave function for the diffraction theory .
replacing @xmath125 by @xmath126 the atom density at the center , @xmath127 contains a singularity at @xmath128 which determines the approximate position of the focal plane . to obtain a finite value for the atomic density , spherical aberration ( _ i.e. , _ anharmonic terms in the potential ) must be considered .
expanding @xmath129 to the @xmath130 term , omitting the phase factor @xmath131 and choosing the scaled position @xmath132 as an integration variable , we find that the asymptotic wave function in the vicinity of @xmath96 is [ a11 ] @xmath133 the main contribution to the wave function comes from the region where the integrand phase originating from the anharmonic term is of the order of unity @xmath134 to determine the validity of eqs .
( [ a11 ] ) , the next anharmonic term ( @xmath135 ) must produce a small addition to the integrand phase in this region , implying @xmath136 or @xmath137 this condition defines the proper asymptotic limit - a wave function correction of relative weight @xmath138 in eq .
( [ a11a ] ) could be included to increase the calculation s accuracy .
we assume condition ( [ a111 ] ) holds . from eqs .
( [ a11 ] ) the time - dependent density , [ a12 ] @xmath139 contains both the slowly - varying @xmath140 dependence and the sharp dependence of @xmath141 in the vicinity of @xmath142 . at the focal center ( @xmath143 ) , the integral ( [ a11b ] ) can be expressed analytically ( see ref .
@xcite , no .
3.696 ) through fractional bessel functions as @xmath144 \right . \nonumber \\ & & \times \left [ j_{-\,^{1}/_{4}}\left ( \tilde{\omega}^{2}/8\right ) + i^{^{3}/_{2}}sign\left ( \tilde{\omega}\right ) j_{^{1}/_{4}}\left ( \tilde{% \omega}^{2}/8\right ) \right ] , \label{a13}\end{aligned}\ ] ] giving an analytical asymptotic expression for the density as a function of time along @xmath145@xmath146 the lens parameters in this asymptotic diffraction theory are defined in the same way as they were for the exact fourier theory above .
they are found graphically or numerically by evaluating eqs .
( [ a11 ] ) , ( [ a12 ] ) , and ( [ a13a ] ) . the focal time @xmath94 for a given @xmath0 ( which is not given by @xmath147 can be found as the first maximum of @xmath148 from eq .
( [ a13a ] ) . using this focal time @xmath94
, we further characterize the sw lens by the peak density @xmath149 evaluating eq .
( [ a13a ] ) , by the density profile at the focus @xmath150 found from eqs .
( [ a11 ] ) and ( [ a12 ] ) , by the spot size @xmath100 ( hwhm of the density profile @xmath151 ) , and by the depth of focus @xmath152 found from eqs .
( [ a12a ] ) and ( [ a13a ] ) .
calculations performed with these asymptotic diffraction results , only restricting @xmath15 to be near @xmath153 , are referred to as method 2 .
in addition to the method 2 asymptotic results , we can derive expressions which are valid both for @xmath154 and for the time restricted to be near the focus , @xmath155 namely within the depth of focus .
this further approximation , which is referred to as method 3 , leads to analytical expressions for the lens parameters in the asymptotic limit . in order to proceed , we need to determine the peak of the function @xmath156 from eq .
( [ a12a ] ) near @xmath157 for @xmath158 .
first , times within the depth of focus @xmath159 are roughly determined by the requirement @xmath160 or @xmath161 and scale as @xmath162 @xmath138 times the focal plane position . for @xmath163 eq .
( [ a12c ] ) justifies the inequality @xmath164 therefore , it is sufficient to replace the slowly - varying time dependence @xmath165 by @xmath0 in eqs .
( [ a11c ] ) and ( [ a12a ] ) , taking the lowest order limit of ( [ a11d ] ) near @xmath157 . in other words
we replace eqs .
( [ a11c ] ) and ( [ a11d ] ) by [ 6p ] @xmath166 respectively , to arrive at the density near the focus , @xmath167 \text{. } \label{7}\ ] ] the subscript @xmath168 reminds us these results are asymptotic .
equation ( [ 7 ] ) can be evaluated numerically . unlike eqs .
( [ a12a ] ) and ( [ a13a ] ) , eq .
( [ 7 ] ) is independent of an explicit dependence on @xmath120 and , therefore , is a universal ( scaled ) density function . inserting eqs .
( [ a13 ] ) and ( [ 6p ] ) into ( [ 7 ] ) , one finds that the asymptotic time evolution of the density along @xmath96 can be written analytically as @xmath169 and is shown in fig .
[ fig01 ] .
approximate expressions for the focal parameters are derived by finding the maximum of the function @xmath170 either numerically using eq .
( [ 7 ] ) or graphically from fig .
[ fig01 ] .
we find that the maximum @xmath171 occurs at @xmath172 .
hence , the asymptotic focal time is given by inverting eq .
( [ 6p ] ) , @xmath173 \simeq \theta ^{-1}\left [ 1 + 1.27\theta ^{-\,^{1}/_{2}}% \right ] .
\label{a14}\ ] ] the asymptotic atom distribution at the focal plane , @xmath174 putting @xmath175 into eq .
( [ 7 ] ) , is also plotted in fig .
[ fig01 ] . for the peak density
we have @xmath176 the asymptotic focal spot size @xmath100 can be expressed through the half - width @xmath177 of the function @xmath178 which is found to be @xmath179 .
inverting eq .
( [ 6p ] ) gives the expression @xmath180 the asymptotic depth of focus , @xmath101 , defined by eqs .
( [ depth1 ] ) and ( [ depth2 ] ) , is evaluated using eq .
( [ 7 ] ) by finding the two values @xmath181 on either side of @xmath172 for which @xmath182 .
these are @xmath183 and @xmath184 giving a depth of focus from inverting eq .
( [ 6p ] ) of @xmath185 equations ( [ 6p])-([a151 ] ) constitute the method 3 results describing the sw lens near @xmath186 and @xmath187 .
= 8 cm = 11.95 cm in much the same way , we can use methods 2 and 3 to derive expressions for the background density , defined as the density @xmath188 at the midpoints between periodic focuses , @xmath189 @xmath190 .
we need to find the wave function , eq .
( [ a7 ] ) , at the point @xmath189 @xmath112 .
since no singularity arises in the integral for @xmath191 @xmath94 when the integrand is expanded around @xmath192 , only the harmonic term is needed to find the wave function , where @xmath193 .
we find the approximate wave function , @xmath194 , and background density , @xmath195 correct to order @xmath196 .
this is the method 2 result .
this expression allows us to find the asymptotic density contrast of importance for lithography , @xmath197 .
the contrast is defined@xcite to be the ratio of the atomic density at the focal points @xmath121 to the background@xmath198 @xmath199 where @xmath97 is given by eqs .
( [ a13 ] ) and ( [ a13a ] ) for method 2 .
in addition , putting the asymptotic focal time ( [ a14 ] ) from method 3 into eq .
( [ back ] ) , the asymptotic background density at the focus is @xmath200 or @xmath201 > from eq .
( [ contrast ] ) the asymptotic contrast ratio at the focus is ( within the accuracy of this calculation ) @xmath202 equations ( [ 102 ] ) and ( [ 103 ] ) are a method 3 result .
= 8 cm = 8.11 cm to summarize , we now have three methods to find the lens parameters , each with different levels of numerical effort required to use it .
the relevant equations for these methods are listed in table 1 : table 1 . techniques and equations for three different methods of calculating thin sw lens parameters .
the numbers in the table correspond to equations in the text .
@xmath203 = 8 cm = 16 cm in fig .
[ fig02 ] the exact atom density at the center @xmath148 and atom distribution at the focal plane @xmath204 , as calculated by method 1 , are compared with the approximate expressions from method 2 . the convergence of the method 2 result as @xmath0 increases is evident . in figs .
[ fig03 ] the pulse area dependences of the focal parameters , as calculated by the three methods , are compared .
the accuracy of method 2 is generally better than that of method 3 , but all methods converge for @xmath205 .
these results are discussed further in sec .
v , including the easily explained , constant offsets of the asymptotic peak densities in fig .
[ fig03]a .
we now consider the density modulation and focusing caused by a resonant standing wave field acting on a two - level atom .
if the pulse duration satisfies the inequality @xmath206 [ the field area @xmath0 for a resonant sw field is defined below by eq .
( [ 7p ] ) ] , then during the interaction the ground and excited state wave functions , @xmath207 and @xmath208 , evolve in an interaction representation according to [ 10 ] @xmath209 by satisfying eq . ( [ 9 ] ) , we are assured that the resonant atom - field interaction is in the raman - nath regime and that the pulse duration is shorter than the excited state lifetime , avoiding the complications of saturation and momentum space diffusion .
if before the interaction the incident wave function is uniform in the ground state , @xmath210 then just after the interaction @xmath211 $ \\ $ \cos [ \left ( \theta /2\right ) \cos \left ( kx\right ) ] $ % \end{tabular } \right ) , \label{a2}\ ] ] where the effective pulse area for the resonant atom - field interaction is defined as @xmath212 the subsequent free space evolution and radiative decay of the system were recently analyzed@xcite using the fourier method for the closed two - level scheme ( _ i.e. _ , the excited state @xmath22 decays only to @xmath213)@xcite .
however , the results that are relevant to the problem discussed here pertain to any excited state decay scheme . in the following paragraphs
we summarize these results in order to apply the three methods of sec .
ii to the resonant case .
our goal is to separate two terms in the free evolution of the ground state : ( 1 ) the strict hamiltonian evolution , resulting from the ground state amplitude produced coherently by the sw pulse , and ( 2 ) the terms which result from spontaneous decay of the excited state .
such a separation is straightforward by the fourier technique , method 1 .
once this is done , we can use a diffraction analysis to obtain asymptotic results for the focal density caused by the hamiltonian term , @xmath214 in eq .
( [ a2 ] ) .
the diffraction analysis , methods 2 and 3 from above , leads to analytic expressions for the resonant sw lens parameters and , anticipating the results , sheds light on the rabi - like oscillation of the lens parameters as a function of @xmath0 .
after the excited state decays ( in a time @xmath215 following the pulse ) , the total atomic density @xmath216 involves only the ground state and is a periodic function of @xmath15 having period @xmath7 .
the atomic system can no longer be described simply by a wave function since we must trace the full density matrix over the spontaneously emitted photons to form the reduced density matrix @xmath217 .
however , the part of the density matrix which was produced coherently by the pulse can still be described by evolving the wave function @xmath214 in eq .
( [ a2 ] ) into @xmath207 .
this was done in ref.@xcite .
the total density @xmath216 , which is formally the diagonal component @xmath218 of the reduced density matrix , can be expanded in a set of fourier components , [ 11 ] @xmath219 e^{inqx } \nonumber \\ & = & \rho ^{\left ( s\right ) } \left ( x , t\right ) + \rho ^{\left ( d\right ) } \left ( x , t\right ) \label{11a } \\ \rho ^{\left ( s\right ) } \left ( x , t\right ) & = & \sum_{n=-\infty } ^{\infty } \rho _ { n}^{\left ( s\right ) } \left ( t\right ) e^{inqx } \label{11b } \\
\rho ^{\left ( d\right ) } \left ( x , t\right ) & = & \sum_{n=-\infty } ^{\infty } \rho _ { n}^{\left ( d\right ) } \left ( t\right ) e^{inqx}\text{. } \label{11c}\end{aligned}\ ] ] each ground state fourier component , @xmath220 consists of two terms : a stimulated term , @xmath221 , caused by the free evolution of @xmath214 , and a spontaneous term , @xmath222 .
after fourier expansion of @xmath214 , the stimulated fourier components of the density are given by squaring @xmath223 to form@xcite @xmath224 \right .
\nonumber \\ & & + \left .
( -1)^{n}j_{2n}\left [ \theta \cos \left ( n\omega _ { q}t/2\right ) % \right ] \right\ } \text{. } \label{12}\end{aligned}\ ] ] in particular , the average stimulated density ( @xmath225 ) is @xmath226 \text{. } \label{12'}\ ] ] each of the stimulated terms , @xmath227 is periodic in time with the talbot period , @xmath228 the @xmath229 spontaneous term , @xmath222 , results directly from the decay of the @xmath229 fourier component of the excited state population . since this decay is accompanied by an atomic recoil , each harmonic s spontaneous term for @xmath230 acquires an additional doppler phase which depends on the momentum @xmath231 of the emitted photon , @xmath232 .
the resulting inhomogeneous dephasing after integrating over spontaneous emission directions leads to a decay of the spontaneous part of the @xmath230 fourier components on a time scale of the order of @xmath233@xmath234 in the spatial domain this is equivalent to a length scale on the order of the talbot distance , @xmath14 . while we could use the complete results of ref .
@xcite to analyze focusing including this spontaneous term for times @xmath235 , we choose instead to simplify the problem by waiting exactly @xmath11 talbot periods @xmath236 , where @xmath11 is a positive integer greater than one . as a result
, one can neglect @xmath237 in eqs .
( [ 11 ] ) . the @xmath225 spontaneous fourier component does not decay . for @xmath225
it follows from the fourier expansion of @xmath238 , for example , that @xmath239 \text{. } \label{spon}\ ] ] thus , in accordance with the conservation of probability , the total average density is one by summing eqs .
( [ 12 ] ) and ( [ spon ] ) , @xmath240 combining these results , the total density is @xmath241 we can now use eqs .
( [ 11 ] ) , ( [ 12 ] ) , ( [ spon ] ) , and ( [ dens ] ) as exact , method 1 expressions to compare to an approximate diffraction theory . to present a diffraction theory of resonant focusing ,
we consider only the stimulated ground state wave function @xmath207 evolving from @xmath242 [ eq . ( [ a2 ] ) ] , remembering that it is first necessary to add an integer number of talbot periods , @xmath243 , to reach an initial
time when the modulated spontaneous contribution can be neglected .
moreover , we must add on the additional background density from the spontaneous term , eq .
( [ spon ] ) , to the final result .
the wave function , @xmath244,$ ] is amplitude modulated ; however , @xmath245 superposes two terms , each of which is phase modulated , [ a3 ] @xmath246 \text{. } \label{a3b}\end{aligned}\ ] ] these states can be seen as evolving independently , leading to the components of @xmath207 we denote as @xmath247 therefore , the problem of focusing after this specific type of amplitude modulation can be mapped onto the problem of focusing by a phase grating . for times
@xmath248 , @xmath249 , near the even antinodes , @xmath250 the component @xmath251 is responsible for focusing while the component @xmath252 evolves smoothly as a background density . near the odd antinodes ,
@xmath253 the role of the components @xmath254 is reversed .
the stimulated density is then given by @xmath255 and the total density , adding on the spontaneous background @xmath256 , by @xmath257 + \rho ^{\left ( s\right ) } \left ( x , t\right ) \text{. } \label{tot}\ ] ] consider the atom density near @xmath96 ( the even antinodes ) for the development of the asymptotic theory . since the period of the component wave functions @xmath254 is twice as large as the far detuned case [ compare eqs .
( [ a6 ] ) and ( [ a3b ] ) ] , the dimensionless coordinate and time become @xmath258 noting that @xmath259 and @xmath260 .
we again assume that @xmath261 . for the @xmath262 component , free evolution in the coordinate representation including the lowest - order spherical aberration gives @xmath263 where the function @xmath264 and variables @xmath265 and @xmath266 are given by eqs .
( [ a11b ] - [ a11d ] ) . to calculate @xmath267 , it is sufficient to consider only the harmonic part of the potential near @xmath145 just as we did for the background term ( [ back ] ) in the phase modulation case . replacing @xmath268 in the fresnel - kirchhoff equation ( [ a7 ] ) by @xmath269 and using the expansion @xmath270
one finds [ a19 ] @xmath271 ^{-\,^{1}/_{2}}x .
\label{a19b}\end{aligned}\ ] ] this contribution remains of order unity at the focus , i.e. , it has relative weight @xmath272 compared to eq .
( [ a18 ] ) . since eq .
( [ a18 ] ) is of higher accuracy than eq .
( [ a19a ] ) ( relative corrections to eq .
( [ a18 ] ) are of order @xmath138 ) , it is valid to add eqs .
( [ a18 ] ) and ( [ a19a ] ) to form the total wave function and calculate the density to absolute order @xmath273 .
interference between @xmath262 and @xmath274 leads to a new effect , a rabi - like oscillation of the focused atom distribution . squaring @xmath275 to form @xmath276
, we find that the asymptotic stimulated atom density near @xmath96 is [ a20 ] @xmath277 ^{^{1}/_{2 } } \nonumber \\ & & \times \left [ \cos \left ( \theta -\tilde{x}_{-}^{2}\right ) f_{c}\left ( \tilde{x},\tilde{\omega}\right ) \right .
\nonumber \\ & & -\left .
\sin \left ( \theta -\tilde{x}_{-}^{2}\right ) f_{s}\left ( \tilde{x},\tilde{\omega}\right ) \right ] \right\ } , \label{a20a } \\ f_{c}\left ( \tilde{x},\tilde{\omega}\right ) + if_{s}\left
( \tilde{x},\tilde{% \omega}\right ) & \equiv & \left ( 2\pi /i\right ) ^{^{1}/_{2}}3^{-\,^{1}/_{4}}f\left ( \tilde{x},\tilde{\omega}\right ) .
\label{a20b}\end{aligned}\ ] ] due to the symmetry between @xmath278 and @xmath279 , this is the density near the focus at both the even and odd antinodes , giving a total density of spatial period @xmath280 . equations ( [ a20 ] ) are the method 2 result for the resonant lens . after accounting for the relevant coordinate and time scales ( [ a17 ] ) , the dominant term , @xmath281 , looks exactly like the far detuned result of sec .
ii and leads to the same lens parameters if taken alone .
the interference term has a relative amplitude @xmath282 near the focus ( @xmath157 ) when compared to the @xmath281 term and oscillates sinusoidally with @xmath283 now we employ method 3 , the limit of this asymptotic result for @xmath284 . considering the interference term as a small correction to @xmath285 ) , @xmath286 , and @xmath179 ] and taylor expand eqs .
( [ a20 ] ) around @xmath288 and @xmath289 .
the following asymptotic expressions can be derived for the focal parameters : [ a21 ] @xmath290 ^{-1}\left [ \cos \left ( \theta \right ) \right .
\nonumber \\ & & \times \partial f_{c}\left ( 0,\tilde{\omega}\right ) /\partial \tilde{\omega% } \nonumber \\ & & -\left .
\sin \left ( \theta \right ) \partial f_{s}\left ( 0,\tilde{% \omega}\right ) /\partial \tilde{\omega}\right ] \right\ } _ { \tilde{\omega}=% \tilde{\omega}_{f } } \nonumber \\ &
\simeq & \theta ^{-1}\left\ { 1 + 1.27\theta ^{-\,^{1}/_{2}}\right . + \theta ^{-\,^{3}/_{4 } } \nonumber \\ & & \times \left [ -0.465\cos \left ( \theta \right ) \right .
\nonumber \\ & & + \left . \left . 0.110\sin \left ( \theta \right ) \right ] \right\ } , \label{a21a } \\
\rho ^{\left ( s\right ) } \left ( 0,t_{f}\right ) & \sim & 3^{^{1}/_{2}}\left ( 4\pi \right ) ^{-1}\theta ^{^{1}/_{2}}\left\ { f\left ( 0,\tilde{\omega}% _ { f}\right ) \right .
\nonumber \\ & & + \theta ^{-\,^{1}/_{4}}\left\ { \left [ \cos \left ( \theta \right ) f_{c}\left ( 0,\tilde{\omega}_{f}\right ) \right .
\right .
\nonumber \\ & & -\left .
\sin \left ( \theta \right ) f_{s}\left ( 0,\tilde{\omega}% _ { f}\right ) \right ] \right\ } \nonumber \\ & \simeq & 0.957\theta ^{^{1}/_{2}}\left\ { 1+\theta ^{-\,^{1}/_{4}}\left [ 0.166\cos \left ( \theta \right ) \right .
\right .
\nonumber \\ & & + \left .
0.703\sin \left ( \theta \right ) \right ] \right\ } , \label{a21b } \\ w & \sim & 3^{-\,^{1}/_{4}}\theta ^{-\,^{3}/_{4}}\left\ { \tilde{x}\right .
\nonumber \\ & & + \theta ^{-\,^{1}/_{4}}\left [ \partial f\left ( \tilde{x},\tilde{\omega}% \right ) /\partial \tilde{x}\right ] ^{-1 } \nonumber \\ & & \times \left [ \cos \left ( \theta \right ) \left ( ^{1}/_{2}f_{c}\left ( 0,% \tilde{\omega}\right ) -f_{c}\left ( \tilde{x},\tilde{\omega}\right ) \right .
. \nonumber \\ & & + \left ( \partial f_{c}\left ( 0,\tilde{\omega}\right ) /\partial \tilde{% \omega}\right ) \left ( \partial f\left ( \tilde{x},\tilde{\omega}\right ) /\partial \tilde{\omega}\right ) \nonumber \\ & & \times \left .
\left ( \partial ^{2}f\left ( 0,\tilde{\omega}\right ) /\partial \tilde{x}^{2}\right ) ^{-1}\right ) \nonumber \\ & & -\left ( \text{the same with replacements } \right .
\nonumber \\ & & \left .
\cos \left ( \theta \right ) \left .
\rightarrow \right .
\sin \left ( \theta \right ) , \,\,f_{c}\left .
\rightarrow \right .
f_{s}\right ) \right ] \right\ } _ { \tilde{\omega}=\tilde{\omega}_{f},\tilde{x}=\tilde{x}_{f } } \nonumber \\ & \simeq & 0.744\theta ^{-\,^{3}/_{4}}\left\ { 1+\theta ^{-\,^{1}/_{4}}\left [ 0.198\cos \left ( \theta \right ) \right .
\right .
\nonumber \\ & & + \left . \left .
0.162\sin \left ( \theta \right ) \right ] \right\ } .
\label{a21c}\end{aligned}\ ] ] = 8 cm = 9.77 cm in fig .
[ fig04 ] the exact time dependence of the peak stimulated density , @xmath276 , and the stimulated density profile at the focal plane , @xmath291 , as calculated numerically by method 1 using eqs .
( [ 11b ] ) and ( [ 12 ] ) for several pulse areas , are compared with the approximate diffraction expression of method 2 , eq .
( [ a20a ] ) .
a comparison of the focal parameters as functions of the pulse area for the three different methods is plotted in fig .
[ fig05 ] , clearly showing the oscillation of the exact and approximate results with @xmath0 .
while a diffraction theory may be unnecessary since an exact result can be calculated by the fourier method , the physical origin of the rabi - like term , which is masked by the fourier result , has been revealed by the diffraction theory to be the interference of the @xmath278 and @xmath279 components of the wave function .
further discussion is again reserved for sec . v. = 8 cm = 9.77 cm
chromatic aberration , the dispersion or wave length dependence of a lens properties , results from a finite distribution of longitudinal velocities ( _ i.e. _ , de broglie wave lengths ) in an atom beam . in trap experiments this type of aberration is avoided . however , for beam experiments detected in the laboratory frame , the atomic distribution would be probed or deposited at a certain distance from the interaction region , @xmath44 . as a result , the total atom density at @xmath292 will be an average of each velocity s density at @xmath292 over the flux of atoms with that velocity .
in general , to calculate the density , we must know the longitudinal velocity distribution or at least its statistical properties .
previous theoretical and experimental publications that have discussed chromatic aberration in the context of atom optical lenses have used heuristic , numerical , or monte carlo simulation approaches @xcite .
we attack the problem from a different perspective , based on averaging the exact fourier components over the atomic flux distribution .
the technique is easy to apply , the results are simple to understand , and the conclusions have a physical interpretation . for the remainder of the paper , the _ ideal _ case will refer to the monovelocity atomic beam with longitudinal velocity @xmath17 and infinitely narrow longitudinal and transverse velocity distributions . in the previous sections we have presented a full characterization of the thin sw lens acting on the ideal beam .
two types of beams with more realistic longitudinal distributions are typical and will be considered here , ( 1 ) a velocity narrowed beam centered around some average velocity @xmath293 or ( 2 ) a thermal beam with average speed @xmath293 . in this section
we consider only focusing by the far - detuned standing wave .
for this type of lens , the exact expression for the density , eq .
( [ 5b ] ) , can be averaged over the proper flux distribution function for a specified @xmath292 by numerical integration , provided the paraxial approximation still holds : @xmath294 .
in addition , the narrow velocity distribution case allows for an approximate analytical solution to compare to the numerically integrated result : by expanding the lens parameters around the average values of the flux distribution , we can account for chromatic aberration analytically using the density of the fourier method . in the laboratory frame the atoms
interact with a sw field with a fixed width @xmath295 along the @xmath16 direction and are then detected at a fixed distance @xmath292 from the lens . as a result ,
the pulse area @xmath0 and the time of free flight @xmath120 after the lens depend on the atomic longitudinal velocity @xmath17 as @xmath296 .
the atomic density is defined as an average over the single - particle atomic flux distribution @xmath297 .
this distribution is a measure of the probability , @xmath298 to find an atom with velocity between @xmath17 and @xmath299 passing through a plane perpendicular to the @xmath16 axis .
( one would divide @xmath297 by the atomic spatial density to get a true flux distribution in inverse units of particles per unit time per unit area . ) to derive quantitative results , we need to specify @xmath297 for a given experiment . in this section
we denote flux averages @xmath300 by @xmath301 taking into account that only atoms having positive velocity , @xmath302 , have to be included . using this definition ,
the average velocity @xmath293 , the ( dimensionless ) relative velocity @xmath303 , and the ( dimensionless ) flux width @xmath304 are given , respectively , as [ 16 ] @xmath305 ^{\,^{1}/_{2}}\text{. } \label{16c}\end{aligned}\ ] ] the pulse area @xmath0 and the time of flight @xmath120 for each velocity class @xmath306 can be defined by reference to the average velocity @xmath293 , @xmath307 where @xmath308 and @xmath309would be computed using @xmath293 . the exact , flux - averaged density @xmath310 follows immediately from eq .
( [ 5b ] ) , @xmath311 \text{. } \label{therm}\end{aligned}\ ] ] the integrals can be evaluated numerically , term - by - term , giving the flux - averaged fourier components of the density .
if the distribution @xmath312 has a non - zero width @xmath304 , then the focusing by the sw lens is degraded .
we show this graphically below .
we start with the narrow velocity distribution case , @xmath313 and derive approximate , analytical expressions from eq .
( [ therm ] ) to compare to an exact numerical integration .
our goal is to determine the effect of a flux width @xmath304 on the focus when compared to the ideal beam results ( @xmath314 ) .
since @xmath315 we can expand the field area and time , eq .
( [ 162 ] ) , as @xmath316 substituting values ( [ 19 ] ) into @xmath216 of eq .
( [ 5b ] ) , expanding in powers of @xmath317 and applying eqs .
( [ therm ] ) and ( [ 16 ] ) , one finds to order @xmath318 that [ 20 ] @xmath319 j_{n}^{\prime } \left [ \theta _ { 0}\sin \left ( nt\right ) \right ] + \theta _ { 0}^{2}\left [ \sin \left ( nt\right ) \right . \nonumber \\ & & + \left . \left .
nt\cos \left ( nt\right ) \right ]
^{2}j_{n}^{\prime \prime } % \left [ \theta _ { 0}\sin \left ( nt\right ) \right ] \right\ } e^{inx } , \label{20b}\end{aligned}\ ] ] where @xmath320 is the lowest - order correction to the density at @xmath292 as a result of the flux distribution . to order @xmath321 four of the focal parameters ( the focal distance @xmath322 defined as the first maximum of the function @xmath323 the peak density @xmath324 the spot size @xmath325 defined as the lowest root of the equation @xmath326 , and the depth of focus @xmath327 defined by @xmath328 ) ,
are given by [ 21 ] @xmath329 _ { t = t_{0,f } } ; \label{21b } \\ \bar{\rho}\left ( 0,t_{f}\right ) & \simeq & \rho \left ( 0,t_{0,f}\right ) + v^{2}\rho _ { 1}\left ( 0,t_{0,f}\right ) ; \label{21c } \\ w & \simeq & w_{0}+v^{2}w_{1 } , \label{21d } \\
w_{1 } & = & \left\ { \left [ \partial \rho \left ( w , t\right ) /\partial w\right ] ^{-1}\left [ ^{1}/_{2}\rho _ { 1}\left ( 0,t\right ) \right .
\right .
\nonumber \\ & & -\rho _ { 1}\left ( w , t\right ) \nonumber \\ & & -\left .
t_{1}\partial \rho \left ( w , t\right ) /\partial t\right ] \right\ } _ { t = t_{0,f}.w = w_{0 } } , \label{21e } \\
\delta t & = & \delta t_{0}+v^{2}\delta t_{1 } , \label{21f } \\
\delta t_{1 } & = & \frac{\,^{1}/_{2}\rho _
{ 1}\left ( 0,t_{0,f}\right ) -\rho _ { 1}\left ( 0,t_{0,+}\right ) } { \left [ \partial \rho \left ( 0,t\right ) /\partial t\right ] _ { t = t_{0,+ } } } \nonumber \\ & & -\frac{\,^{1}/_{2}\rho _
{ 1}\left ( 0,t_{0,f}\right ) -\rho _ { 1}\left ( 0,t_{0,-}\right ) } { \left [ \partial \rho \left ( 0,t\right ) /\partial t\right ] _ { t = t_{0,- } } } , \label{21g}\end{aligned}\ ] ] where @xmath330 , @xmath331 and @xmath332 are the focal distance , focal spot size , and the depth of focus , respectively , for @xmath333 as calculated in sec .
ii for the monovelocity beam .
it is significant that eqs .
( [ 20 ] ) and ( [ 21 ] ) depend only on the dimensionless flux width @xmath304 and not on the exact form of the distribution @xmath334 .
the pulse area dependences of the correction coefficients , @xmath335and @xmath336 , are shown in fig .
[ fig06 ] . to verify eqs .
( [ 20 ] ) and ( [ 21 ] ) for small @xmath304 , we can now quantitatively compare the approximate expressions to the exact , numerically averaged density , eq .
( [ therm ] ) , for a physically reasonable and mathematically convenient distribution function .
= 8 cm = 7.31 cm for example , we can find the flux distribution @xmath337 corresponding to the local maxwellian velocity distribution , @xmath338 .
\label{171}\ ] ] the flux distribution , average velocity , and ( dimensionless ) flux width that follow from @xmath339 can be written with accuracy @xmath340 as [ 172 ] @xmath341 , \label{172a } \\
u_{0 } & = & \widetilde{u}_{0}\left ( 1+v^{2}/2\widetilde{u}_{0}^{2}\right ) , \label{172b } \\ v & = & \frac{v}{\widetilde{u}_{0}}\frac{\sqrt{1-v^{2}/2\widetilde{u}_{0}^{2}}}{% 1+v^{2}/2\widetilde{u}_{0}^{2}}. \label{172c}\end{aligned}\ ] ] as long as @xmath342 we can be assured that @xmath343 .
distribution ( [ 172a ] ) can be inserted into eq .
( [ therm ] ) for @xmath344 , and the exact density can be evaluated at the time @xmath330 for different values of the distribution width @xmath304 using the parametrization of eq .
( [ 172c ] ) .
alternatively , the precise , flux - averaged focal time , @xmath345 , can be found numerically from the first maximum of @xmath346 in time and used to evaluate the density and lens parameters .
the peak reduction and spot size increase for two pulses areas at @xmath330 are plotted as a function of @xmath304 in fig .
[ fig07 ] . in fig .
[ fig08 ] the density profile at the true focus , @xmath347 , for one of these pulse areas is shown with and without chromatic aberration .
both figures are explained in detail in sec . v. = 8 cm = 6.44 cm = 8 cm = 5.5 cm next , we consider the other case of experimental interest , the maxwellian thermal beam having a flux distribution @xmath348 where @xmath349 is now a thermal velocity . for a thermal flux ,
the distribution width @xmath304 , still defined by eq .
( [ 16c ] ) , is not small , @xmath350 on the other hand , since @xmath351 , we expect that the order of magnitude of the density s peak amplitude , the spot size ( or spatial resolution ) of the lens , and the focal position should be the same . in analogy with the analysis above ,
we consider the atomic density at a distance @xmath292 from the sw lens , where @xmath352 is now defined by the free evolution time @xmath353 for an atom with thermal velocity @xmath349 .
( the average thermal flux velocity @xmath293 is given by @xmath354 ) we can now use the thermal flux distribution ( [ 22 ] ) in eq .
( [ therm ] ) to calculate the various , exact lens parameters . to find the focus for this distribution , instead of monitoring the peak atomic density @xmath355
, we monitor the thermally - averaged density contrast @xmath356 from eq ( [ contrast ] ) .
an example of @xmath357 for @xmath358 is shown in fig .
[ fig09 ] .
the focal plane position is defined by the time when @xmath359 is largest , @xmath360 ( @xmath361 ) . the atomic density at @xmath94 , @xmath347 , for this pulse area is also shown in fig .
[ fig09 ] .
several lens parameters as functions of @xmath333 are shown in fig .
[ fig10 ] , including the focal time @xmath94 , the contrast at the focal plane @xmath362 , and the spot size @xmath100 .
= 8 cm = 8 cm = 8 cm = 7.23 cm transverse velocity distributions put an additional constraint on the thin sw lens performance and have been a limiting factor in fresnel atom optics and interferometry .
for an atom beam , the angular divergence is equivalent to the transverse velocity distribution of atoms in a cold trap , again allowing a time domain treatment .
the transverse and longitudinal beam velocities are generally related linearly by the angle in the paraxial limit .
this is the case both for atomic beams collimated by apertures and for beams that are laser - cooled in the transverse direction .
given that we have already discussed chromatic aberration in detail , here we present comparative focal parameters and densities which include either transverse effects alone or simultaneous longitudinal and transverse velocity effects .
only the narrow longitudinal velocity atom beam will be considered in detail since this type of beam leads to a better focus .
we will look at two cases . for both cases
the fourier method suggests a straightforward interpretation of the effects of a transverse velocity distribution in terms of an inhomogeneous decay of the density spatial harmonics .
first , we treat the case where the beam is monochromatic as in sec .
ii and the transverse velocity distribution is either a uniform or one - dimensional maxwellian ( thermal ) distribution of transverse velocities .
the uniform distribution might apply to an atomic beam collimated by a slit or a pair of slits while the thermal distribution might apply to a laser - cooled beam or to atoms released from a laser - cooled trap .
the resulting focal parameters and densities for the slit - collimated and laser - cooled atoms are similar qualitatively .
then , we consider a more general beam case where the transverse velocity and longitudinal flux distributions are _ decoupled , _ but the longitudinal velocity dependence of the time _ couples _ the transverse integration to the longitudinal .
this applies to atomic beams that are laser cooled before interacting with the sw lens .
the theoretical treatment of each of these cases is facilitated by the fourier method .
our first case is the monovelocity longitudinal beam ( @xmath314 ) with a transverse velocity distribution @xmath363 . to unify the longitudinal and transverse results , we start with eq .
( [ therm ] ) and take @xmath364 this is equivalent to starting from eq .
( [ 5a ] ) or ( [ 5b ] ) and taking @xmath365 , @xmath366 and @xmath367 .
assuming we are describing the atomic beam rather than the trap , the atoms are again detected , deposited , or used for lithography at @xmath368 .
the inclusion of an initial transverse velocity @xmath369 during the atom - field interaction will lead to a doppler shift of the field frequency .
( note that the dimensionless transverse velocity scale that is consistent with the dimensionless variables ( [ a8 ] ) is the single - photon recoil velocity , @xmath370 . )
we rigorously account for the doppler shift by taking @xmath371 and averaging the density over @xmath363 , @xmath372 \int du_{x}p(u_{x})e^{inu_{x}t_{0}}\text{. } \label{doppler}\ ] ] the assumption is that the initial transverse density matrix is diagonal in momentum space and can be described by a normalized velocity distribution @xmath373 . in terms of the longitudinal velocity @xmath293 of the beam , the transverse velocity can be written as @xmath374 where @xmath375 is the propagation angle with respect to the @xmath16-axis .
the two transverse distributions of interest , the uniform distribution @xmath376 and maxwellian distribution @xmath377 , take the following form for @xmath378 : [ trans ] @xmath379\text { , } \label{trans1 }
\\ 2.\text { } p_{2}(u_{x } ) & = & ( \sqrt{\pi } v_{x})^{-1}\exp [ -(u_{x}/v_{x})^{2}]% \text{. } \label{trans2}\end{aligned}\ ] ] each can be written as a distribution over beam angles @xmath375 , if desired , by defining the divergence angle @xmath380 .
the first distribution might be formed when a thermal transverse distribution , peaked at @xmath381 @xmath382 with velocity width much greater than @xmath383 is collimated by a pair of slits which select a small range of propagation angles up to @xmath384 .
this beam has a @xmath385 transverse velocity spread of @xmath386 .
the second distribution might be formed by transverse laser cooling , with or without previous slit collimation , and has a @xmath385 velocity width of @xmath387 .
( technically , the laser cooling process can create a non - diagonal density matrix with transverse momentum state coherences that are not accounted for here but can be incorporated into a more general fourier result .
we assume the distribution is narrow and diagonal . ) fortuitously , both distributions again give closed - form fourier coefficients for the density . inserting eqs .
( [ trans ] ) into eq .
( [ doppler ] ) , one finds [ transdens ] @xmath388}{nv_{x}t_{0 } } \nonumber \\ & & \times j_{n}\left [ \theta _ { 0}\sin \left ( nt_{0}\right ) \right ] ; \label{transdens1 } \\ 2.\text { } \rho \left ( x , t_{0}\right ) & = & \sum_{n=-\infty } ^{\infty } \exp [ inx-(nv_{x}t_{0}/2)^{2 } ] \nonumber \\ & & \times j_{n}\left [ \theta _ { 0}\sin \left ( nt_{0}\right ) \right ] \text{. } \label{transdens2}\end{aligned}\ ] ] the interpretation is the same as for an atomic free induction decay experiment@xcite .
the harmonics of the total density undergo an inhomogeneous decay as each velocity class @xmath369 evolves with its own doppler - shifted frequency .
the @xmath389 result was recently explained in detail as it is isomorphic to the evolution of a model , one - dimensional bec or a degenerate fermi gas after interacting with a sw pulse@xcite . a similar expression to the @xmath390 result
was used recently to explain the decay of periodic echoes@xcite and the decay of a quasiperiodic atomic focusing and talbot scheme@xcite . for both distributions the harmonic decay functions of eqs .
( [ transdens ] ) depend on @xmath391 near the focus , @xmath392 for @xmath393 only harmonics greater than @xmath394 undergo significant damping , so it appears that the transverse velocity width can be up to the order of the recoil velocity to avoid a significant decay of the high - order fourier components necessary for a highly - peaked focus
. however , the restriction @xmath393 may be sufficient but not necessary .
we have found in the previous sections that the lens parameters scale with @xmath395 and , therefore , in sec .
v we show that the restriction on @xmath396 is somewhat relaxed . the focal distance @xmath94 , defined as the first maximum of eq .
( [ transdens1 ] ) or ( [ transdens2 ] ) , moves closer to the lens for @xmath397 when compared to the focal distance for @xmath398 this makes sense since the transverse distribution causes a decay in the modulated density in time .
however , this shift of the focal distance is secondary in importance to the changes in the peak density at the focus and the spot size . moreover , as @xmath399 increases , the focal distance including the transverse velocity spread converges rapidly to the ideal value near @xmath400 .
we have calculated the peak density and spot size for selected values of @xmath396 and compared them to the ideal case . in fig .
[ fig11 ] the peak focal density , @xmath401 and spot size @xmath100 are plotted versus @xmath402 for @xmath403 ( ideal ) , @xmath404 ( recoil limit)@xmath198 and @xmath405 ( twice the recoil limit ) . for each of the five curves , the exact , different focal time at each value of @xmath402
is used to calculate the parameters . for the same @xmath385 velocity spread
, the @xmath406 distribution gives a narrower , more peaked density spot than the @xmath407 distribution .
it is significant that the percentage error in the peak density and spot size is decreasing as @xmath402 increases .
this result is expected since inhomogeneous decay is less pronounced at earlier times , @xmath408 . to treat the full problem of focusing an atomic beam with velocity distributions more generally , we must account for the angular relation between the longitudinal and transverse velocities . the atomic velocity in a beam , @xmath409 can be written @xmath410 for small angles @xmath411 where @xmath412 to lowest order in @xmath375 in the paraxial approximation . since atomic beams
are inherently formed by a process which truncates either the divergence angle or the transverse velocity distribution , it is not a simple matter to describe them theoretically .
however , if the beam is localized around some large longitudinal speed and then laser cooled in the transverse direction before the sw lens , the longitudinal and transverse distributions are effectively independent . we can model the total ( normalized ) flux probability @xmath413 as a product of the two independent distributions , @xmath414 using the local maxwellian longitudinal flux distribution @xmath415 eq .
( [ 172a ] ) , and thermal transverse velocity distribution @xmath416 eq .
( [ trans2 ] ) .
= 8 cm = 7.34 cm since the time of flight , @xmath417 to the detection plane , @xmath418 depends on the longitudinal velocity , the doppler phase @xmath419 does as well .
inserting the integration of the doppler phase over the distribution @xmath377 into eq .
( [ therm ] ) , we find the velocity - averaged density to be [ densp ] @xmath420 \nonumber \\ & & \times \left\ { j_{n}\left [ \left ( \theta _ { 0}u_{0}/u\right ) \sin \left ( nt_{0}u_{0}/u\right ) \right ] \right\ } .
\label{densp2b}\end{aligned}\ ] ] the fourier component of the density now has an additional gaussian multiplier to integrate over longitudinal velocities .
if the longitudinal width @xmath304 is small but non - negligible and we want @xmath393 ( laser cooling to the recoil limit ) , the argument of the gaussian is on the order of the argument of the sine function . as a result
, we can not neglect the gaussian decay in the integral by assuming @xmath297 is sharply peaked around @xmath293 and thus evaluating it at @xmath365 ( @xmath314 limit ) . in sec .
v we treat two examples of experimental interest with this expression , an atomic beam which is laser - cooled to the transverse recoil limit , both with and without a longitudinal flux width @xmath421 this is shown in fig .
[ fig08 ] along with the ideal and purely longitudinally - broadened cases .
in this article we have developed a theory of atom focusing by standing wave light fields in the thin lens regime .
we have shown that the exact analytical expressions for the fourier components of the atomic density , recently used in the theory of atom interferometry , are useful for numerical calculations of the focusing effect and all its relevant lens parameters , including the effects of chromatic aberration and angular divergence .
we call the fourier technique method 1 .
thin lens focusing becomes especially effective for a large field area @xmath110 .
when @xmath422 for the monovelocity case , using the fresnel - kirchhoff diffraction integral , we can consider only the first anharmonic term of the phase shift created by the sw light field when expanding around the potential wells at @xmath423 .
this procedure , labeled method 2 , leads to a manageable expression ( [ a12 ] ) for the atomic density distribution which can be used to evaluate the focal parameters by numerical integration . within a small vicinity of the focal time ,
( [ a12 ] ) simplifies further to the asymptotic density profile ( [ 7 ] ) .
this is method 3 .
the analytical correction to the focal time and analytical expressions for the peak focal density , spot size , and depth of focus as functions of the pulse area @xmath0 follow from this result .
the last three lens parameters scale as @xmath424 , @xmath425 and @xmath426 respectively , while the focal time scales as @xmath427 [ see eqs .
( [ a14 ] ) , ( [ a15 ] ) , ( [ a16 ] ) , and ( [ a151 ] ) ] . from fig .
[ fig03 ] one sees that in the range @xmath428 the accuracy of the method 3 asymptotic expressions for the peak focal density , focal time , and spot size can be as small as 17% , 1.9% and 18% respectively .
the asymptotic parameters obtained by method 2 using eqs .
( [ a12 ] ) are generally more precise , giving accuracies for the same parameters as small as 5.8% , 0.24% and 4% , respectively . the accuracy of the asymptotic result for the focal time @xmath94 is an order of magnitude better than for other focusing parameters . evidently , this is a consequence of the fact that a finite value for the focal time , @xmath429 , comes out of the harmonic approximation for the lens potential .
the first anharmonic term then leads to a relative correction of the order of @xmath430 [ see eq . ( [ a14 ] ) ] , and higher - order anharmonic corrections ( not included in our consideration ) should have relative weights @xmath431 and larger inverse powers of @xmath0 . as for the other parameters , which can be derived only by including the first anharmonic term , the next order corrections should have a larger relative weight of @xmath430 .
for example , the constant offsets of the asymptotic peak densities from the exact result in fig .
[ fig03 ] can be explained as follows .
a correction to the asymptotic wave function at @xmath96 [ eq .
( [ a13 ] ) ] of relative weight @xmath430 would result from including the next order spherical aberration , the @xmath432 term in the expansion of @xmath433 in eq .
( [ a11a ] ) . when the wave function is squared to form the density , this relative correction of weight @xmath430 multiplies the lowest - order peak density that grows as @xmath424 to give a constant term , which is automatically accounted for by the exact fourier method .
hence , the constant offset of the asymptotic peak density is not included in our asymptotic expressions . in this article we have also analyzed a new type of atom focusing which arises when the sw field is resonant with the atomic transition .
the density profiles and focal parameters for the resonant lens are shown in figs .
[ fig04 ] and [ fig05 ] , respectively . to gain insight into the different interactions that can lead to focusing
, we can consider the far - detuned and resonant cases from a more general point of view than that taken in secs .
ii and iii , still assuming that the field envelope is short enough to neglect spontaneous emission during the pulse .
when an atom interacts with a pulsed light field , the states of the system can be decomposed into a set of semiclassical dressed states of the atom plus field . if the atom - field interaction is adiabatic or the pulse turns on instantaneously , these dressed states are instantaneous eigenstates of the total hamiltonian and therefore undergo a phase modulation
when the light field modes have a modulated intensity , the dressed state energies and , therefore , phase evolutions are also spatially modulated .
focusing can occur near the intensity extrema , which correspond to spatial minima of the dressed state energies .
if initially the bare atom was in a pure state such as the ground state @xmath434 , the atomic wave function after the interaction is generally a coherent admixture of the two dressed states , which have an energy separation that is spatially modulated .
therefore , an interference term in squaring the bare state amplitudes can lead to a rabi - like oscillation of the total density and its properties .
this rabi - like oscillation can be seen in eq .
( [ a20a ] ) and fig .
[ fig05 ] for a resonant sw field .
nonadiabatic , detuned atom - field couplings can lead to a similar effect .
the evolution of the pure and dressed states in this regime has been considered previously without center - of - mass spatial effects@xcite . for the far detuned standing wave @xmath435 or an adiabatic turn - on of the field
, the ground state evolves into only one of the dressed states while the other dressed state has a negligible amplitude . as a result
, one does not observe any interference or field dependent oscillation in this case . up to this point in the discussion
, we have reviewed the ideal situation of the ideal atomic beam having no angular divergence and no longitudinal velocity distribution .
our calculations in the thin lens regime have shown that focusing results in a relatively large peak focal density .
for the most part , this result is in contrast to previous experiments carried out in the thick lens regime , where it proved difficult to achieve such large density peaks .
the exceptions are certain experiments in na at @xmath436 on the @xmath437 transition@xcite . in these thick lens experiments for a thermal beam having thermal velocity @xmath438 the spot size resolution and contrast were @xmath439 @xmath440@xcite and @xmath441 @xmath442@xcite , respectively . as an exercise
, we can compare that experimental data using a thermal beam and thick sw lens with our results using a ideal atomic beam and thin lens . from fig .
[ fig03 ] or an exact calculation , a pulse area of @xmath443 ( @xmath444 is needed for a resolution of @xmath445 ( @xmath445 ) , or @xmath446 @xmath447 ( @xmath448 @xmath447 ) in sodium . if we set the beam velocity @xmath17 equal to @xmath449 then focusing for @xmath443 occurs at a distance @xmath450 the atomic density profile at this distance is shown with the thin solid line in fig .
[ fig08 ] .
the contrast at the focus , defined by eq .
( [ contrast ] ) as the ratio of focal atomic densities at @xmath451 and @xmath452 ( not shown in fig .
[ fig08 ] ) , is @xmath453 we can determine the standing wave power needed to achieve a pulse area of @xmath454 for the two - level system .
this system can be realized in na using optical pumping to the ground state sublevel having magnetic quantum number @xmath455 a circularly polarized standing wave drives only the @xmath456 transition which has the dipole moment matrix element @xmath457 @xmath458 . to achieve the intensity necessary for atomic focusing
, the sw field must be formed by focused laser beams . in the experiments of refs.@xcite
, the fields have been focused to a circular spot having radius @xmath459@xcite . for experiments in the thin lens regime , it is sufficient to focus the field only in the @xmath16-direction , along which the atomic beam propagates .
we assume that the laser field intensity has a gaussian profile in the @xmath460 plane , where @xmath295 and @xmath462 are field radii along the @xmath16- and @xmath463-axes , respectively ( i.e. , the directions perpendicular to the laser beam propagation along the @xmath15 axis ) . for homogeneous atomic focusing into a set of lines , one should choose @xmath462 to be larger than the atomic beam radius . from the data in refs.@xcite , @xmath462 should be as large as @xmath464 the general field mode ( [ 25 ] ) can be created using cylindrical optical lenses .
the far - detuned pulse area @xmath0 ( [ 2 ] ) can be re - expressed through the constituent traveling wave field powers of eqs .
( [ 1 ] ) and ( [ 25 ] ) , @xmath465 , as @xmath466 for further estimates we assume that @xmath467 , @xmath468 and @xmath469 then , for a field area of @xmath470 the required laser power is @xmath471 even this power level is four times less than the power used in the experiment of ref.@xcite . to compare directly to the circular focus of the field in ref.@xcite , we must take @xmath472 giving us a required power of @xmath473 these relatively low power levels are sufficient for our consideration of thin lens focusing .
we can now discuss the influence of chromatic aberration assuming no angular divergence .
a longitudinal velocity distribution in the atomic beam leads to a degradation of the focus .
but even for the thermal beam , the normalized distribution width @xmath304 is still less than @xmath474 [ see eq .
( [ 221 ] ) ] , and we have shown that the order of magnitude of the focusing parameters is the same as for the ideal beam ( @xmath475 ) . for narrow flux distributions , @xmath315 the lowest - order corrections to the atom density behave as @xmath318 [ see eqs .
( [ 20 ] ) ] if @xmath476 .
this is a simple consequence of the fact that terms linear in the deviation from the average atomic flux vanish after averaging over velocities for any symmetric , or asymmetric , distribution .
in particular , this principle applies to the analytical fourier components of the density multiplied by the narrow , local maxwellian flux distribution @xmath477 , eq .
( [ therm ] ) , for small @xmath304 . in fig .
[ fig07 ] we show the exact ( using @xmath337 ) and approximate reductions of the density at the ideal beam focal time , @xmath478@xcite@xmath198 as functions of @xmath304 for two values of @xmath402 , @xmath479 and @xmath480 .
again , these choices for @xmath402 are the pulse areas required to achieve @xmath481 and @xmath482 @xmath447 spot sizes @xcite , respectively , using ideal sodium beams . the quadratic dependence , consistent with eqs .
( [ 21c ] ) , is evident for small @xmath304 , but the exact density peak actually reduces more slowly ( linearly ) at larger values of @xmath304 where the small @xmath304 approximation breaks down . at this point
the expansions of the bessel functions of eq .
( [ therm ] ) near @xmath483 in terms of @xmath303 are invalid since @xmath484 , or equivalently @xmath485 : a term proportional to @xmath486 would appear in an expansion around the average velocity .
we take this as evidence that near this value of @xmath304 for a fixed @xmath487 the focal properties go from being dominated by spherical aberration to a regime where a more complicated combination of spherical and chromatic aberrations is important . to test this idea ,
we have also examined the dependence of the spot size increase , @xmath488 on @xmath304 for @xmath489 and @xmath480 using the exact density for @xmath337 . in fig .
[ fig07 ] the spot size s quadratic dependence for small @xmath304 turns into a linear dependence near the same value of @xmath304 that the peak density reduction deviates from the approximate theory .
we see that the chromatic aberration has a smaller effect on the spot size than on the peak density .
note that from a ray tracing argument for far - field focusing through an effective parabolic lens aperture of full width @xmath490 , a strong , @xmath0-independent , linear dependence of @xmath491 ( or equivalently @xmath492 ) would be expected for a focus dominated by chromatic aberration in the absence of spherical aberration@xcite .
however , this is not the case for the sw field lens , where spherical aberration determines the limiting resolution up to large values of @xmath402 .
thus , the chromatic aberration we find is less severe and depends strongly on @xmath333 .
we have not explored this @xmath402 dependence further . however , from fig .
[ fig07 ] we deduce that if we fix @xmath304 ( and by implication the atomic beam properties ) , a threshold pulse area , @xmath493 , must exist for each @xmath304 to mark the breakdown of the small @xmath304 expansion .
for @xmath494 the effects of chromatic aberration should be calculated exactly . to further demonstrate the small @xmath304 results for a narrow longitudinal distribution
, we can define @xmath495 to be the percentage reduction of the peak focal density as a function of @xmath402 and @xmath304 .
the flux distribution width for a given @xmath402 and @xmath495 is therefore @xmath496 from eq .
( [ 21c ] ) . from fig .
[ fig06 ] or [ fig07 ] for @xmath497 , to limit the peak density reduction to @xmath498 , we can estimate that a normalized flux width of @xmath499 would be needed .
( for the local maxwellian distribution , @xmath415 a flux width of @xmath500 is found to be sufficient by an exact calculation with eq .
( [ therm ] ) . ) using @xmath499 to calculate the other parameters approximately from eqs .
( [ 21b ] ) and ( [ 21e ] ) , the percentage corrections to the focal distance and spot size are even less than @xmath501 @xmath502 and @xmath503 respectively . a longitudinal width
, @xmath504 is typical of the supersonic beams produced by seeding an inert ( noble ) gas supersonic expansion with sodium@xcite .
while the mean longitudinal velocity of a beam produced recently from a bec of sodium atoms by bragg scattering@xcite was only @xmath505 or @xmath506 @xmath507 the @xmath385 longitudinal and transverse velocity widths were only approximately @xmath508 and @xmath509 , respectively , giving @xmath510 and justifying the paraxial approximation .
in these experiments , mean longitudinal velocities up to @xmath511 @xmath512 or @xmath513 @xmath514 , with similar widths were also achieved using higher - order bragg scattering , implying @xmath515 .
since experiments producing atomic beams of this type are in their infancy , we expect that achieving relative velocity widths of @xmath516 will shortly become routine . in fig .
[ fig08 ] the atomic distribution at the focus @xmath517 ( dotted line ) , as calculated exactly using @xmath337 in eq .
( [ therm ] ) for @xmath489 and our extreme case from fig .
[ fig07 ] of @xmath518 ( @xmath519)@xmath198 is compared to the ideal case ( thin solid line ) .
while the reduction of the peak density from @xmath520 to @xmath521 is a significant @xmath522 on a scale where @xmath523 is the background at the focus , the half - width defining the spot size is broadened by only @xmath524 ( or @xmath525 @xmath447 in sodium ) , and the coherent oscillations of the density along @xmath15 are still prevalent .
this suggests that even for @xmath526 and a focal peak reduced by chromatic aberration , a high resolution , large contrast focus is possible with the thin sw lens .
in addition , using the flux distribution @xmath527 with @xmath528 we have also calculated that the field area , @xmath529 is required to restore a spot size of @xmath445 , or @xmath530 @xmath447 in sodium .
this corresponds to the laser power , @xmath531 for @xmath532 , @xmath533 @xmath534 and @xmath535 .
we have also considered the focusing of a thermal atomic beam with flux @xmath536 .
the results of the calculations are graphed in figs .
[ fig09 ] and [ fig10 ] . for this part of the paper only , we changed the definition of the focal distance to be one where the contrast ( [ contrast ] ) , and not the density , is optimized .
the contrast , @xmath537 for @xmath538 at a distance @xmath292 from the sw field @xmath539 is shown in fig .
[ fig09 ] .
we see that @xmath540 contains three local maxima of approximately equal weight near the focus .
they arise as a result of the time - oscillations of the flux - averaged , background atom density near @xmath360 , @xmath541 .
these contrast maxima
compete with one another in some sense as different values of the pulse area correspond to different velocity classes and therefore different focal distances .
this effect results in the discontinuities in the pulse area dependences of the lens parameters seen in fig .
[ fig10 ] . for the longitudinal thermal beam and @xmath358 ,
the focal contrast is equal to @xmath542 while the spot size or spatial resolution is again @xmath543 for the same experimental parameters as above , @xmath532 , @xmath544 @xmath534 and @xmath545 , a laser field power of @xmath546 is needed to produce a pulse area of @xmath358 .
finally , the atomic beam angular divergence has a crucial impact on the focus , independent of the chromatic aberration .
for a beam with no chromatic aberration , @xmath365 , if atoms move at an angle @xmath375 with respect to the @xmath16-axis , the density profile displaces a distance @xmath547 along the @xmath15 axis .
this displacement should ideally be much smaller than the fundamental spot size @xmath100 that accounts for spherical aberration .
we can use the previously generated , exact spot sizes and focal times for each @xmath402 to estimate the maximum allowed @xmath385 transverse velocity from this argument , @xmath548 .
this corresponds to an angular divergence @xmath549 .
we expect that allowing @xmath550 to be @xmath551 will at most double the spot size . alternatively , from eqs .
( [ a14 ] ) and ( [ a16 ] ) for dimensionless @xmath552 and @xmath100 , the asymptotic restriction on the angular divergence , @xmath553 is @xmath554 it is interesting to note that the asymptotic focusing restriction is @xmath555 times less severe than the recoil limit , @xmath556 which is the condition required to observe atom interference in the single interaction region geometry ( for example , the atomic talbot effect@xcite ) .
condition ( [ 27 ] ) arises from the requirement that atoms moving at an angle @xmath375 not be displaced more than @xmath7 at the talbot distance @xmath557 . even though effective focusing requires the displacement to be @xmath558 times smaller asymptotically ( since @xmath559 ) , this displacement occurs at a distance that is @xmath0-times shorter than the talbot distance ( since @xmath560 ) .
this argument is consistent with eqs .
( [ transdens ] ) and fig .
[ fig11 ] .
this figure suggests that small spot sizes with sharp density peaks can be achieved for laser powers in the microwatts to milliwatts range even for transverse distribution widths larger than the recoil limit .
. [ fig11 ] , note that @xmath561 corresponds to a @xmath562 @xmath447 spot size in sodium . ) for example , taking @xmath563 and remembering that the slit - collimated beam has a @xmath385 transverse velocity @xmath564 ( @xmath565 ) in units of @xmath566 , the exact spot size for @xmath567 is @xmath568 from eq .
( [ transdens1 ] ) .
now , assuming the divergence is the maximum allowed , we set @xmath569 for @xmath563 and find that the spot size is broadened to @xmath570 ( @xmath571 @xmath447 in sodium ) , or @xmath572
. the spot size has less than doubled for a transverse width almost twice the recoil limit .
similarly , the asymptotic predictions for @xmath563 give an increase from a spot size of @xmath573 ( not @xmath574 ) for @xmath567 to a spot size of @xmath575 for @xmath576 from eq .
( [ 26 ] ) in eq .
( [ transdens1 ] ) , a @xmath577 increase .
the broadening is even less severe for maxwellian transverse distributions .
we now take a case of experimental interest to demonstrate the combined effects of transverse and longitudinal broadening on the focal density . the pulse area , @xmath489 ,
is used again .
we show results for the @xmath406 ( maxwellian ) transverse distribution , eq .
( [ trans2 ] ) , since laser cooling has been shown to improve the focus in thick lens experiments@xcite . in fig .
( [ fig08 ] ) we plot the density at the focus @xmath98 ( dashed line ) from eq .
( [ transdens2 ] ) for a monochromatic beam ( @xmath314 ) cooled to the @xmath385 single - photon recoil limit , @xmath578 @xmath579 in addition , we show the atomic focal density ( lower solid line ) from eq .
( [ densp2b ] ) for @xmath497 , @xmath580 ( @xmath519 in @xmath337 ) , and @xmath581 .
the latter curve combines the attributes of the other curves in this figure to provide a real picture of the focus .
the following table summarizes the results for fig .
[ fig08 ] : our work suggests emphatically that the thin sw lens can focus atoms effectively if the angular divergence is cooled near the recoil limit , even in the presence of strong chromatic aberration .
this makes physical sense if one considers that the longitudinal velocity average over slightly different focal regions for each velocity subclass is a slowly - varying integral over the fourier _ amplitudes _ when compared to the more sensitive transverse velocity integral , an average over the doppler _ phases_. atomic beams of the type considered throughout this article can be made in the laboratory with current technology .
in addition to the experimental possibilities for standard atomic beams that limit angular divergence by slit collimation or transverse laser cooling , recent experiments that use higher - order bragg diffraction of a bec to form atomic beams with transverse and longitudinal @xmath385 velocities smaller than the recoil velocity@xcite offer a promising avenue to observe and characterize thin lens focusing effects .
we would even argue that focal patterns nearly identical to those described in this paper have already been achieved but not recognized in recent condensate experiments on the so - called momentum space talbot effect@xcite . in these experiments
a cloud of cold atoms from a condensate interacts with two off - resonant sw pulses , separated by a variable time delay .
the detection scheme is insensitive to fresnel effects as the researchers image the far - field diffraction pattern and therefore the atomic momentum distribution .
however , the accurate fit of the data in that experiment to our raman - nath theory suggests that the thin lens focusing effect is occurring after the first pulse along the sw grating direction .
a detection scheme sensitive to the fresnel density pattern may be necessary to image the focusing atoms in the cloud .
is indebted to prof .
tycho sleator and the nyu physics department for providing him with the visiting scholar appointment during which this work was completed .
this work is supported by the national science foundation under grants no .
phy-9414020 and phy-9800981 , by the u.s .
army research office under grant no .
daag55 - 97 - 0113 and aasert no .
daah04 - 96 - 0160 , and by the university of michigan rackham predoctoral fellowship . j.j .
mcclelland ,
nanofabrication via atom optics _ , _ nist electronic physics group , preprint , to appear in _ handbook of nanostructured materials and nanotechnology , _ ed . by h. s. nalwa , to be published by academic press .
berggren , a. bard , j.l .
wilbur , j.d .
gillaspy , a.g .
heig , j.j .
mcclelland , s.l .
rolston , w.d .
phillips , m. prentiss , and g.m .
whitesides , science * 269 * , 1255 , ( 1995 ) ; s. nowak , t. pfau and j. mlynek , appl . phys .
b * 63 * , 203 ( 1996 ) ; k.s .
johnson , k.k .
berggren , a. black , c. black , a.p .
chu , n. dekker , d. ralph , j.h .
thywissen , r. younkin , m. prentiss , m.tinkham , and g. whitesides , appl .
. lett . * 69 * , 2773 ( 1996 ) other definitions of the focal time are possible , such as the earliest time that maximizes the ratio or the difference between the peak and background densities , or the point where the peak - width product is a maximum .
the definition we have chosen , the first maximum of @xmath589 is sufficient to characterize the lens fully .
furthermore , this definition is consistent with the optical lens focus and leads to nearly identical results as a definition where the difference in central and background densities , @xmath590 is maximized .
we also use the term _ spot size _ instead of line size or linewidth to refer to the hwhm of the focused line created by the one - dimensional sw lens . note that by fixing @xmath330 we are not actually using the time @xmath591 where the function @xmath592 is a maximum when evaluating @xmath593 .
the correction is small even for the largest value of @xmath304 plotted .
see , for example , j. schmiedmayer , m.s .
chapman , c.r .
ekstrom , t.d .
hammond , d.a .
kokorowski , a. lenef , r.a .
rubinstein , e.t .
smith , and d.e .
pritchard _
_ in _ atom interferometry _ , ed . by p.r
berman , academic press , san diego ( 1997 ) . | the focusing of two - level atoms in a beam or trap after interacting with both far - detuned and resonant standing wave fields in the thin lens and paraxial approximations is considered theoretically .
the thin lens approximation is discussed quantitatively from a quantum perspective .
exact quantum expressions for the fourier components of the density ( that include all spherical aberration ) are used to study the focusing numerically .
the following lens parameters and density profiles are calculated as functions of the pulsed field area @xmath0 : the position of the focal plane , peak atomic density , atomic density pattern at the focus , focal spot size , depth of focus , and background density .
the lens parameters are compared to asymptotic , analytical results derived from a scalar diffraction theory for which spherical aberration is small but non - negligible ( @xmath1 ) . within the diffraction theory analytical expressions show that the focused atoms in the far detuned case have an approximately constant background density @xmath2 while the peak density behaves as @xmath3 , the focal distance or time as @xmath4 , the focal spot size as @xmath5 , and the depth of focus as @xmath6 .
focusing by the resonant standing wave field leads to similar results .
however , resonant focusing is also accompanied by a new effect , a rabi - like oscillation of the atom density . for the far - detuned lens , chromatic aberration caused by the longitudinal velocity distribution in an atom beam
is studied quantitatively with the exact fourier results .
similarly , the degradation of the focus that results from angular divergence in beams or thermal velocity distributions in traps is studied quantitatively with the exact fourier method and understood analytically using the asymptotic results .
overall , we show that strong thin lens focusing is possible with modest laser powers and with atomic beam characteristics that are currently achievable in the laboratory . 0.4 in epsf 2 |
neurons in the visual cortex form spatial representations or maps of several stimulus features . how are different spatial representations of visual information coordinated in the brain ? in this paper , we study the hypothesis that the coordinated organization of several visual cortical maps can be explained by joint optimization .
previous attempts to explain the spatial layout of functional maps in the visual cortex proposed specific optimization principles ad hoc . here , we systematically analyze how optimization principles in a general class of models impact on the spatial layout of visual cortical maps . for each considered optimization principle
we identify the corresponding optima and analyze their spatial layout .
this directly demonstrates that by studying map layout and geometric inter - map correlations one can substantially constrain the underlying optimization principle .
in particular , we study whether such optimization principles can lead to spatially complex patterns and to geometric correlations among cortical maps as observed in imaging experiments .
neurons in the primary visual cortex are selective to a multidimensional set of visual stimulus features , including visual field position , contour orientation , ocular dominance , direction of motion , and spatial frequency @xcite . in many mammals ,
these response properties form spatially complex , two - dimensional patterns called visual cortical maps @xcite . the functional advantage of a two dimensional mapping of stimulus selectivities is currently unknown @xcite . what determines the precise spatial organization of these maps ?
it is a plausible hypothesis that natural selection should shape visual cortical maps to build efficient representations of visual information improving the fitness of the organism .
cortical maps are therefore often viewed as optima of some cost function .
for instance , it has been proposed that cortical maps optimize the cortical wiring length @xcite or represent an optimal compromise between stimulus coverage and map continuity @xcite .
if map structure was largely genetically determined map structure might be optimized through genetic variation and darwinian selection on an evolutionary timescale .
optimization may , however , also occur during the ontogenetic maturation of the individual organism for instance by the activity - dependent refinement of neuronal circuits .
if such an activity - dependent refinement of cortical architecture realizes an optimization strategy its outcome should be interpreted as the convergence towards a ground state of a specific energy functional .
this hypothesized optimized functional , however , remains currently unknown . as several different functional maps coexist in the visual cortex candidate energy functionals are expected to reflect the multiple response properties of neurons in the visual cortex .
in fact , consistent with the idea of joint optimization of different feature maps cortical maps are not independent of each other @xcite .
various studies proposed a coordinated optimization of different feature maps @xcite .
coordinated optimization appears consistent with the observed distinct spatial relationships between different maps such as the tendency of iso - orientation lines to intersect od borders perpendicularly or the preferential positioning of orientation pinwheels at locations of maximal eye dominance @xcite .
specifically these geometric correlations have thus been proposed to indicate the optimization of a cost function given by a compromise between stimulus coverage and continuity @xcite , a conclusion that was questioned by carreira - perpinan and goodhill @xcite .
+ visual cortical maps are often spatially complex patterns that contain defect structures such as point singularities ( pinwheels ) @xcite or line discontinuities ( fractures ) @xcite and that never exactly repeat @xcite .
it is conceivable that this spatial complexity arises from geometric frustration due to a coordinated optimization of multiple feature maps in which not all inter - map interactions can be simultaneously satisfied @xcite . in many optimization models , however , the resulting map layout is spatially not complex or lacks some of the basic features such as topological defects @xcite . in other studies coordinated optimization
was reported to preserve defects that would otherwise decay @xcite .
an attempt to rigorously study the hypothesis that the structure of cortical maps is explained by an optimization process thus raises a number of questions : i ) what are the genuine ground states of candidate energy functionals and how can they be calculated with precision and rigor ?
ii ) how do differences in candidate optimization principles impact on the predicted map structure and conversely what can be learned about an hypothetical underlying optimization principle from observations on map structure ?
iii ) is there a way to analyze the coordinated organization of cortical maps predicted by optimization principles in general ?
if theoretical neuroscience was able to answer these questions with greater confidence , the interpretation and explanation of visual cortical architecture could build on a more solid foundation than currently available .
to start laying such a foundation , we examined how symmetry principles in general constrain the form of optimization models and developed a formalism for analyzing map optimization independent of the specific energy functional assumed .
+ minima of a given energy functional can be found by gradient descent which is naturally represented by a dynamical system describing a formal time evolution of the maps .
response properties in visual cortical maps are arranged in repetitive modules of a typical spatial length called hypercolumn .
optimization models that reproduce this typical length scale are therefore effectively pattern forming systems with a so - called cellular or finite wavelength instability , see @xcite .
in the theory of pattern formation , it is well understood that symmetries play a crucial role @xcite .
some symmetries are widely considered biologically plausible for cortical maps , for instance the invariance under spatial translations and rotations or a global shift of orientation preference @xcite . in this paper
we argue that such symmetries and an approach that utilizes the analogy between map optimization and pattern forming systems can open up a novel and systematic approach to the coordinated optimization of visual cortical representations .
+ a recent study found strong evidence for a common design in the functional architecture of orientation columns @xcite .
three species , galagos , ferrets , and tree shrews , widely separated in evolution of modern mammals , share an apparently universal set of quantitative properties .
the average pinwheel density as well as the spatial organization of pinwheels within orientation hypercolumns , expressed in the statistics of nearest neighbors as well as the local variability of the pinwheel densities in cortical subregions ranging from 1 to 30 hypercolumns , are found to be virtually identical in the analyzed species .
however , these quantities are different from random maps .
intriguingly , the average pinwheel density was found to be statistical indistinguishable from the mathematical constant @xmath0 up to a precision of 2% .
such apparently universal laws can be reproduced in relatively simple self - organization models if long - range neuronal interactions are dominant @xcite . as pointed out by kaschube and coworkers ,
these findings pose strong constraints on models of cortical functional architecture @xcite .
many models exhibiting pinwheel annihilation @xcite or pinwheel crystallization @xcite were found to violate the experimentally observed layout rules . in @xcite
it was shown that the common design is correctly predicted in models that were based only on intrinsic op properties . alternatively , however , it is conceivable that they result from geometric frustration due to inter - map interactions and joint optimization . in the current study
we therefore in particular examined whether the coordinated optimization of the op map and another feature map can reproduce the quantitative laws defining the common design .
+ the presentation of our results is organized as follows .
first we introduce a formalism to model the coordinated optimization of complex and real valued scalar fields .
complex valued fields can represent for instance orientation preference ( op ) or direction preference maps @xcite .
real valued fields may represent for instance ocular dominance ( od ) @xcite , spatial frequency maps @xcite or on - off segregation @xcite .
we construct several optimization models such that an independent optimization of each map in isolation results in a regular op stripe pattern and , depending on the relative representations of the two eyes , od patterns with a regular hexagonal or stripe layout .
a model - free , symmetry - based analysis of potential optimization principles that couple the real and complex valued fields provides a comprehensive classification and parametrization of conceivable coordinated optimization models and identifies representative forms of coupling energies . for analytical treatment of the optimization problem
we adapt a perturbation method from pattern formation theory called weakly nonlinear analysis @xcite .
this method is applicable to models in which the spatial pattern of columns branches off continuously from an unselective homogeneous state . it reduces the dimensionality of the system and leads to amplitude equations as an approximate description of the system near the symmetry breaking transition at which the homogeneous state becomes unstable .
we identify a limit in which inter - map interactions that are formally always bidirectional become effectively unidirectional . in this limit
, one can neglect the backreaction of the complex map on the layout of the co - evolving scalar feature map .
we show how to treat low and higher order versions of inter - map coupling energies which enter at different order in the perturbative expansion .
+ second we apply the derived formalism by calculating optima of two representative low order examples of coordinated optimization models and examine how they impact on the resulting map layout .
two higher order optimization models are analyzed in text s1 . for concreteness and
motivated by recent topical interest @xcite , we illustrate the coordinated optimization of visual cortical maps for the widely studied example of a complex op map and a real feature map such as the od map .
op maps are characterized by pinwheels , regions in which columns preferring all possible orientations are organized around a common center in a radial fashion @xcite . in particular , we address the problem of pinwheel stability in op maps @xcite and calculate the pinwheel densities predicted by different models .
as shown previously , many theoretical models of visual cortical development and optimization fail to predict op maps possessing stable pinwheels @xcite .
we show that in case of the low order energies , a strong inter - map coupling will typically lead to op map suppression , causing the orientation selectivity of all neurons to vanish .
for all considered optimization models , we identify stationary solutions of the resulting dynamics and mathematically demonstrate their stability .
we further calculate phase diagrams as a function of the inter - map coupling strength and the amount of overrepresentation of certain stimuli of the co - evolving scalar feature map .
we show that the optimization of any of the analyzed coupling energies can lead to spatially relatively complex patterns
. moreover , in case of op maps , these patterns are typically pinwheel - rich .
the phase diagrams , however , differ for each considered coupling energy , in particular leading to coupling energy specific ground states .
we therefore thoroughly analyze the spatial layout of energetic ground states and in particular their geometric inter - map relationships .
we find that none of the examined models reproduces the experimentally observed pinwheel density and spatially aperiodic arrangements .
our analysis identifies a seemingly general condition for interaction induced pinwheel - rich op optima namely a substantial bias in the response properties of the co - evolving scalar feature map . in case of the low order inter - map coupling energies
strong inter - map coupling leads to a suppression of op selectivity .
this suppressive effect can be avoided by restricting coupling strengths .
one aim of this article is to test different optimization principles and potentially rule out some optimization principles .
when comparing our results from different optimization principles to biological data such parameter tuning reduces the practicability . in this supporting information we complement our study using the high order inter - map coupling energies .
we show that in this case a suppression of op selectivity can not occur .
we derive coupled amplitude equations which , however , involve several mathematical assumptions .
a systematic treatment as it is shown in the main article would imply that low order and higher order inter - map coupling energies are in general non - zero .
low order energy terms would enter at third order in the expansion and higher order corrections could potentially alter the stability properties .
in addition , higher order inter - map coupling energies can affect the stability of patterns .
in the following , we assume all low order inter - map coupling energies to be zero and that contributions entering the amplitude equations at higher orders can be neglected .
the obtained results are confirmed numerically in part ( ii ) of this study .
we model the response properties of neuronal populations in the visual cortex by two - dimensional scalar order parameter fields which are either complex valued or real valued @xcite .
a complex valued field @xmath1 can for instance describe op or direction preference of a neuron located at position @xmath2 .
a real valued field @xmath3 can describe for instance od or the spatial frequency preference .
although we introduce a model for the coordinated optimization of general real and complex valued order parameter fields we consider @xmath1 as the field of op and @xmath3 as the field of od throughout this article . in this case , the pattern of preferred stimulus orientation @xmath4 is obtained by @xmath5 the modulus @xmath6 is a measure of the selectivity at cortical location @xmath2 .
+ op maps are characterized by so - called _ pinwheels _ , regions in which columns preferring all possible orientations are organized around a common center in a radial fashion .
the centers of pinwheels are point discontinuities of the field @xmath7 where the mean orientation preference of nearby columns changes by 90 degrees .
pinwheels can be characterized by a topological charge @xmath8 which indicates in particular whether the orientation preference increases clockwise or counterclockwise around the pinwheel center , @xmath9 where @xmath10 is a closed curve around a single pinwheel center at @xmath11 . since @xmath4 is a cyclic variable in the interval
@xmath12 $ ] and up to isolated points is a continuous function of @xmath2 , @xmath13 can only have values @xmath14 where @xmath15 is an integer number @xcite .
if its absolute value @xmath16 , each orientation is represented only once in the vicinity of a pinwheel center . in experiments , only pinwheels with a topological charge of @xmath17 are observed , which are simple zeros of the field @xmath1 .
+ od maps can be described by a real valued two - dimensional field @xmath3 , where @xmath18 indicates ipsilateral eye dominance and @xmath19 contralateral eye dominance of the neuron located at position @xmath2 .
the magnitude indicates the strength of the eye dominance and thus the zeros of the field correspond to the borders of od .
+ in this article , we view visual cortical maps as optima of some energy functional @xmath20 .
the time evolution of these maps can be described by the gradient descent of this energy functional .
the field dynamics thus takes the form @xmath21 \nonumber \\ \partial_t \ , o(\mathbf{x},t ) & = & g[z(\mathbf{x},t),o(\mathbf{x},t ) ] , \end{aligned}\ ] ] where @xmath22 $ ] and @xmath23 $ ] are nonlinear operators given by @xmath22=-\frac{\delta e}{\delta \overline{z}}$ ] , @xmath23=-\frac{\delta e}{\delta o}$ ] .
the system then relaxes towards the minima of the energy @xmath20 .
the convergence of this dynamics towards an attractor is assumed to represent the process of maturation and optimization of the cortical circuitry .
various biologically detailed models have been cast to this form @xcite .
+ all visual cortical maps are arranged in repetitive patterns of a typical wavelength @xmath24 .
we splitted the energy functional @xmath20 into a part that ensures the emergence of such a typical wavelength for each map and into a part which describes the coupling among different maps .
a well studied model reproducing the emergence of a typical wavelength by a pattern forming instability is of the swift - hohenberg type @xcite .
many other pattern forming systems occurring in different physical , chemical , and biological contexts ( see for instance @xcite ) have been cast into a dynamics of this type .
its dynamics in case of the op map is of the form @xmath25 with the linear swift - hohenberg operator @xmath26 @xmath27 , and @xmath28 $ ] a nonlinear operator .
the energy functional of this dynamics is given by @xmath29 in fourier representation , @xmath30 is diagonal with the spectrum @xmath31 the spectrum exhibits a maximum at @xmath32.for @xmath33 , all modes are damped since @xmath34 and only the homogeneous state @xmath35 is stable .
this is no longer the case for @xmath36 when modes on the _ critical circle _ @xmath32 acquire a positive growth rate and now start to grow , resulting in patterns with a typical wavelength @xmath37 .
thus , this model exhibits a supercritical bifurcation where the homogeneous state looses its stability and spatial modulations start to grow .
+ the coupled dynamics we considered is of the form @xmath38 where @xmath39 , and @xmath40 is a constant . to account for the species differences in the wavelengths of the pattern we chose two typical wavelengths @xmath41 and @xmath42 .
the dynamics of @xmath43 and @xmath44 is coupled by interaction terms which can be derived from a coupling energy @xmath45 . in the uncoupled case
this dynamics leads to pinwheel free op stripe patterns .
how many inter - map coupling energies @xmath45 exist ?
using a phenomenological approach the inclusion and exclusion of various terms has to be strictly justified .
we did this by symmetry considerations .
the constant @xmath40 breaks the inversion symmetry @xmath46 of inputs from the ipsilateral ( @xmath18 ) or contralateral ( @xmath19 ) eye .
such an inversion symmetry breaking could also arise from quadratic terms such as @xmath47 . in the methods section
we detail how a constant shift in the field @xmath3 can eliminate the constant term and generate such a quadratic term . including either a shift or a quadratic term thus already represents the most general case .
the inter - map coupling energy @xmath45 was assumed to be invariant under this inversion .
otherwise orientation selective neurons would , for an equal representation of the two eyes , develop different layouts to inputs from the left or the right eye .
the primary visual cortex shows no anatomical indication that there are any prominent regions or directions parallel to the cortical layers @xcite . besides invariance under translations @xmath48 and rotations @xmath49 of both maps we required that the dynamics should be invariant under orientation shifts @xmath50
note , that the assumption of shift symmetry is an idealization that uncouples the op map from the map of visual space .
bressloff and coworkers have presented arguments that euclidean symmetry that couples spatial locations to orientation shift represents a more plausible symmetry for visual cortical dynamics @xcite , see also @xcite . the existence of orientation shift symmetry , however , is not an all or none question .
recent evidence in fact indicates that shift symmetry is only weakly broken in the spatial organization of orientation maps @xcite .
a general coupling energy term can be expressed by integral operators which can be written as a volterra series @xmath51 with an @xmath52-th .
order integral kernel @xmath53 .
inversion symmetry and orientation shift symmetry require @xmath54 to be even and that the number of fields @xmath55 equals the number of fields @xmath56 .
the lowest order term , mediating an interaction between the fields @xmath57 and @xmath55 is given by @xmath58 i.e. @xmath59 next , we rewrite eq .
( [ eq : volterra4 ] ) as an integral over an energy density @xmath45 .
we use the invariance under translations to introduce new coordinates @xmath60 this leads to @xmath61 the kernel @xmath62 may contain local and non - local contributions .
map interactions were assumed to be local . for local interactions
the integral kernel is independent of the locations @xmath63 .
we expanded both fields in a taylor series around @xmath64 @xmath65 for a local energy density we could truncate this expansion at the first order in the derivatives .
the energy density can thus be written @xmath66 due to rotation symmetry this energy density should be invariant under a simultaneous rotation of both fields . from all possible combinations of eq .
( [ eq : t4fields_part3a ] ) only those are invariant in which the gradients of the fields appear as scalar products . the energy density can thus be written as @xmath67 where we suppress the argument @xmath64 .
all combinations @xmath68 can also enter via their complex conjugate .
the general expression for @xmath69 is therefore @xmath70 from all possible combinations we selected those which are invariant under orientation shifts and eye inversions .
this leads to @xmath71 the energy densities with prefactor @xmath72 to @xmath73 do not mediate a coupling between od and op fields and can be absorbed into the single field energy functionals .
the densities with prefactors @xmath74 and @xmath75 ( also with @xmath76 and @xmath77 ) are complex and can occur only together with @xmath78 ( @xmath79 ) to be real . these energy densities , however , are not bounded from below as their real and imaginary parts can have arbitrary positive and negative values .
the lowest order terms which are real and positive definite are thus given by @xmath80 the next higher order energy terms are given by @xmath81 here the fields @xmath3 and @xmath1 enter with an unequal power .
in the corresponding field equations these interaction terms enter either in the linear part or in the cubic nonlinearity .
we will show in this article that interaction terms that enter in the linear part of the dynamics can lead to a suppression of the pattern and possibly to an instability of the pattern solution .
therefore we considered also higher order interaction terms .
+ these higher order terms contain combinations of terms in eq .
( [ eq : t4 ] ) and are given by @xmath82 as we will show below examples of coupling energies @xmath83 form a representative set that can be expected to reproduce experimentally observed map relationships . for this choice of energy the corresponding interaction terms in the dynamics eq .
( [ eq : dynamicsshcoupled_ref ] ) are given by @xmath84+n_\beta[o , o , z]+n_\epsilon[o , o , o , o , z , z,\overline{z}]+n_\tau[o , o , o , o , z , z,\overline{z}]\nonumber \\ & = & -\alpha o^2z+\beta\nabla \left(a\nabla o\right ) + \epsilon \
, 2\nabla \left(|a|^2a\nabla o\right ) -2\tau \,o^4|z|^2z , \nonumber \\ -\frac{\delta u}{\delta o}&= & \widetilde{n}_\alpha[o , z,\overline{z}]+\widetilde{n}_\beta[o , z,\overline{z}]+ \widetilde{n}_\epsilon[o , o , o , z , z,\overline{z},\overline{z}]+ \widetilde{n}_\tau[o , o , o , z , z,\overline{z},\overline{z}]\nonumber \\ & = & -\alpha o|z|^2+\beta \nabla \left ( \overline{a } \nabla z \right ) + \epsilon \ , 2 \nabla \left(|a|^2\overline{a } \nabla z \right)-2\tau\ , o^3|z|^4+c.c . \end{aligned}\ ] ] with @xmath85 and @xmath86 denoting the complex conjugate . in general , all coupling energies in @xmath87 , and @xmath88 can occur in the dynamics and we restrict to those energies which are expected to reproduce the observed geometric relationships between op and od maps .
it is important to note that with this restriction we did not miss any essential parts of the model .
when using weakly nonlinear analysis the general form of the near threshold dynamics is insensitive to the used type of coupling energy and we therefore expect similar results also for the remaining coupling energies .
+ numerical simulations of the dynamics eq .
( [ eq : dynamicsshcoupled_ref ] ) , see @xcite , with the coupling energy eq .
( [ eq : energy_all ] ) and @xmath89 are shown in fig .
[ fig : numerics ] . .
color code of op map with zero contours of od map superimposed . * * @xmath90 * * @xmath91 * * and * * @xmath92 .
initial conditions identical in * * - * * , @xmath93 . ]
the initial conditions and final states are shown for different bias terms @xmath40 and inter - map coupling strengths @xmath94 .
we observed that for a substantial contralateral bias and above a critical inter - map coupling pinwheels are preserved from random initial conditions or are generated if the initial condition is pinwheel free . without a contralateral bias
the final states were pinwheel free stripe solutions irrespective of the strength of the inter - map coupling .
we studied eq .
( [ eq : dynamicsshcoupled_ref ] ) with the low order inter - map coupling energies in eq .
( [ eq : energy_all ] ) using weakly nonlinear analysis .
we therefore rewrite eq .
( [ eq : dynamicsshcoupled_ref ] ) as @xmath95 -n_{3,c}[z , o , o ] \nonumber \\ \partial_t \ , o(\mathbf{x},t ) & = & r_o o(\mathbf{x},t ) -\hat{l}_o^0 \ , o(\mathbf{x},t ) + n_{2,u}[o , o]-n_{3,u}[o , o , o ] -\tilde{n}_{3,c}[o , z,\overline{z}]\ , , \end{aligned}\ ] ] where we shifted both linear operators as @xmath96 , @xmath97 . the constant term @xmath40 in eq .
( [ eq : dynamicsshcoupled_ref ] ) is replaced by a quadratic interaction term @xmath98=\tilde{\gamma } o^2 $ ] with @xmath99 , see methods .
the uncoupled nonlinearities are given by @xmath100 , @xmath101 while @xmath102 and @xmath103 are the nonlinearities of the low order inter - map coupling energy eq .
( [ eq : interactionterm ] ) .
we study eq .
( [ eq : fieldeq_foramplderiv ] ) close to the pattern forming bifurcation where @xmath104 and @xmath105 are small .
we therefore expand both control parameters in powers of the small expansion parameter @xmath106 @xmath107 close to the bifurcation the fileds are small and thus nonlinearities are weak . we therefore expand both fields as @xmath108 we further introduced a common slow timescale @xmath109 and insert the expansions in eq .
( [ eq : fieldeq_foramplderiv ] ) and get @xmath110\right ) \nonumber \\ & & + \mu^3\left(- n_{3,c}[z_1,o_1,o_1]\right ) \nonumber \\ & & \vdots \end{aligned}\ ] ] and @xmath111 \right)\nonumber\\ & & + \mu^3\left(- r_{z2 } \partial_t o_1+r_{o2 } o_1+r_{o1 } o_2- r_{z1 } \partial_t
o_2-\hat{l}^0o_3-\tilde{n}_{3,u}[o_1,o_1,o_1]\right ) \nonumber \\ & & + \mu^3\left(- \tilde{n}_{3,c}[o_1,z_1,\overline{z}_1]\right ) \nonumber \\ & & \vdots \end{aligned}\ ] ] we consider amplitude equations up to third order as this is the order where the nonlinearity of the low order inter - map coupling energy enters first . for eq .
( [ eq : expansion_z_coupled ] ) and eq .
( [ eq : expansion_o_coupled ] ) to be fulfilled each individual order in @xmath106 has to be zero . at linear order in @xmath106
we get the two homogeneous equations @xmath112 thus @xmath113 and @xmath114 are elements of the kernel of @xmath115 and @xmath116 .
both kernels contain linear combinations of modes with a wavevector on the corresponding critical circle i.e. @xmath117 with the complex amplitudes @xmath118 , @xmath119 and @xmath120 , @xmath121 . in view of the hexagonal or stripe layout of the od pattern shown in fig . [ fig : numerics ] , @xmath122 is an appropriate choice . since in cat visual cortex the typical wavelength for od and op maps are approximately the same @xcite i.e. @xmath123 the fourier components of the emerging pattern are located on a common critical circle . to account for species differences we also analyzed models with detuned op and od wavelengths in part ( ii ) of this study .
+ at second order in @xmath106 we get @xmath124 as @xmath113 and @xmath114 are elements of the kernel @xmath125 . at third order , when applying the solvability condition ( see methods ) , we get @xmath126 - \hat{p}_c n_{3,c}[z_1,o_1,o_1 ] \\ r_{z2 } \partial_t \ , o_1&= & r_{o2 } o_1-\sqrt{r_{o2 } } \ , \hat{p}_c
\tilde{n}_{2,u}[o_1,o_1]- \hat{p}_c \tilde{n}_{3,u}[o_1,o_1,o_1 ] - \hat{p}_c \tilde{n}_{3,c}[o_1,z_1,\overline{z}_1]\ , .
\nonumber\end{aligned}\ ] ] we insert the leading order fields eq .
( [ eq_z1o1 ] ) and obtain the amplitude equations @xmath127 for simplicity we have written only the simplest inter - map coupling terms . depending on the configuration of active modes additional contributions may enter the amplitude equations .
in addition , for the product - type coupling energy , there are coupling terms which contain the constant @xmath128 , see methods and eq .
( [ eq : potential_cubic_part3 ] ) .
the coupling coefficients are given by @xmath129 + n_{3,u}[e^{\imath \vec{k}_j \cdot \vec{x}},e^{\imath \vec{k}_i \cdot \vec{x } } , e^{-\imath \vec{k}_j \cdot \vec{x } } ] \right)=2 \nonumber \\ g_{ii}= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } \ , n_{3,u}[e^{\imath \vec{k}_i \cdot \vec{x } } , e^{\imath \vec{k}_i \cdot \vec{x } } , e^{-\imath \vec{k}_i \cdot \vec{x } } ] = 1\nonumber \\ g_{ij^-}= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } \left ( n_{3,u}[e^{\imath \vec{k}_i \cdot \vec{x } } , e^{-\imath \vec{k}_j \cdot \vec{x}},e^{\imath \vec{k}_j \cdot \vec{x } } ] + n_{3,u}[e^{-\imath \vec{k}_j \cdot \vec{x}},e^{\imath \vec{k}_i \cdot \vec{x } } , e^{\imath \vec{k}_j \cdot \vec{x } } ] \right)=2 \nonumber \\ f_{ij}= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } \left ( n_{3,u}[e^{\imath \vec{k}_j \cdot \vec{x } } , e^{-\imath \vec{k}_j \cdot \vec{x}},e^{\imath \vec{k}_i \cdot \vec{x}}]+ n_{3,u}[e^{-\imath \vec{k}_j \cdot \vec{x } } , e^{\imath \vec{k}_j \cdot \vec{x}},e^{\imath \vec{k}_i \cdot \vec{x } } ] \right)=2 \nonumber \\
f_{ii}=&\ , 0 \nonumber \\ h_{ij}= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } n_{3,c}[e^{\imath \vec{k}_i \cdot \vec{x}},e^{\imath \vec{k}_j \cdot \vec{x } } , e^{-\imath \vec{k}_j \cdot \vec{x } } ] \nonumber \\ h_{ii}= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } \ , n_{3,c}[e^{\imath \vec{k}_i \cdot \vec{x } } , e^{\imath \vec{k}_i \cdot \vec{x } } , e^{-\imath \vec{k}_i \cdot \vec{x } } ] \nonumber \\ h= & \ , e^{-\imath \vec{k}_i \cdot \vec{x } } \ , n_{3,c}[e^{-\imath \vec{k}_i \cdot \vec{x } } , e^{\imath \vec{k}_i \cdot \vec{x } } , e^{\imath \vec{k}_i \cdot \vec{x } } ] \ , . \ ] ] from eq .
( [ eq : withkappa ] ) we see that inter - map coupling has two effects on the modes of the op pattern .
first , inter - map coupling shifts the bifurcation point from @xmath104 to @xmath130 .
this can cause a potential destabilization of pattern solutions for large inter - map coupling strength .
second , inter - map coupling introduces additional resonant interactions that for instance couple the modes @xmath131 and their opposite modes @xmath132 . in case of @xmath133
the inter - map coupling terms in dynamics of the modes @xmath134 are small . in this limit
the dynamics of the modes @xmath134 decouples from the modes @xmath135 and we can use the uncoupled od dynamics , see methods .
when we scale back to the fast time variable and set @xmath136 , @xmath137 we obtain @xmath138 the amplitude equations are truncated at third order .
if pattern formation takes place far from threshold fifth order , seventh order , or even higher order corrections have to be considered .
moreover , at seventh order the nonlinearities of the higher order inter - map coupling terms enter the expansion .
a derivation of amplitude equation with higher order inter - map coupling energies is presented in text s1 .
using symmetry considerations we derived inter - map coupling energies up to eighth order in the fields , see eq .
( [ eq : t4 ] ) , eq.([eq : t6 ] ) , and eq.([eq : t8 ] ) .
which of these various optimization principles could reproduce realistic inter - map relationships such as a uniform coverage of all stimulus features ? we identified two types of optimization principles that can be expected to reproduce realistic inter - map relationships and good stimulus coverage .
first , product - type coupling energies of the form @xmath139 .
these energies favor configurations in which regions of high gradients avoid each other and thus leading to high coverage .
second , gradient - type coupling energies of the form @xmath140 . in experimentally obtained maps ,
iso - orientation lines show the tendency to intersect the od borders perpendicularly .
perpendicular intersection angles lead to high coverage as large changes of the field @xmath55 in one direction lead to small changes of the field @xmath57 in that direction . to see that the gradient - type coupling energy favors perpendicular intersection angles we decompose the complex field @xmath1 into the selectivity @xmath141 and the preferred orientation @xmath4 .
we obtain @xmath142 if the orientation selectivity is locally homogeneous , i.e. @xmath143 , then the energy is minimized if the direction of the iso - orientation lines ( @xmath144 ) is perpendicular to the od borders . in our symmetry - based analysis
we further identified terms that are expected to lead to the opposite behavior for instance mixture terms such as @xmath145 . + pinwheels are prominent features in op maps .
we therefore also analyze how different optimization principles impact on the pinwheel positions with respect to the co - evolving feature maps .
the product - type coupling energies are expected to favor pinwheels at od extrema .
pinwheels are zeros of @xmath55 and are thus expected to reduce this energy term .
they will reduce energy the most when @xmath146 is maximal which should repel pinwheels from od borders , where @xmath3 is zero .
also the gradient - type coupling energy is expected to couple the od pattern with the position of pinwheels . to see this we decompose the field @xmath55 into its real and imaginary part @xmath147 at pinwheel centers the zero contours of @xmath148 and @xmath149 cross .
since there @xmath150 and @xmath151 are almost constant and not parallel the energy can be minimized only if @xmath152 is small at the pinwheel centers , i.e. the extrema or saddle - points of @xmath3 .
+ from the previous considerations we assume all coupling coefficients of the energies to be positive .
a negative coupling coefficient can be saturated by higher order inter - map coupling terms . in the following ,
we discuss the impact of the low order inter - map coupling energies on the resulting optima of the system using the derived amplitude equations .
the corresponding analysis for higher order inter - map coupling energies is provided in text s1 . as indicated by numerical simulations and weakly nonlinear analysis of the uncoupled od dynamics
, see methods , we discussed the influence of the od stripe , hexagon , and constant solutions on the op map using the coupled amplitude equations derived in the previous section .
a potential backreaction onto the dynamics of the od map can be neglected if the modes @xmath153 of the op map are much smaller than the modes @xmath154 of the od map .
this can be achieved if @xmath155 .
we first give a brief description of the uncoupled op solutions .
next , we study the impact of the low order coupling energies in eq .
( [ eq : energy_all ] ) on these solutions .
we demonstrate that these energies can lead to a complete suppression of orientation selectivity . in the uncoupled case
there are for @xmath156 two stable stationary solutions to the amplitude equations eq .
( [ eq : ampleq_coupled_thirdorder ] ) , namely op stripes @xmath157 and op rhombic solutions @xmath158 with @xmath159 , @xmath160 , @xmath161 an arbitrary phase , and @xmath162 . in the uncoupled case
the angle @xmath163 between the fourier modes is arbitrary .
the stripe solutions are pinwheel free while the pinwheel density for the rhombic solutions varies as @xmath164 and thus @xmath165 . for the rhombic solutions pinwheels
are located on a regular lattice .
we therefore refer to these and other pinwheel rich solutions which are spatially periodic as pinwheel crystals ( pwc ) .
in particular , we refer to pinwheel crystals with as rhombic spatial layout as rpwc solutions and pinwheel crystals with a hexagonal layout as hpwc solutions . without inter - map coupling ,
the potential of the two solutions reads @xmath166 , thus the stripe solutions are always energetically preferred compared to rhombic solutions .
+ in the following we study three scenarios in which inter - map coupling can lead to pinwheel stabilization .
first , a deformation of the op stripe solution can lead to the creation of pinwheels in this solution .
second , inter - map coupling can energetically prefer the ( deformed ) op rhombic solutions compared to the stripe solutions .
finally , inter - map coupling can lead to the stabilization of new pwc solutions .
+ for the low order interaction terms the amplitude equations are given by @xmath167 , @xmath168 with the potential @xmath169 with the uncoupled contributions @xmath170 the coupling coefficients read @xmath171 , @xmath172 , @xmath173 , @xmath174 , where @xmath175 is the angle between the wavevector @xmath176 and @xmath177 .
we first studied the impact of the low order product - type coupling energy . here , the constant @xmath179 enters explicitly in the amplitude equations , see eq .
( [ eq : potential_cubic_part3 ] ) and eq .
( [ eq : delta ] ) . in the case of od stripes , see methods , with @xmath180 we get the following amplitude equations @xmath181 where nct
. refers to non inter - map coupling terms @xmath182 , resulting from the potential @xmath183 , see eq .
( [ eq : potential_uncoupled ] ) . the equations for the modes @xmath132
are given by interchanging the modes @xmath131 and @xmath132 as well as interchanging the modes @xmath184 and @xmath185 .
the op stripe solution in case of inter - map coupling is given by @xmath186 with @xmath187 , @xmath188 , and + @xmath189 and the phase relation + .
in the uncoupled case ( @xmath190 ) they reduce to @xmath191 and @xmath192 . with increasing inter - map coupling the amplitude @xmath193 grows and the solutions are transformed , reducing the representation of all but two preferred orientations .
the parameter dependence of this solution is shown in fig .
[ fig : amplitudes_alpha ] * * for different values of the bias @xmath40 .
both amplitudes become identical at @xmath194 with @xmath195 this pattern solution finally vanishes at @xmath196 this existence border is thus independent of the od bias @xmath40 . above this coupling strength
only the trivial solution @xmath197 is stable .
+ in addition to the op stripe patterns there exist rhombic op solutions , see fig . [
fig : amplitudes_alpha]**. these rhombic solutions are pinwheel rich with a pinwheel density of @xmath198 but are energetically not preferred compared to the stripe solutions , see fig .
[ fig : amplitudes_alpha]**. the rhombic solutions in the uncoupled case @xmath199 , @xmath200 are transformed by inter - map coupling .
the phase relations are given by @xmath201 where @xmath161 is an arbitrary phase .
stationary amplitudes are given by @xmath202 with @xmath203 .
with increasing inter - map coupling strength @xmath204 the amplitudes @xmath205 are suppressed , see fig .
( [ fig : amplitudes_alpha])**. in addition , for nonzero bias @xmath40 , there is an increase of the amplitudes @xmath206 . the amplitudes @xmath207 and @xmath208 collapse at @xmath209
. a further increase of the inter - map coupling strength leads to a suppression of these amplitudes and finally to the op stripe pattern where @xmath210 .
+ in the case the od map is a constant , eq .
( [ eq : odconst ] ) , the amplitude equations simplify to @xmath211 thus inter - map coupling in this case only renormalizes the bifurcation parameter and the energetic ground state is thus a stripe pattern with an inter - map coupling dependent reduction of the amplitudes @xmath212 therefore at @xmath213 the stripe pattern ceases to exist and the only stable solution is the trivial one i.e. @xmath214 .
in addition , there is the rhombic solution with the stationary amplitudes @xmath215 in the case of od hexagons @xmath216 , eq .
( [ eq : hex ] ) , the amplitude equations read @xmath217 where the indices are cyclic i.e. @xmath218 .
these amplitude equations have stripe - like solutions as well as solutions with a rhombic layout of the form @xmath199 , @xmath219 . for both solutions
the stationary phases depend on the inter - map coupling strength @xmath204 . in this case
stationary solutions of eq .
( [ eq : amp_alp_odhex ] ) are calculated numerically using a newton method and initial conditions close to these solutions .
in contrast to the case of od stripes and od constant solutions the amplitude equations ( [ eq : amp_alp_odhex ] ) have an additional type of pwc solution which have uniform amplitudes , @xmath220 .
the dynamics of their phases is given by @xmath221 when solving the amplitude equations numerically we observe that the phase relations vary with the inter - map coupling strength for non - uniform solutions .
but for the uniform solution the phase relations are independent of the inter - map coupling strength .
the phases of the uniform solution are determined up to a free phase @xmath222 which results from the orientation shift symmetry @xmath223 of eq .
( [ eq : dynamicsshcoupled_ref ] ) .
we therefore choose @xmath224 . as an ansatz for the uniform solutions we use @xmath225 where @xmath226 is the kronecker delta and @xmath227 a constant parameter .
note , that @xmath1 can not become real since @xmath228 .
the equation for the uniform amplitudes is then given by @xmath229 while the phase dynamics reads @xmath230 the stationarity condition is fulfilled for an arbitrary @xmath128 only if @xmath231 or @xmath232 .
the corresponding amplitudes are given by solving the stationarity condition for the real part and read @xmath233 we calculate the stability properties of all solutions by linear stability analysis considering perturbations of the amplitudes @xmath234 , @xmath235 and of the phases @xmath236 , @xmath237 .
this leads to a perturbation matrix @xmath238 .
amplitude and phase perturbations in general do not decouple .
we calculated the eigenvalues of the perturbation @xmath238 matrix numerically and checked the results by direct numerical simulation of the amplitude equations . in case of the uniform solutions eq .
( [ eq : uniformsol_part3 ] ) the perturbation matrix @xmath238 is explicitly stated in text s3 .
+ the stability of the @xmath239and @xmath232 uniform solutions depends on the coupling strength @xmath204 and on the sign of @xmath240 . as the solution of @xmath241 is given by @xmath242 in the stability range of od hexagons there is only one possible stable uniform solution , the @xmath232 uniform solution .
this solution ceases to exist at @xmath243 .
this existence border is in fact independent of the bias @xmath40 and given by @xmath244 thus the limit @xmath245 makes the uniform solution unstable for smaller and smaller coupling strengths . the case of op solutions when interacting with od stripes is shown in fig .
[ fig : amplitudes_alpha]**. in case of op stripes inter - map coupling suppresses the amplitude @xmath246 of the stripe pattern while increasing the amplitude of the opposite mode @xmath247 .
this transformation reduces the representation of all but two preferred orientations .
when both amplitudes collapse the resulting op map is selective only to two orthogonal orientations namely @xmath248 and @xmath249 .
we refer to these unrealistic solutions as _ orientation scotoma solutions_. the phase relations ensure that od borders that run parallel to the op stripes are located at the op maxima and minima i.e. in the center of the orientation scotoma stripes .
with increasing inter - map coupling , this orientation scotoma pattern is suppressed until finally all amplitudes are zero and only the homogeneous solution is stable . in case of op rhombs inter - map coupling makes the rhombic pattern more stripe - like by reducing the amplitude @xmath205 . the mode @xmath206 which is zero in the uncoupled case increases and finally collapses with the mode @xmath207 .
increasing inter - map coupling more suppresses all but the two modes @xmath250 , leading again to the orientation scotoma stripe pattern . + the parameter dependence of op solutions when interacting with od hexagons is shown in fig .
[ fig : amplitudes_alpha]**. op stripe solutions became above a critical inter - map coupling strength unstable against pwc solutions .
this critical coupling strength strongly depended on the od bias .
op rhombic solutions also became unstable against pwc but for a lower coupling strength than the op stripes .
thus there is at intermediate coupling strength a bistability between stripe - like solutions and pwc solutions .
the potential of the op stripe and op rhombic solutions is shown in fig .
[ fig : amplitudes_alpha]**. stripes are energetically preferred in the uncoupled case as well as for small inter - map coupling strength for which they are stable .
+ to summarize , stripe solutions were deformed but no pinwheels were created for this solution .
the rhombic solutions were energetically not preferred for low inter - map coupling whereas for intermediate inter - map coupling these solutions lose pinwheels and became stripe solutions . instead , additional pinwheel rich solutions with a crystal layout became stable for intermediate inter - map coupling . for large inter - map coupling orientation selectivity was completely suppressed . , @xmath251 .
solid ( dashed ) lines : stable ( unstable ) solutions to eq .
( [ eq : amp_alp_odstripes ] ) ( od stripes ) and eq .
( [ eq : amp_alp_odhex ] ) ( od hexagons ) . * * od stripes , ( blue ) , @xmath242 ( green ) , @xmath252 ( orange ) . * * od hexagons , @xmath252 ( blue ) , @xmath253 ( red ) . * * transition from op stripe solutions , * * transition from op rhombic solutions .
* * potential for op stripes ( red ) and op rhombs ( blue ) interacting with od stripes , @xmath254 .
* * potential for op stripes and op rhombs interacting with od hexagons .
arrows indicate corresponding lines in the phase diagram , fig .
( [ fig : pd_alpha ] ) . ]
the phase diagram as a function of the od bias @xmath40 and the inter - map coupling strength @xmath204 for this coupling energy is shown in fig . [
fig : pd_alpha ] . , @xmath255 .
vertical lines : stability range of od stripes , hexagons , and constant solution .
magenta ( orange ) line : transition of stripes ( rhombs ) to the orientation scotoma solution .
blue line : stability border for the @xmath232 uniform solution ( hpwc ) .
green line : stability line of stripe - like solutions .
red line : pattern solutions ceases to exist , see eq .
( [ eq : alpha_crit_stripes ] ) and eq .
( [ eq : alpha_crit_hex ] ) .
blue region : stability region of hpwc , gray region : no pattern solution exists . ]
when rescaling the inter - map coupling strength as @xmath256 the phase diagram is independent of the bifurcation parameter of the op map @xmath104 .
thus for fixed @xmath257 the phase diagram depends only on two control parameters @xmath258 and @xmath256 .
the phase diagram contains the stability borders of the uncoupled od solutions @xmath259 .
they correspond to vertical lines , as they are independent of the inter - map coupling in the limit @xmath155 . at @xmath242 hexagons become stable .
stripe solutions become unstable at @xmath260 . at @xmath261
the homogeneous solution becomes stable while at @xmath262 hexagons lose their stability . in the units @xmath258 the borders @xmath263 vary slightly with @xmath105 , see fig .
[ fig : constterm ] , and are drawn here for @xmath251 .
colored lines correspond to the stability and existence borders of op solutions . in the region of stable od stripes
the op stripes run parallel to the od stripes . with increasing inter - map coupling strength the orientation preference of all but two orthogonal orientations
is suppressed . in the region of stable od hexagons stripe - like
op solutions dominate for low inter - map coupling strength . above a critical bias dependent coupling strength
the @xmath232 uniform solution becomes stable ( blue line ) .
there is a region of bistability between stripe - like and uniform solutions until the stripe - like solutions lose their stability ( orange line ) .
op rhombic solutions lose their stability when the uniform solution becomes stable .
thus there is no bistability between op rhombs and op uniform solutions . as in the case of od stripes
the uniform solution becomes unstable at @xmath264 .
also in the case of od hexagons the inter - map coupling leads to a transition towards the trivial solution where there is no op pattern at all . in case of the od constant solution
the op map is a stripe solution .
pinwheel rich solutions thus occur only in the region of stable od hexagons . in the following
we discuss the properties of these solutions .
the uniform solution eq .
( [ eq : uniformsol_part3 ] ) with @xmath232 is illustrated in fig .
[ fig : unif_d = pi_full ] . .
dashed lines mark the unit cell of the regular pattern . * * distribution of preferred orientations . ] for all stationary solutions the positions of the pinwheels are fixed by the od map and there are no translational degrees of freedom .
the unit cell ( dashed line ) contains 6 pinwheels which leads to a pinwheel density of @xmath265 .
two of them are located at od maxima ( contra center ) while one is located at an od minimum ( ipsi center ) .
the remaining three pinwheels are located at od saddle - points .
therefore , all pinwheels are located where the gradient of the od map is zero .
the pinwheel in the center of the op hexagon is at the ipsilateral od peak .
because these pinwheels organize most of the map while the others essentially only match one op hexagon to its neighbors we refer to this pinwheel crystal as the _ ipsi - center pinwheel crystal_. the iso - orientation lines intersect the od borders ( gray ) exactly with a right angle .
the intersection angles are , within the stability range of od hexagons , independent of the bias @xmath40 .
the remarkable property of perfect intersection angles can not be deduced directly from the coupling energy term .
the solution is symmetric under a combined rotation by @xmath266 and an orientation shift by @xmath267 .
the symmetry of the pattern is reflected by the distribution of preferred orientations , see fig .
[ fig : unif_d = pi_full]**. although the pattern is selective to all orientations the six orientations @xmath268 , @xmath269 are slightly overrepresented .
+ to summarize , the low order product - type inter - map coupling leads in case of od hexagons to a transition from pinwheel free stripe solutions towards pinwheel crystals .
the design of the pwc is an example of an orientation hypercolumn dominated by one pinwheel . with increasing inter - map coupling the pwc solution is suppressed until only the homogeneous solution is stable . in case of od stripes or the constant solution
the op solutions are pinwheel free stripe pattern .
when using a gradient - type inter - map coupling energy the interaction terms are independent of the od shift @xmath128 . in this case , the coupling strength can be rescaled as @xmath271 and is therefore independent of the bias @xmath40 .
the bias in this case only determines the stability of od stripes , hexagons or the constant solution .
a coupling to od stripes is easy to analyze in the case of a gradient - type inter - map coupling .
the energetically preferred solutions are op stripes with the direction perpendicular to the od stripes for which @xmath272 .
this configuration corresponds to the hubel and wiesel ice - cube model @xcite .
in addition there are rpwc solutions with the stationary amplitudes @xmath273 , @xmath274 , @xmath200 , and the stationary phases as in eq .
( [ eq : phases_alp_odstripes_oprhombs ] ) .
increasing inter - map coupling strength @xmath94 leads to an increase of the amplitudes @xmath275 while decreasing the amplitudes @xmath276 thus making the rhombic solution more stripe - like .
+ in the case the od map is a constant , eq .
( [ eq : odconst ] ) , the gradient - type inter - map coupling leaves the op dynamics unaffected .
the stationary states are therefore op stripes with an arbitrary direction and rpwc solutions as in the uncoupled case .
+ in the case of od hexagons the amplitude equations read @xmath277 using @xmath278 we obtain the phase equations @xmath279 these amplitude equations have stripe - like and rhombic solutions with inter - map coupling dependent phase relations .
we therefore calculate their stationary phases and amplitudes numerically using a newton method and initial conditions close to these solutions . besides stripe - like and rhombic solutions these amplitude equations have uniform solutions .
again we find that the ansatz eq .
( [ eq : uniformsol_part3 ] ) can satisfy the stationarity condition .
the phase dynamics in this case reads @xmath280 as in the case of the product - type inter - map coupling energy stationary solutions are @xmath231 and @xmath232 with the stationary amplitudes @xmath281 we studied the stability properties of both stationary solutions by linear stability analysis where amplitude and phase perturbations in general do not decouple .
the stability matrix of the uniform solutions is given in text s3 .
the eigenvalues are calculated numerically .
it turned out that the @xmath232 solution is unstable for @xmath282 while the @xmath231 solution becomes stable for @xmath283 .
the @xmath231 solution loses its stability above @xmath284 from thereon only the homogeneous solution @xmath285 is stable .
the course of the stationary amplitudes when interacting with od hexagons is shown in fig .
[ fig : amplitudes_beta ] . and od hexagons
. solid ( dashed ) lines : stable ( unstable ) solutions to eq .
( [ eq : amp_bet_odhex ] ) .
transition from op stripes towards the uniform solution ( red ) , transition from op rhombs towards the uniform solution ( blue ) .
arrows indicate corresponding lines in the phase diagram , fig .
( [ fig : pd_beta ] ) . ]
the op rhombic solution is almost unchanged by inter - map coupling but above a critical coupling strength the rhombs decay into a stripe - like solution .
the amplitude of the op stripe solution is suppressed by inter - map coupling and finally becomes unstable against the @xmath231 uniform solution .
thus for large inter - map coupling only the uniform solution is stable .
a further increase in the inter - map coupling suppresses the amplitude of this uniform solution until finally only the homogeneous solution is stable .
the phase diagram of this coupling energy is shown in fig . [
fig : pd_beta ] . , @xmath155 . vertical black lines : stability range of od stripes , hexagons , and constant solution .
blue line : stability border for the @xmath231 uniform solution .
green line : rhombic solutions become unstable .
orange line : stripe - like solutions become unstable .
red line : pattern solutions cease to exist , see eq .
( [ eq : unif_d_beta ] ) .
gray region : no pattern solution exists . ] we rescaled the inter - map coupling strength as @xmath286 , where @xmath287 is the stationary amplitude of the od hexagons .
the stability borders are then independent of the od bias in the od solutions and further independent of the bifurcation parameter @xmath104 .
this simplifies the analysis since the op solutions and their stability depend on @xmath40 only indirect via the amplitudes @xmath134 . in case of od stripes or od constant solution
there is no pinwheel crystallization .
instead the op solutions are pinwheel free stripes . in case of od hexagons
hpwcs become stable above @xmath283 ( blue line ) .
rhombic op patterns become unstable at @xmath288 and decay into a stripe - like solution ( green line ) . at @xmath289
these stripe - like solutions become unstable ( orange line ) . thus above @xmath289
the hpwc is the only stable solution . at @xmath290
the pattern solution ceases to exist ( red line ) . the uniform solution eq . ( [ eq : uniformsol_part3 ] ) , @xmath231 is illustrated in fig .
[ fig : unif_d=0_full ] . uniform solution of eq .
( [ eq : uniformsol_part3 ] ) . *
* op map , superimposed are the od borders ( gray ) , 90% ipsilateral eye dominance ( black ) , and 90% contralateral eye dominance ( white ) , @xmath291 .
dashed lines mark the unit cell of the regular pattern . * * distribution of preferred orientations . ]
this pwc contains only three pinwheels per unit cell leading to a pinwheel density of @xmath292 .
two of the three pinwheels are located at maxima of the od map ( contra peak ) while the remaining pinwheel is located at the minimum ( ipsi peak ) of the od map .
a remarkable property of this solution is that the pinwheel located at the od minimum , carries a topological charge of 1 such that each orientation is represented twice around this pinwheel .
pinwheels of this kind have not yet been observed in experimentally recorded op maps .
this kind of uniform solution corresponds to the structural pinwheel model by braitenberg @xcite .
we therefore refer to this solution as the _ braitenberg pinwheel crystal_. + the iso - orientation lines are again perfectly perpendicular to od borders and this is independent of the bias @xmath40 .
the solution is symmetric under a combined rotation by @xmath293 and an orientation shift by @xmath294 .
further it is symmetric under a rotation by @xmath295 .
the pattern is selective to all orientations but the distribution of represented orientations is not uniform .
the three orientations @xmath296 , @xmath297 are overrepresented , see fig . [
fig : unif_d=0_full]**. + overall this op map is dominated by uniform regions around hyperbolic points . in contrast to the ipsi center pwc all pinwheels in this op map organize a roughly similar fraction of the cortical surface .
* summary of results * + in this study we presented a symmetry - based analysis of models formalizing that visual cortical architecture is shaped by the coordinated optimization of different functional maps .
in particular , we focused on the question of whether and how different optimization principles specifically impact on the spatial layout of functional columns in the primary visual cortex .
we identified different representative candidate optimization principles .
we developed a dynamical systems approach for analyzing the simultaneous optimization of interacting maps and examined how their layout is influenced by coordinated optimization .
in particular , we found that inter - map coupling can stabilize pinwheel - rich layouts even if pinwheels are intrinsically unstable in the weak coupling limit .
we calculated and analyzed the stability properties of solutions forming spatially regular layouts with pinwheels arranged in a crystalline array .
we analyzed the structure of these pinwheel crystals in terms of their stability properties , spatial layout , and geometric inter - map relationships .
for all models , we calculated phase diagrams showing the stability of the pinwheel crystals depending on the od bias and the inter - map coupling strength .
although differing in detail and exhibiting distinct pinwheel crystal phases for strong coupling , the phase diagrams exhibited many commonalities in their structure .
these include the general fact that the hexagonal pwc phase is preceded by a phase of rhombic pwcs and that the range of od biases over which pinwheel crystallization occurs is confined to the stability region of od patch solutions .
+ * comparison to previous work * + our analytical calculations of attractor and ground states close a fundamental gap in the theory of visual cortical architecture and its development .
they rigorously establish that models of interacting op and od maps in principle offer a solution to the problem of pinwheel stability @xcite .
this problem and other aspects of the influence of od segregation on op maps have previously been studied in a series of models such as elastic net models @xcite , self - organizing map models @xcite , spin - like hamiltonian models @xcite , spectral filter models @xcite , correlation based models @xcite , and evolving field models @xcite .
several of these simulation studies found a higher number of pinwheels per hypercolumn if the op map is influenced by strong od segregation compared to the op layout in isolation or the influence of weak od segregation @xcite .
in such models , large gradients of op and od avoid each other @xcite . as a result ,
pinwheel centers tend to be located at centers of od columns as seen in experiments @xcite . by this mechanism ,
pinwheels are spatially trapped and pinwheel annihilation can be reduced @xcite . moreover , many models appear capable of reproducing realistic geometric inter - map relationships such as perpendicular intersection angles between od borders and iso - orientation lines @xcite .
tanaka et al . reported from numerical simulations that the relative positioning of orientation pinwheels and od columns was dependent on model parameters @xcite .
informative as they were , almost all of these previous studies entirely relied on simulation methodologies that do not easily permit to assess the progress and convergence of solutions . whether the reported patterns were attractors or just snapshots of transient states and whether the solutions would further develop towards pinwheel - free solutions or other states thus remained unclear .
moreover , in almost all previous models , a continuous variation of the inter - map coupling strength was not possible which makes it hard to disentangle the contribution of inter - map interactions from intrinsic mechanisms .
the only prior simulation study of a coordinated optimization model that tracked the number of pinwheels during the optimization process did not provide evidence that pinwheel annihilation could be stopped but only reported a modest reduction in annihilation efficiency @xcite . from this perspective
, the prior evidence for coordination induced pinwheel stabilization appears relatively limited .
our analytical results leave no room to doubt that map interactions can stabilize an intrinsically unstable pinwheel dynamics .
they also reveale that interaction of orientation preference with a stripe pattern of od is per se not capable of stabilizing pinwheels . +
* the mathematical structure of interaction models * + independent of its predictions , our study clarifies the general mathematical structure of interaction dominated optimization models . to the best of our knowledge
our study for the first time describes an analytical approach for examining the solutions of coordinated optimization models for op and od maps .
our symmetry - based phenomenological analysis of conceivable coupling terms provides a general classification and parametrization of biologically plausible coupling terms . to achieve this we mapped the optimization problem to a dynamical systems problem which allows for a perturbation expansion of fixed points , local minima , and optima . using weakly nonlinear analysis , we derived amplitude equations as an approximate description near the symmetry breaking transition .
we identified a limit in which inter - map coupling becomes effectively unidirectional enabling the use of the uncoupled od patterns .
we studied fixed points and calculated their stability properties for different types of inter - map coupling energies .
this analysis revealed a fundamental difference between high and low order coupling energies .
for the low order versions of these energies , a strong inter - map coupling typically leads to op map suppression , causing the orientation selectivity of all neurons to vanish .
in contrast , the higher order variants of the coupling energies do generally not cause map suppression but only influence pattern selection , see text s1 . we did not consider an interaction with the retinotopic map .
experimental results on geometric relationships between the retinotopic map and the op map are ambiguous . in case of ferret visual cortex high gradient regions of both maps
avoid each other @xcite . in case of cat , however , high gradient regions overlap @xcite .
such positive correlations can not be easily treated with dimension reduction models , see @xcite .
it is noteworthy that our phenomenological analysis identified coupling terms that could induce an attraction of high gradient regions .
such terms contain the gradient of only one field and can thus be considered as a mixture of the gradient and the product - type energy . + * conditions for pinwheel stabilization * + our results indicate that a patchy layout of a second visual map interacting with the op map is important for the effectiveness of pinwheel stabilization by inter - map coupling . such a patchy layout can be easily induced by an asymmetry in the representation of the corresponding stimulus feature such as eye dominance or spatial frequency preference . in spatial frequency maps , for instance , low spatial frequency patches tend to form islands in a sea of high spatial frequency preference @xcite . in our model , the patchy layout results from the overall dominance of one eye . in this case , od domains form a system of hexagonal patches rather than stripes enabling the capture and stabilization of pinwheels by inter - map coupling .
the results from all previous models did not support the view that od stripes are capable of stabilizing pinwheels @xcite .
our analysis shows that od stripes are indeed not able to stabilize pinwheels , a result that appears to be independent of the specific type of map interaction . in line with this
, several other theoretical studies , using numerical simulations @xcite , indicated that more banded od patterns lead to less pinwheel rich op maps .
for instance , in simulations using an elastic net model , the average pinwheel density of op maps interacting with a patchy od layout was reported substantially higher ( about 2.5 pinwheels per hypercolumn ) than for op maps interacting with a more stripe - like od layout ( about 2 pinwheels per hypercolumn ) @xcite . +
* experimental evidence for pinwheel stabilization by inter - map coupling * + several lines of biological evidence appear to support the picture of interaction induced pinwheel stabilization . supporting the notion that pinwheels might be stabilized by the interaction with patchy od columns , visual cortex
is indeed dominated by one eye in early postnatal development and has a pronounced patchy layout of od domains @xcite .
further support for the potential relevance of this picture comes from experiments in which the od map was artificially removed resulting apparently in a significantly smoother op map @xcite . in this context
it is noteworthy that macaque visual cortex appears to exhibit all three fundamental solutions of our model for od maps : stripes , hexagons , and a monocular solution , which are stable depending on the od bias . in the visual cortex of macaque monkeys ,
all three types of patterns are found near the transition to the monocular segment , see @xcite and fig .
( [ fig : constsolutions ] ) . here , od domains form bands in the binocular region and a system of ipsilateral eye patches at the transition zone to the monocular region where the contralateral eye gradually becomes more dominant .
if pinwheel stability depends on a geometric coupling to the system of od columns one predicts systematic differences in pinwheel density between these three zones of macaque primary visual cortex .
because od columns in the binocular region of macaque visual cortex are predominantly arranged in systems of od stripes our analysis also indicates that pinwheels in these regions are either stabilized by other patchy columnar systems or intrinsically stable .
+ * the geometry of interaction induced pinwheel crystals * + one important general observation from our results is that map organization was often not inferable by simple qualitative considerations on the energy functional .
the organization of interaction induced hexagonal pinwheel crystals reveals that the relation between coupling energy and resulting map structure is quite complex and often counter intuitive .
we analyzed the stationary patterns with respect to intersection angles and pinwheel positions . in all models ,
intersection angles of iso - orientation lines and od borders have a tendency towards perpendicular angles whether the energy term mathematically depends on this angle , as for the gradient - type energies , or not , as for the product - type energies .
intersection angle statistics thus are not a very sensitive indicator of the type of interaction optimized .
mathematically , these phenomena result from the complex interplay between the single map energies and the interaction energies . in case of the low order gradient - type inter - map coupling energy all pinwheels
are located at od extrema , as expected from the used coupling energy . for other analyzed coupling energies , however , the remaining pinwheels are located either at od saddle - points ( low order product - type energy ) or near od borders ( higher order gradient - type energy ) , in contrast to the expection that od extrema should be energetically preferred .
remarkably , such correlations , which are expected from the gradient - type coupling energies , occur also in the case of the product - type energies .
remarkably , in case of product type energies pinwheels are located at od saddle - points . which is not expected per se and presumably result from the periodic layout of op and od maps .
correlations between pinwheels and od saddle - points have not yet been studied quantitatively in experiments and may thus provide valuable information on the principles shaping cortical functional architecture . + * how informative is map structure ? * + our results demonstrate that , although distinct types of coupling energies can leave distinguishing signatures in the structure of maps shaped by interaction ( as the op map in our example ) , drawing precise conclusions about the coordinated optimization principle from observed map structures is not possible for the analyzed models . in the past numerous studies
have attempted to identify signatures of coordinated optimization in the layout of visual cortical maps and to infer the validity of specific optimization models from aspects of their coordinated geometry @xcite .
it was , however , never clarified theoretically in which respect and to which degree map layout and geometrical factors of inter - map relations are informative with respect to an underlying optimization principle . because our analysis provides complete information of the detailed relation between map geometry and optimization principle for the different models our results enable to critically assess whether different choices of energy functionals specifically impact on the predicted map structure and conversely what can be learned about the underlying optimization principle from observations of map structures .
+ we examined the impact of different interaction energies on the structure of local minima and ground states of models for the coordinated optimization of a complex and a real scalar feature map such as op and od maps .
the models were constructed such that in the absence of interactions , the maps reorganized into simple stripe or blob pattern .
in particular , the complex scalar map without interactions would form a periodic stripe pattern without any phase singularity . in all models ,
increasing the strength of interactions could eventually stabilize qualitatively different more complex and biologically more realistic patterns containing pinwheels that can become the energetic ground states for strong enough inter - map interactions .
the way in which this happens provides fundamental insights into the relationships between map structure and energy functionals in optimization models for visual cortical functional architecture .
+ our results demonstrate that the structure of maps shaped by inter - map interactions is in principle informative about the type of coupling energy . the organization of the complex scalar map that optimizes the joined energy functional was in general different for all different types of coupling terms examined .
we identified a class of hpwc solutions which become stable for large inter - map coupling .
this class depends on a single parameter which is specific to the used inter - map coupling energy .
furthermore , as shown in text s1 , pinwheel positions in rpwcs , tracked while increasing inter - map coupling strength , were different for different coupling terms examined and thus could in principle serve as a trace of the underlying optimization principle .
this demonstrates that , although pinwheel stabilization is not restricted to a particular choice of the interaction term , each analyzed phase diagram is specific to the used coupling energy . in particular , in the strong coupling regime substantial information can be obtained from a detailed inspection of solutions .
+ in the case of the product - type coupling energies , the resulting phase diagrams are relatively complex as stationary solutions and stability borders depend on the magnitude of the od bias . here , even quantitative values of model parameters can in principle be constrained by analysis of the map layout .
in contrast , for the gradient - type coupling energies , the bias dependence can be absorbed into the coupling strength and only selects the stationary od pattern .
this leads to relatively simple phase diagrams . for these models
map layout is thus uninformative of quantitative model parameters .
we identified several biologically very implausible op patterns .
in the case of the product - type energies , we found orientation scotoma solutions which are selective to only two preferred orientations . in the case of the low order gradient - type energy , we found op patterns containing pinwheels with a topological charge of 1 which have not yet been observed in experiments .
if the relevant terms in the coupling energy could be determined by other means , the parameter regions in which these patterns occur could be used to constrain model parameters by theoretical bounds .
+ the information provided by map structure overall appears qualitative rather than quantitative . in both low order inter - map coupling energies ( and the gradient - type higher order coupling energy ,
see text s1 ) , hpwc patterns resulting from strong interactions were fixed , not exhibiting any substantial dependence on the precise choice of interaction coefficient . in principle , the spatial organization of stimulus preferences in a map is an infinite dimensional object that could sensitively depend in distinct ways to a large number of model parameters .
it is thus not a trivial property that this structure often gives essentially no information about the value of coupling constants in our models .
the situation , however , is reversed when considering the structure of rpwcs .
these solutions exist and are stable although energetically not favored in the absence of inter - map interactions .
some of their pinwheel positions continuously depend on the strength of inter - map interactions . these solutions and
their parameter dependence nevertheless are also largely uninformative about the nature of the interaction energy .
this results from the fact that rpwcs are fundamentally uncoupled system solutions that are only modified by the inter - map interaction . as pointed out before , preferentially orthogonal intersection angles between iso - orientation lines and od borders appear to be a general feature of coordinated optimization models in the strong coupling regime .
although the detailed form of the intersection angle histogram is solution and thus model specific , our analysis does not corroborate attempts to use this feature to support specific optimization principles , see also @xcite .
the stabilization of pinwheel crystals for strong inter - map coupling appears to be universal and provides per se no specific information about the underlying optimization principle .
in fact , the general structure of the amplitude equations is universal and only the coupling coefficients change when changing the coupling energy .
it is thus expected that also for other coupling energies , respecting the proposed set of symmetries , pwc solutions can become stable for large enough inter - map coupling .
+ * conclusions * + our analysis conclusively demonstrates that od segregation can stabilize pinwheels and induce pinwheel - rich optima in models for the coordinated optimization of op and od maps when pinwheels are intrinsically unstable in the uncoupled dynamics of the op map .
this allows to systematically assess the possibility that inter - map coupling might be the mechanism of pinwheel stabilization in the visual cortex .
the analytical approach developed here is independent of details of specific optimization principles and thus allowed to systematically analyze how different optimization principles impact on map layout .
moreover , our analysis clarifies to which extend the observation of the layout in physiological maps can provide information about optimization principles shaping visual cortical organization .
+ the common design observed in experimental op maps @xcite is , however , not reproduced by the optima of the analyzed optimization principles . whether this is a consequence of the applied weakly nonlinear analysis or of the low number of optimized feature maps or
should be considered a generic feature of coordinated optimization models will be examined in part ( ii ) of this study @xcite . in part
( ii ) we complement our analytical studies by numerical simulations of the full field dynamics .
such simulations allow to study the rearrangement of maps during the optimization process , to study the timescales on which optimization is expected to take place , and to lift many of the mathematical assumptions employed by the above analysis .
in particular , we concentrate on the higher order inter - map coupling energies for which the derived amplitude equations involved several simplifying conditions , see text s1 .
we studied the intersection angles between iso - orientation lines and od borders .
the intersection angle of an od border with an iso - orientation contour @xmath298 is given by @xmath299 where @xmath2 denotes the position of the od zero - contour lines .
a continuous expression for the op gradient is given by @xmath300 .
we calculated the frequency of intersection angles in the range @xmath301 $ ] . in this way those parts of the maps are emphasized from which the most significant information about the intersection angles can be obtained @xcite .
these are the regions where the op gradient is high and thus every intersection angle receives a statistical weight according to @xmath302 .
for an alternative method see @xcite .
we studied how the emerging od map depends on the overall eye dominance .
to this end we mapped the uncoupled od dynamics to a swift - hohenberg equation containing a quadratic interaction term instead of a constant bias .
this allowed for the use of weakly nonlinear analysis to derive amplitude equations as an approximate description of the shifted od dynamics near the bifurcation point .
we identified the stationary solutions and studied their stability properties .
finally , we derived expressions for the fraction of contralateral eye dominance for the stable solutions . here
we describe how to map the swift - hohenberg equation @xmath303 to one with a quadratic interaction term . to eliminate the constant term we shift the field by a constant amount @xmath304 .
this changes the linear and nonlinear terms as @xmath305 we define the new parameters @xmath306 and @xmath307 .
this leads to the new dynamics @xmath308 the condition that the constant part is zero is thus given by @xmath309 for @xmath310 the real solution to eq .
( [ eq : mapping_deltaeq_part3 ] ) is given by @xmath311 with @xmath312 .
for small @xmath40 this formula is approximated as @xmath313 the uncoupled od dynamics we consider in the following is therefore given by @xmath314 this equation has been extensively studied in pattern formation literature @xcite .
we studied eq .
( [ eq : od_uncoupled ] ) using weakly nonlinear analysis .
this method leads to amplitude equation as an approximate description of the full field dynamics eq .
( [ eq : od_uncoupled ] ) near the bifurcation point @xmath315 .
we summarize the derivation of the amplitude equations for the od dynamics which is of the form @xmath316 -n_3[o , o , o]\ , , \ ] ] with the linear operator @xmath317 . in this section
we use for simplicity the variables @xmath318 instead of @xmath319 .
the derivation is performed for general quadratic and cubic nonlinearities but are specified later according to eq .
( [ eq : dynamicsshcoupled_ref ] ) as @xmath320=o^3 $ ] and @xmath321=\gamma \ , o^2 $ ] . for the calculations in the following , it is useful to separate @xmath105 from the linear operator @xmath322 therefore the largest eigenvalue of @xmath323 is zero .
the amplitude of the field @xmath44 is assumed to be small near the onset @xmath324 and thus the nonlinearities are small .
we therefore expand both the field @xmath44 and the control parameter @xmath105 in powers of a small expansion parameter @xmath106 as @xmath325 and @xmath326 the dynamics at the critical point @xmath327 becomes arbitrarily slow since the intrinsic timescale @xmath328 diverges at the critical point . to compensate we introduce a rescaled time scale @xmath329 as @xmath330 in order for all terms in eq .
( [ eq : oduncoupled ] ) to be of the same order the quadratic interaction term @xmath331 must be small .
we therefore rescale @xmath331 as @xmath332 .
this preserves the nature of the bifurcation compared to the case @xmath333 . + we insert the expansion eq .
( [ eq : expand_o ] ) and eq .
( [ eq : expand_mu ] ) in the dynamics eq .
( [ eq : oduncoupled ] ) and get @xmath334\right ) \nonumber \\ & + & \ , \mu^3 \left(-\hat{l}^0 o_3+r_1 \left(o_2-\partial_t o_2 \right ) + r_2 \left(o_1-\partial_t o_1 \right)-n_3[o_1,o_1,o_1 ] \right)\nonumber \\ & \vdots & \end{aligned}\ ] ] we sort and collect all terms in order of their power in @xmath106 . the equation can be fulfilled for @xmath335 only if each of these terms is zero .
we therefore solve the equation order by order . in the leading order
we get the homogeneous equation @xmath336 thus @xmath114 is an element of the kernel of @xmath323 .
the kernel contains linear combinations of modes with wavevector @xmath177 on the critical circle @xmath337 . at this level
any of such wavevectors is possible .
we choose @xmath338 where the wavevectors are chosen to be equally spaced @xmath339 and the complex amplitudes @xmath119 .
the homogeneous equation leaves the amplitudes @xmath154 undetermined .
these amplitudes are fixed by the higher order equations .
besides the leading order homogeneous equation we get inhomogeneous equations of the form @xmath340 to solve this inhomogeneous equation we first apply a solvability condition .
we thus apply the _ fredholm alternative theorem _ to eq .
( [ eq : inhom_o_part3 ] ) . since the operator @xmath323 is self - adjoint @xmath341 , the equation is solvable if and only if @xmath342 is orthogonal to the kernel of @xmath323 i.e. @xmath343 the orthogonality to the kernel can be expressed by a projection operator @xmath344 onto the kernel and the condition @xmath345 can be rewritten as @xmath346 .
+ at second order we get @xmath347 applying the solvability condition eq .
( [ eq : solvecond_o ] ) we see that this equation can be fulfilled only for @xmath348 . at third order
we get @xmath349 -n_3[o_1,o_1,o_1 ] \ , .\ ] ] the parameter @xmath350 sets the scale in which @xmath114 is measured and we can set @xmath351 .
we apply the solvability condition and get @xmath352-\hat{p}_c n_3[o_1,o_1,o_1 ] \ , .\ ] ] we insert our ansatz eq .
( [ eq : kernelod ] ) which leads to the amplitude equations at third order @xmath353 -\hat{p}_i \sum_{j , k } b_j b_k b_l e^{-\imath \vec{k}_i \vec{x } } n_3[e^{\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_k \vec{x}},e^{\imath \vec{k}_l \vec{x } } ] \ , , \ ] ] where @xmath354 is the projection operator onto the subspace @xmath355 of the kernel .
@xmath354 picks out all combinations of the modes which have their wavevector equal to @xmath176 . in our case
the three active modes form a so called triad resonance @xmath356 .
the quadratic coupling terms which are resonant to the mode @xmath357 are therefore given by @xmath358 + n_2[e^{-\imath \vec{k}_3 \vec{x}},e^{-\imath \vec{k}_2 \vec{x } } ] \right)\ , .\ ] ] resonant contributions from the cubic nonlinearity result from terms of the form @xmath359 .
their coupling coefficients are given by @xmath360 + n_3[e^{\imath \vec{k}_i \vec{x}},e^{-\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_j \vec{x}}]+ n_3[e^{\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_i \vec{x}},e^{-\imath \vec{k}_j \vec{x}}]+\nonumber \\ & & n_3[e^{-\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_i \vec{x}},e^{\imath \vec{k}_j \vec{x}}]+ n_3[e^{\imath \vec{k}_j \vec{x}},e^{-\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_i \vec{x}}]+ n_3[e^{-\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_j \vec{x}},e^{\imath \vec{k}_i \vec{x}}]\ , , \end{aligned}\ ] ] and @xmath361 + n_3[e^{\imath \vec{k}_i \vec{x}},e^{-\imath \vec{k}_i \vec{x}},e^{\imath \vec{k}_i \vec{x}}]+ n_3[e^{-\imath \vec{k}_i \vec{x}},e^{\imath \vec{k}_i \vec{x}},e^{\imath \vec{k}_i \vec{x}}]\ , .\ ] ] when specifying the nonlinearities eq .
( [ eq : dynamicsshcoupled_ref ] ) the coupling coefficients are given by @xmath362 . finally , the amplitude equations ( here in the shifted variables ( @xmath363 ) are given by @xmath364 where we scaled back to the original time variable @xmath365 .
equations for @xmath366 and @xmath367 are given by cyclic permutation of the indices .
the amplitude equations ( [ eq : ampl_od_uncoupled ] ) have three types of stationary solutions , namely od stripes @xmath368 with @xmath369 , hexagons @xmath370 with the resonance condition @xmath371 and @xmath372 .
finally , there is a homogeneous solution with spatially constant eye dominance @xmath373 the spatial average of all solutions is @xmath374 .
the course of @xmath375 , @xmath376 , and of @xmath179 is shown in fig .
[ fig : constsolutions ] .
( [ eq : stripes ] ) ( blue ) , @xmath377 eq . ( [ eq : hex ] ) ( red ) , and of @xmath179 eq . ( [ eq : delta ] ) ( green ) for @xmath251 .
the solutions are plotted in solid lines within their stability ranges . * * od map of macaque monkey . adapted from @xcite . * * details of * * with stripe - like , patchy , and homogeneous layout . ]
we decomposed the amplitude equations ( [ eq : ampl_od_uncoupled ] ) into the real and imaginary parts . from the imaginary part
we get the phase equation @xmath378 and equations for @xmath379 by cyclic permutation of the indices .
the stationary phases are given by @xmath380 .
the phase equation can be derived from the potential @xmath381=-2\widetilde{\gamma } \cos ( \psi_1+\psi_2+\psi_3)$ ] .
we see that the solution @xmath382 is stable for @xmath383 and the solution @xmath384 is stable for @xmath385 .
+ we calculate the stability borders of the od stripe , hexagon , and constant solution in the uncoupled case .
this treatment follows @xcite . in case of stripes
the three modes of the amplitude equations are perturbed as @xmath386 assuming small perturbations @xmath387 , and @xmath388 .
this leads to the linear equations @xmath389 with the stability matrix @xmath390 the corresponding eigenvalues are given by @xmath391 this leads to the two borders for the stripe stability @xmath392 in terms of the original variables @xmath393 the borders are given by ( @xmath394 ) @xmath395 to derive the stability borders for the hexagon solution @xmath396 we perturb the amplitudes as @xmath397 the stability matrix is then given by @xmath398 and the corresponding eigenvalues are given by @xmath399 the stability borders for the hexagon solution are given by @xmath400 in terms of the original variables we finally get @xmath401 the phase diagram of this model is depicted in fig .
[ fig : constterm]**. ) . dashed lines : stability border of hexagon solutions , solid line : stability border of stripe solution , gray regions : stability region of constant solution * * percentage of neurons dominated by the contralateral eye plotted for the three stationary solutions .
circles : numerically obtained values , solid lines : @xmath402 and @xmath403 . ] it shows the stability borders @xmath404 , and @xmath405 for the three solutions obtained by linear stability analysis . without a bias term the od map is either constant , for @xmath406 , or has a stripe layout , for @xmath407 . for positive
@xmath105 and increasing bias term there are two transition regions , first a transition region from stripes to hexagons and second a transition region from hexagons to the constant solution .
+ the spatial layout of the od hexagons consists of hexagonal arrays of ipsilateral eye dominance blobs in a sea of contralateral eye dominance , see fig .
[ fig : constterm]**. to compare the obtained solutions with physiological od maps we quantified the fraction of neurons selective to the contralateral eye inputs . for stripe and hexagon solutions we thus calculated
the fraction of contralateral eye dominated territory @xmath402 and @xmath403 . in case of stripes
this is a purely one - dimensional problem .
the zeros of the field are given by @xmath408 with the solution @xmath409 as the field has a periodicity of @xmath0 the area fraction is given by @xmath410 in case of hexagons we observe that the territory of negative @xmath3 values is approximately a circular area .
we obtain the fraction of negative @xmath3 values by relating this area to the area of the whole hexagonal lattice . in case of hexagons the field is given by @xmath411 as an approximation we project the field onto the @xmath412-axis and choose for simplicity @xmath413 .
the field has its maximum at the origin @xmath414 .
the projection leads to @xmath415 the zeros @xmath416 are located at @xmath417 the circular area of positive @xmath396 values is now given by @xmath418 .
the periodicity of the hexagonal pattern is given by @xmath419 .
this leads to @xmath420 .
the area of the hexagon is therefore given by @xmath421 .
the contra fraction is finally given by @xmath422 the course of the fractions @xmath402 and @xmath403 is shown in fig .
[ fig : constterm]**. at the border @xmath242 , where hexagons become stable @xmath423 . at the border @xmath262 , where hexagons loose stability @xmath424 .
both quantities are independent of @xmath105 .
we confirmed our results by direct numerical calculation of the fraction of positive @xmath396 pixel values .
deviations from the result eq .
( [ eq : chex ] ) are small .
for @xmath425 the zeros of eq .
( [ eq : zerosofhex ] ) are not that well approximated with a circular shape and the projection described above leads to the small deviations which decrease with increasing bias @xmath40 . *
text s1 : supporting information for + coordinated optimization of visual cortical maps + ( i ) symmetry - based analysis *
we studied the coupled swift - hohenberg equations @xmath426 -n_{7,c}[z , z,\overline{z},o , o , o , o ] \nonumber \\
\partial_t \ , o(\mathbf{x},t ) & = & r_o o(\mathbf{x},t ) -\hat{l}_o^0 \ , o(\mathbf{x},t ) + n_{2,u}[o , o]-n_{3,u}[o , o , o ] -\tilde{n}_{7,c}[o , o , o , z , z,\overline{z},\overline{z}]\ , , \end{aligned}\ ] ] with the higher order inter - map coupling energies @xmath427 using weakly nonlinear analysis .
we study eq .
( [ eq : fieldeq_foramplderivs1 ] ) close to the pattern forming bifurcation where @xmath104 and @xmath105 are small .
we therefore expand both control parameters in powers of the small expansion parameter @xmath106 @xmath107 close to the bifurcation the fileds are small and thus nonlinearities are weak .
we therefore expand both fields as @xmath108 we further introduced a common slow timescale @xmath109 and insert the expansions in eq .
( [ eq : fieldeq_foramplderivs1 ] ) and get @xmath428\right ) \nonumber \\ & & \vdots \nonumber \\ & & + \mu^7 \left ( -\hat{l}^0 z_7 + r_{z2 } z_5 + r_{z4}z3+r_{z6}z_1 + \dots + n_{3,u}[z_5,z_1,\overline{z}_1 ] \right ) \nonumber\\ & & + \mu^7 \left(- n_{7,c}[z_1,z_1,\overline{z}_1,o_1,o_1,o_1,o_1]\right ) \nonumber \\ & & \vdots\end{aligned}\ ] ] and @xmath429 \right)\nonumber\\ & & + \mu^3\left(- r_{z2 } \partial_t o_1+r_{o2 } o_1+r_{o1 } o_2- r_{z1 } \partial_t
o_2-\hat{l}^0o_3-\tilde{n}_{3,u}[o_1,o_1,o_1]\right ) \nonumber \\ & & \vdots \nonumber \\ & & + \mu^7 \left(-\hat{l}^0 o_7+r_{o2 } o_5+r_{o4 } o_3+r_{o6 } o_1+\dots -\tilde{n}_{3,u}[o_5,o_1,o_1]-\tilde{n}_{2,u}[o_1,o_5]-\dots \right)\nonumber \\ & & + \mu^7\left(- \tilde{n}_{7,c}[o_1,o_1,o_1,z_1,z_1,\overline{z}_1,\overline{z}_1]\right ) \nonumber \\ & & \vdots\end{aligned}\ ] ] we consider amplitude equations up to seventh order as this is the order where the nonlinearity of the higher order coupling energy enters first .
( [ eq : expansion_z_coupled ] ) and eq .
( [ eq : expansion_o_coupled ] ) to be fulfilled each individual order in @xmath106 has to be zero . at linear order in @xmath106
we get the two homogeneous equations @xmath430 thus @xmath113 and @xmath114 are elements of the kernel of @xmath115 and @xmath116 .
both kernels contain linear combinations of modes with a wavevector on the critical circle i.e. @xmath431 with the complex amplitudes @xmath432 , @xmath433 and @xmath120 , @xmath121 . in view of the hexagonal or stripe layout of the od pattern shown in fig . [ fig : numerics ] , @xmath122 is an appropriate choice . since in cat
visual cortex the typical wavelength for od and op maps are approximately the same @xcite i.e. @xmath123 the fourier components of the emerging pattern are located on a common critical circle . to account for species differences we also analyzed models with detuned op and od wavelengths in part ( ii ) of this study .
+ at second order in @xmath106 we get @xmath124 as @xmath113 and @xmath114 are elements of the kernel @xmath125 . at third order , when applying the solvability condition ( see methods ) , we get @xmath434 \nonumber \\ r_{z2 } \partial_t \ , o_1&= & r_{o2 } o_1-\sqrt{r_{o2 } } \ , \hat{p}_c \tilde{n}_{2,u}[o_1,o_1]- \hat{p}_c \tilde{n}_{3,u}[o_1,o_1,o_1 ] \ , .
\end{aligned}\ ] ] we insert the leading order fields eq .
( [ eq_z1o1 ] ) and obtain the amplitude equations @xmath435 these uncoupled amplitude equations obtain corrections at higher order .
there are fifth order , seventh order and even higher order corrections to the uncoupled amplitude equations .
in addition , at seventh order enters the nonlinearity of the higher order inter - map coupling energies .
the amplitude equations up to seventh order are thus derived from @xmath436 \nonumber \\ r_{z2 } \partial_t \ , z_3 & = & r_{z2 } z_3- \dots - \hat{p } n_{3,u}[z_1,z_1,\overline{z_3 } ] \\ r_{z2 } \partial_t \ , z_5 & = & r_{z2 } z_5-\dots - \hat{p}_c n_{3,u}[z_3,z_1,\overline{z_3}]- \hat{p}_c n_{7,c}[z_1,z_1,\overline{z}_1,o_1,o_1,o_1,o_1 ] \
, , \nonumber\end{aligned}\ ] ] and corresponding equations for the fields @xmath114 , @xmath437 , and @xmath438 .
the field @xmath113 is given in eq .
( [ eq_z1o1 ] ) and its amplitudes @xmath439 and @xmath440 are determined at third order .
the field @xmath441 contains contributions from modes off the critical circle @xmath442 , @xmath443 and on the critical circle i.e. @xmath444 .
its amplitude @xmath445 are determined at fifth order .
the field @xmath446 also contains contributions from modes off the critical circle @xmath447 and on the critical circle i.e. @xmath448 .
its amplitude @xmath449 are determined at seventh order .
this leads to a series of amplitude equations @xmath450 which are solved order by order .
we set @xmath136 and @xmath137 and rescale to the fast time .
this leads to @xmath451 we can combine the amplitude equations up to seventh order by introducing the amplitudes @xmath452 and @xmath453 .
this leads to the amplitude equations @xmath454 for simplicity we have written only the simplest inter - map coupling terms .
depending on the configuration of active modes additional contributions may enter the amplitude equations .
in addition , for the product - type coupling energy , there are coupling terms which contain the constant @xmath128 , see methods . in case of @xmath133 the inter - map coupling terms in dynamics of the modes @xmath134 are small . in this limit
the dynamics of the modes @xmath134 decouples from the modes @xmath135 and we can use the uncoupled od dynamics , see methods . in the following ,
we use the effective inter - map coupling strength @xmath455 ( and @xmath456 ) .
in this article we demonstrated that the low order coupling terms can lead to a complete suppression of op selectivity i.e. vanishing magnitude of the order parameter @xmath141 .
as the coupling terms are effectively linear they not only influence pattern selection but also whether there is a pattern at all .
this is in general not the case for higher order coupling energies using the amplitude equations eq .
( [ eq : ampleq_coupled_seventhorder ] ) . in this case
the coupling is an effective cubic interaction term and complete selectivity suppression is impossible .
moreover , as in the low - order energy case , we could identify the limit @xmath155 in which the backreaction onto the od map formally becomes negligible .
the potential is of the form @xmath457 where @xmath226 denotes the kronecker delta and the uncoupled contributions @xmath170 amplitude equations can be derived from the potential by @xmath458 .
we have not written terms involving the modes @xmath459 or @xmath460 .
the complete amplitude equations involving all modes and the corresponding coupling coefficients are given in text s2 .
as for the low order coupling energies terms involving the constant @xmath128 depend only on the coupling coefficient of the product - type energy @xmath461 . in the following we specify the amplitude equations for negligible backreaction where @xmath462 , @xmath463 , or @xmath464 .
+ first , we studied the higher order product - type inter - map coupling energy @xmath465 .
as for the lower order version of this coupling energy the shift @xmath179 explicitly enters the amplitude equations resulting in a rather complex parameter dependence , see eq .
( 68 ) in the methods section .
in the case of od stripes the amplitude equations of op modes read @xmath466 where @xmath226 denotes the kronecker delta and @xmath467 the equation for the mode @xmath208 is given by interchanging the modes @xmath207 and @xmath208 in eq .
( [ eq : ampl_tau_odstripes ] ) .
the equations for the modes @xmath132 are given by interchanging the modes @xmath131 and @xmath132 and interchanging the modes @xmath184 and @xmath185 .
+ in this case , at low inter - map coupling the op stripes given by @xmath468 with @xmath469 run parallel to the od stripes .
their stationary amplitudes are given by @xmath470 with @xmath471 , @xmath472 , @xmath473 , @xmath474 .
the parameter dependence of these stripe solutions is shown in fig .
[ fig : amplitudes_tau]**. + at large inter - map coupling the attractor states of the op map consist of a stripe pattern containing only two preferred orientations , namely @xmath248 and @xmath249 .
the zero contour lines of the od map are along the maximum amplitude of orientation preference minimizing the energy term .
+ in addition there are rhombic solutions @xmath475 which exist also in the uncoupled case , see fig .
[ fig : amplitudes_tau]**. however , these rhombic solutions are energetically not favored compared to stripe solutions , see fig .
[ fig : amplitudes_tau]**. the inclusion of the inter - map coupling makes these rhombic solution even more stripe - like .
+ in case of a od constant solution the amplitude equations read @xmath476 with @xmath477 , @xmath478 and @xmath479 .
inter - map coupling thus leads to a renormalization of the uncoupled interaction terms .
stationary solutions are stripes with the amplitude @xmath480 and rhombic solutions with the stationary phases @xmath481 and the stationary amplitudes @xmath482 in the case of od hexagons we identify , in addition to stripe - like and rhombic solutions , uniform solutions @xmath483 . when solving the amplitude equations numerically we have seen that the phase relations vary with the inter - map coupling strength @xmath461 for non - uniform solutions . but for the uniform solution the phase relations are independent of the inter - map coupling strength .
we use the ansatz for uniform solutions @xmath225 where @xmath226 is the kronecker delta and @xmath227 a constant parameter .
this leads to the stationarity condition @xmath484 \sin \delta = 0\ , .\ ] ] four types of stationary solutions exist namely the @xmath485 , which we already observed in case of the low order energies , and the solutions @xmath486 which depends on @xmath287 and @xmath128 and thus on the bias @xmath40 .
the course of eq .
( [ eq : unif_d_gamma ] ) as a function of @xmath40 is shown in fig .
[ fig : pd_tau]**. stationary amplitudes for these solutions are given by @xmath487 we study the stability properties of op stripe - like , rhombic and uniform solutions using linear stability analysis .
the eigenvalues of the stability matrix , see text s2 , are calculated numerically .
linear stability analysis shows that for @xmath488 the @xmath231 solution is unstable for all bias values .
the stability region of the @xmath232 solution and the solution eq .
( [ eq : unif_d_gamma ] ) is bias dependent . the bias dependent solution eq .
( [ eq : unif_d_gamma ] ) is stable for @xmath489 and @xmath490 for which @xmath232 , see fig .
[ fig : pd_tau]**. for larger bias @xmath491 only the @xmath492 uniform solution is stable . the parameter dependence of op solutions when interacting with od stripes is shown in fig .
[ fig : amplitudes_tau]**. similar to the low order variant of this coupling energy the amplitude of the stripes pattern @xmath246 is suppressed while the amplitude of the opposite mode @xmath247 grows . finally both amplitudes collapse , leading to an orientation scotoma solution .
in contrast to the low order variant this stripe pattern is stable for arbitrary large inter - map coupling . in case of op
rhombic solutions inter - map coupling transforms this solution by reducing the amplitudes @xmath205 while increasing the amplitudes @xmath206 . without od bias
this solution is then transformed into the orientation scotoma stripe pattern , similar to the low order variant of this energy .
in contrast to the low order energy , for non - zero bias the amplitudes @xmath207 and @xmath208 stay small but non - zero .
+ the parameter dependence of op solutions when interacting with od hexagons is shown in fig .
[ fig : amplitudes_tau]**. for a small od bias ( @xmath242 ) op rhombic solutions decay into op stripe - like patterns .
these stripe - like patterns stay stable also for large - inter map coupling . in case of a larger od bias ( @xmath493 ) , both the op stripe and the op rhombic solutions decay into the uniform pwc solution .
thus for small bias there is a bistability between stripe - like and uniform pwc solutions while for larger od bias the uniform pwc solution is the only stable solution .
the potential of op stripe and op rhombic solutions is shown in fig .
[ fig : amplitudes_tau]**. in the uncoupled case as well as for small inter - map coupling strength op stripe solutions are for all bias values the energetic ground state .
for large inter - map coupling and a small bias ( @xmath494 ) rhombic solutions are unstable and the stripe - like solutions are energetically preferred compared to pwc solutions . for larger bias ,
however , pwc solutions are the only stable solutions for large inter - map coupling . .
solid ( dashed ) lines : stable ( unstable ) solutions . *
* od stripes , @xmath254 ( blue ) , @xmath242 ( green ) , @xmath495 ( red ) . * * od hexagons , @xmath242 ( blue ) , @xmath253 ( red ) . * * transition from op stripe solutions , * * transition from op rhombic solutions .
* * potential , eq .
( [ eq : potential_septic_part3 ] ) , of op stripes and op rhombs interacting with od stripes . * * potential , eq .
( [ eq : potential_septic_part3 ] ) , of op stripes , op rhombs , and hpwc interacting with od hexagons .
arrows indicate corresponding lines in the phase diagram , fig .
( [ fig : pd_tau ] ) . ]
the stability properties of all stationary solutions are summarized in the phase diagram fig .
[ fig : pd_tau ] .
, @xmath255 . vertical black lines : stability range of od stripes , hexagons , and constant solutions .
magenta ( orange ) line : stability border of orientation scotoma stripes .
green solid line : stability border of rhombic solutions .
red solid line : stability border of pwc solutions , red dashed line : @xmath496 , * * course of eq .
( [ eq : unif_d_gamma ] ) , dashed line : @xmath232 .
* * stability border between eq .
( [ eq : unif_d_gamma ] ) solution and the @xmath232 solution as a function of @xmath105 ( vertical red line in * * ) . ] compared to the gradient - type interaction energy we can not scale out the dependence on @xmath105 .
the phase diagram is thus plotted for @xmath251 .
we rescale the inter - map coupling strength as @xmath497 where @xmath287 is the stationary amplitude of the od hexagons . in the regime of stable od stripes
there is a transition from op stripes towards the orientation scotoma stripe solution . in the regime of stable od hexagons
there is a transition from op stripes towards pwc solutions ( red line ) .
the stability border of pwc solutions is strongly od bias dependent and has a peak at @xmath498 . for small od bias
@xmath40 the uniform solution eq .
( [ eq : unif_d_gamma ] ) is stable . with increasing bias
there is a smooth transition of this solution until at @xmath499 the @xmath492 uniform solution becomes stable . in fig .
[ fig : pd_tau ] * * the stability border @xmath496 between the two types of uniform solutions is plotted as a function of @xmath105 .
we observe that there is only a weak dependence on the control parameter and @xmath500 .
figure [ fig : unif_gamma_depend ] illustrates the uniform solutions eq .
( [ eq : unif_d_gamma ] ) for different values of the od bias @xmath40 . ) * * @xmath242 , * * @xmath501 , * * @xmath502 , * * @xmath503 . op map , superimposed are the od borders ( gray ) , 90% ipsilateral eye dominance ( black ) , and 90% contralateral eye dominance ( white ) , @xmath251 .
dashed lines mark the unit cell of the regular pattern . * * distribution of orientation preference . * * intersection angles between iso - orientation lines and od borders . ] for small bias , the op pattern has six pinwheels per unit cell .
two of them are located at od maxima while one is located at an od minimum .
the remaining three pinwheels are located near the od border . with increasing bias ,
these three pinwheels are pushed further away from the od border , being attracted to the od maxima . with further increasing bias three shifted pinwheels
merge with the one at the od maximum building a single charge 1 pinwheel centered on a contralateral peak .
the remaining two pinwheels are located at an ispi and contra peak , respectively .
note , compared to the braitenberg pwc of the @xmath231 uniform solution the charge 1 pinwheel here is located at the contralateral od peak . finally , the charge 1 pinwheels split up again into four pinwheels . with increasing bias
the solution more and more resembles the ipsi - center pwc ( @xmath232 solution ) which is stable also in the lower order version of the coupling energy . finally , at @xmath504 the ipsi - center pwc becomes stable and fixed for @xmath505 . the distribution of preferred orientations for different values of the bias @xmath40 is shown in fig .
[ fig : unif_gamma_depend ] * * , reflecting the symmetry of each pattern . the distribution of intersection angles is shown in fig .
[ fig : unif_gamma_depend]**. remarkably , all solutions show a tendency towards perpendicular intersection angles .
this tendency is more pronounced with increasing od bias . at about @xmath506 parallel intersection angles
are completely absent and at @xmath504 there are exclusively perpendicular intersection angles .
finally , we examine the higher order version of the gradient - type inter - map coupling .
the interaction terms are independent of the od shift @xmath128 . in this case
the coupling strength can be rescaled as @xmath508 and is therefore independent of the bias @xmath40 .
the bias in this case only determines the stability of od stripes , hexagons or the constant solution . as for its lower order pendant
a coupling to od stripes is relatively easy to analyze .
the energetic ground state corresponds to op stripes with the direction perpendicular to the od stripes for which @xmath272 .
in addition , there are rhombic solutions with the stationary amplitudes @xmath509 . in case
the od map is a constant , @xmath510 , the gradient - type inter - map coupling leaves the op unaffected . as for its lower order pendant
the stationary states are therefore op stripes running in an arbitrary direction and the uncoupled rhombic solutions .
+ in case of od hexagons we identified three types of non - uniform solutions .
besides stripe - like solutions of @xmath1 with one dominant mode we find rpwcs @xmath511 with @xmath512 and distorted rpwcs @xmath513 with @xmath514 .
note , that distorted rpwcs are not stable in case of the product - type coupling energy or the analyzed low - order coupling energies . for these non - uniform solutions
the stationary phases are inter - map coupling strength dependent .
we therefore calculate the stationary phases and amplitudes numerically using a newton method and initial conditions close to these solutions .
+ in case of od hexagons there are further uniform solutions @xmath515 and @xmath516 . the imaginary part of the amplitude equations , see text s1 , leads to equations for the phases @xmath517 . the ansatz eq .
( [ eq : uniformsol_part3 ] ) leads to the stationarity condition @xmath518 the solutions are @xmath485 , and @xmath519 where the stationary amplitude are given by @xmath520 we calculated the stability properties of all stationary solutions by linear stability analysis considering perturbations of the amplitudes @xmath234 , @xmath235 and of the phases @xmath236 , @xmath237 .
this leads to a perturbation matrix @xmath238 . in general amplitude and phase perturbations
do not decouple .
we therefore calculate the eigenvalues of the perturbation @xmath238 matrix numerically .
it turns out that for this type of coupling energy only the uniform solutions with @xmath521 are stable while the @xmath231 and @xmath232 solutions are unstable in general . for increasing inter - map coupling strength the amplitudes of the op stripe and op rhombic solutions
are shown in fig .
[ fig : bif_eps]**. , * * solid ( dashed ) lines : stable ( unstable ) solutions .
blue : rpwc , green : distorted rpwc , red : hpwc .
black lines : stripe - like solutions . *
* potential , eq .
( [ eq : potential_septic_part3 ] ) , of op stripes ( black ) , op rhombs ( blue ) , and hpwc solutions ( red ) .
arrows indicate corresponding lines in the phase diagram , fig .
( s[fig : pdcoupled ] ) . ] in case of stable od hexagons there is a transition from rpwc ( blue ) towards distorted rpwc ( green ) . the distorted rpwcs then decay into the hpwc ( red ) .
in case of op stripes ( black dashed lines ) inter - map coupling leads to a slight suppression of the dominant mode and a growth of the remaining modes .
this growth saturates at small amplitudes and thus the solution stays stripe - like .
this stripe - like solution remains stable for arbitrary large inter - map coupling .
therefore there is a bistability between hpwc solutions and stripe - like solutions for large inter - map coupling .
+ the stability borders for the rpwc and distorted rpwc solutions were obtained by calculating their bifurcation diagram numerically from the amplitude equations , see text s1 .
with increasing map coupling we observe a transition from a rpwc towards a distorted rpwc at @xmath522 ( blue dashed line in fig . [ fig : pdcoupled ] * * ) , see also fig .
[ fig : pwwandering]**. the distorted rpwc loses its stability at @xmath523 ( blue solid line in fig .
[ fig : pdcoupled ] * * ) and from thereon all amplitudes are equal corresponding to the hpwc .
there is a bistability between hpwc , rpwc , and stripe - like solutions . to calculate the inter - map coupling needed for the hexagonal solution to become the energetic ground state we calculated the potential eq .
( [ eq : potential_septic_part3 ] ) for the three solutions . in case of the uniform solution eq .
( [ eq : uniformsol_part3 ] ) the potential is given by @xmath524 the potential in case of the rhombic and stripe - like solutions was obtained by numerically solving the amplitude equations using newtons method and initial conditions close to these solutions . above @xmath525 the hpwc is energetically preferred compared to stripe - like solutions ( red dashed line in in fig . [
fig : pdcoupled ] ) and thus corresponds to the energetic ground state for large inter - map coupling , see fig .
( [ fig : bif_eps])**. we calculated the phase diagram of the coupled system in the limit @xmath155 , shown in fig . [
fig : pdcoupled ] . , for @xmath155 .
vertical lines : stability range of od hexagons , green line : transition from rpwc to distorted rpwc , red line : stability border of hpwc , blue line : stability border of distorted rpwc . above orange line : hpwc corresponds to ground state of energy . ]
the phase diagram contains the stability borders of the uncoupled od solutions @xmath259 .
they correspond to vertical lines , as they are independent of the inter - map coupling in the limit @xmath155 . at @xmath242 hexagons become stable .
stripe solutions become unstable at @xmath260 . at @xmath261
the homogeneous solution becomes stable while at @xmath262 hexagons loose their stability . in units
@xmath258 the borders @xmath263 vary slightly with @xmath105 , see fig .
[ fig : constterm ] , and are drawn here for @xmath251 .
we rescale the inter - map coupling strength as @xmath526 where @xmath287 is the stationary amplitude of the od hexagons .
the stability borders of op solutions are then horizontal lines . for @xmath527 or for @xmath528 pinwheel free orientation stripes are dynamically selected . for @xmath529 and above a critical effective coupling strength @xmath530 hpwc solutions
are stable and become the energetic ground state of eq .
( [ eq : potential_septic_part3 ] ) above @xmath531 . below @xmath523 ,
rpwc solutions are stable leading to a bistability region between rpwc and hpwc solutions .
we find in this region that rhombic solutions transform into distorted rhombic solutions above an effective coupling strength of @xmath522 .
first , we studied the spatial layout of the rhombic solutions which is illustrated in fig .
[ fig : rpwc ] . .
* * selectivity @xmath6 , white : high selectivity , black : low selectivity . ]
the rpwc solutions are symmetric under rotation by 180 degree .
the rhombic solution has 4 pinwheels per unit cell and its pinwheel density is thus @xmath532 .
one may expect that the energy term eq .
( [ eq : energy_ho ] ) favors pinwheels to co - localize with od extrema . in case of the rhombic layout
there is only one pinwheel at an od extremum while the other three pinwheels are located at od saddle - points which are also energetically favorable positions with respect to @xmath45 .
the orientation selectivity @xmath6 for the rpwc is shown in fig .
[ fig : rpwc]**. the pattern of selectivity is arranged in small patches of highly selective regions .
+ the hexagonal layout of the two stable uniform solutions is shown in fig .
[ fig : hpwc ] . .
* * @xmath533 , * * @xmath534 . * * distribution of orientation preference . *
* op map with superimposed od map for three different values of the od bias . * * selectivity @xmath6 , white : high selectivity , black : low selectivity . * * distribution of intersection angles . ]
the @xmath535 solutions have six pinwheels per unit cell .
their pinwheel density is therefore @xmath536 .
three pinwheels of the same topological charge are located at the extrema of the od map .
two of these are located at the od maximum while one is located at the od minimum .
the remaining three pinwheels are not at an od extremum but near the od border .
the distance to the od border depends on the od bias , see fig .
[ fig : hpwc]**. for a small bias ( @xmath494 ) these three pinwheels are close to the od borders and with increasing bias the od border moves away from the pinwheels .
the pinwheel in the center of the op hexagon is at the contralateral od peak . because these pinwheels organize most of the map while the others essentially only match one op hexagon to its neighbors we refer to this pinwheel crystal as the _ contra - center pinwheel crystal_. note , that some pinwheels are located at the vertices of the hexagonal pattern .
pinwheels located between these vertices ( on the edge ) are not in the middle of this edge
. solutions with @xmath537 are therefore not symmetric under a rotation by 60 degree but symmetric under a rotation by 120 degree .
therefore the solution @xmath538 can not be transformed into the solution @xmath534 by a rotation of the od and op pattern by 180 degrees .
this symmetry is also reflected by the distribution of preferred orientations , see fig .
[ fig : hpwc]**. six orientations are slightly overrepresented .
compared to the ipsi - center pwc , which have a @xmath266 symmetry , this distribution illustrates the @xmath293 symmetry of the pattern .
the distribution of intersection angles is continuous , see fig .
[ fig : hpwc]**. although there is a fixed uniform solution with varying od bias the distribution of intersection angles changes .
the reason for this is the bias dependent change in the od borders , see fig . [
fig : hpwc]**. for all bias values there is a tendency towards perpendicular intersection angles , although for low od bias there is an additional small peak at parallel intersection angles .
the orientation selectivity @xmath6 for the hpwc is shown in fig .
[ fig : hpwc]**. the pattern shows hexagonal bands of high selectivity .
+ finally , we study changes in pinwheel positions during the transition from a rpwc towards a hpwc i.e. with increasing inter - map coupling strength . in case of the higher order gradient - type coupling energy
there is a transition towards a contra - center pwc , see fig .
[ fig : pwwandering]**.
. numbers label pinwheels within the unit cell ( dashed lines ) .
blue ( green , red ) points : pinwheel positions for rpwc ( distorted rpwc , hpwc ) solutions . *
* @xmath539 , using stationary amplitudes from fig .
( [ fig : bif_eps])(a ) .
positions of distorted rpwcs move continuously ( pinwheel 1,3,4 ) . * * @xmath540 , using stationary amplitudes from fig .
( [ fig : amplitudes_tau])**. positions of rpwcs move continuously ( pinwheel 5,6 ) . ] in the regime where the distorted rpwc is stable , three of the four pinwheels of the rpwc are moving either from an od saddle - point to a position near an od border ( pinwheel 1 and 3 ) or from an od saddle - point to an od extremum ( pinwheel 4 ) .
one pinwheel ( pinwheel 2 ) is fixed in space . at the transition to the hpwc two additional pinwheels
are created , one near an od border ( pinwheel 5 ) and one at an od extremum ( pinwheel 6 ) .
we compare the inter - map coupling strength dependent pinwheel positions of the gradient - type coupling energy with those of the product type coupling energy , see fig .
( [ fig : pwwandering])**. in this case three ( pinwheel 2,3,4 ) of the four rpwc pinwheels have a inter - map coupling strength independent position . the remaining pinwheel ( pinwheel 1 ) with increasing inter - map coupling strength splits up into three pinwheels . while one of these three pinwheels ( pinwheel 1 ) is fixed in space the remaining two pinwheels ( pinwheel 5,6 ) move towards the extrema of od .
thus for large inter - map coupling , where hpwc solutions are stable , all six pinwheels are located at od extrema .
we derived amplitude equations and analyzed ground states of the higher order inter - map coupling energies .
we calculated local and global optima and derived corresponding phase diagrams . a main difference between phase diagrams for low order and high order coupling energies
consists in the collapse of orientation selectivity above a critical coupling strength that occurs only in the low order models .
in contrast , for the high order versions , orientation selectivity is preserved for arbitrarily strong inter - map coupling . in order to neglect the backreaction on the dynamics of the modes @xmath134 we assumed @xmath133 .
our results , however , show that for the stability of pinwheel crystals a finite amplitude @xmath134 is necessary . a decrease in @xmath134
can not be compensated by another parameter ( as it would be @xmath104 in case of the low order inter - map coupling energies ) . for a finite @xmath134
even higher order corrections to the amplitude equations than those presented here can thus become significant .
such terms are we neglected in the present treatment . in part ( ii ) of this study we numerically confirm our main results for the higher order inter - map coupling energies . + from a practical point of view , the analyzed phase diagrams and pattern properties indicate that the higher order gradient - type coupling energy is the simplest and most convenient choice for constructing models that reflect the correlations of map layouts in the visual cortex . for this coupling ,
intersection angle statistics are reproduced well , pinwheels can be stabilized , and pattern collapse can not occur . * text s2 : supporting information for + coordinated optimization of visual cortical maps + ( i ) symmetry - based analysis *
here , we list the amplitude equations for the op dynamics in case of the high order inter - map coupling energies @xmath541 and @xmath542 .
@xmath543 where the indices are considered to be cyclic i.e. @xmath544 . all sums are considered to run from 1 to 3 . in the article ,
these amplitude equations are specified in case of od stripes , eq .
( 109 ) , od hexagons , eq . ( 110 ) , or a constant od solution , eq ( 111 ) .
in the following , we list the non - zero elements of the coupling coefficients . @xmath545 @xmath546 note , that coupling coefficients involving the constant shift @xmath128 only occur in case of the product - type inter - map coupling energy . * text s3 : supporting information for + coordinated optimization of visual cortical maps + ( i ) symmetry - based analysis *
here , we state the stability matrices @xmath238 used in the linear stability analysis of the coupled amplitude equations . the stability matrix is defined by + @xmath547 for the uniform solutions
we separate the uncoupled contributions from the inter - map coupling contributions i.e. @xmath548 .
the uncoupled contributions are given by @xmath549 @xmath550 @xmath551 @xmath552 where @xmath553 denotes the @xmath554 identity matrix and @xmath555 .
+ the coupling part in case of the low order product - type inter - map coupling energy is given by @xmath556 @xmath557 . + the coupling part in case of the low order gradient - type inter - map coupling energy is given by @xmath558 @xmath559 . + the stationary amplitudes @xmath560 are given in eq .
( 63 ) , eq .
( 68 ) , eq .
( 81 ) , and eq .
the stationary amplitudes @xmath287 and the constant @xmath128 are given by eq .
( 94 ) and eq .
( 116 ) , respectively .
we thank ghazaleh afshar , eberhard bodenschatz , theo geisel , min huang , wolfgang keil , michael schnabel , dmitry tsigankov , and juan daniel flrez weidinger for discussions . xu x , bosking wh , white le , fitzpatrick d , casagrande va ( 2005 ) functional organization of visual cortex in the prosimian bush baby revealed by optical imaging of intrinsic signals .
j neurophysiol 94 : 27482762 .
hoffsmmer f , wolf f , geisel t , lwel s , schmidt ke ( 1996 ) sequential bifurcation and dynamic rearrangement of columnar patterns during cortical development . in :
bower j , editor , computation and neural systems .
hoffsmmer f , wolf f , geisel t , lwel s , schmidt ke ( 1995 ) sequential bifurcation of orientation and ocular dominance maps . in : proceedings of the international conference on artificial neural networks .
paris : ec2 & cie , volume i , p. 535 .
hoffsmmer f , wolf f , geisel t , schmidt ke , lwel s ( 1997 ) sequential emergence of orientation and ocular dominance maps . in : elsner n , menzel r , editors , learning and memory , proceedings of the 23rd gttingen neurobiology conference 1995 .
stutttgart : thieme verlag , p. 97 .
bonhoeffer t , kim ds , malonek d , shoham d , grinwald a ( 1995 ) optical imaging of the layout of functional domains in area 17 and across the area 17/18 border in cat visual cortex .
eur j neurosci 7 : 19731988 .
kaschube m , wolf f , puhlmann m , rathjen s , schmidt ke , et al .
the pattern of ocular dominance columns in cat primary visual cortex : intra- and interindividual variability of column spacing and its dependence on genetic background .
eur j neurosci 18 : 32513266 . | in the primary visual cortex of primates and carnivores , functional architecture can be characterized by maps of various stimulus features such as orientation preference ( op ) , ocular dominance ( od ) , and spatial frequency .
it is a long - standing question in theoretical neuroscience whether the observed maps should be interpreted as optima of a specific energy functional that summarizes the design principles of cortical functional architecture .
a rigorous evaluation of this optimization hypothesis is particularly demanded by recent evidence that the functional architecture of orientation columns precisely follows species invariant quantitative laws @xcite . because it would be desirable to infer the form of such an optimization principle from the biological data , the optimization approach to explain cortical functional architecture raises the following questions : i ) what are the genuine ground states of candidate energy functionals and how can they be calculated with precision and rigor ?
ii ) how do differences in candidate optimization principles impact on the predicted map structure and conversely what can be learned about an hypothetical underlying optimization principle from observations on map structure ?
iii ) is there a way to analyze the coordinated organization of cortical maps predicted by optimization principles in general ? to answer these questions we developed a general dynamical systems approach to the combined optimization of visual cortical maps of op and another scalar feature such as od or spatial frequency preference . from basic symmetry assumptions we obtain a comprehensive phenomenological classification of possible inter - map coupling energies and examine representative examples .
we show that each individual coupling energy leads to a different class of op solutions with different correlations among the maps such that inferences about the optimization principle from map layout appear viable .
we systematically assess whether quantitative laws resembling experimental observations can result from the coordinated optimization of orientation columns with other feature maps . |
the diffuse neutral interstellar medium ( ism ) is known to exist in two flavors : the cold neutral medium ( cnm ) and the warm neutral medium ( wnm ) .
traditionally , the cnm and wnm are understood as being two thermal equilibrium states of the neutral ism with temperature of @xmath5 k and @xmath6 k @xcite .
in addition , for the cnm the spin temperature , @xmath7 is expected from theoretical considerations .
the situation is more complicated in the case of the wnm and @xmath8 could be lower than @xmath9 @xcite .
observationally , the cnm is easily studied with the 21-cm line absorption .
while the cnm properties have been measured extensively , only three direct measurements of the wnm spin temperature exist to date @xcite .
the main reason for this observational paucity is the low optical depth of the wnm , @xmath10 , which creates a need for very sensitive radio instruments .
properties of the wnm are traditionally indirectly inferred only through hi emission line profiles .
out of all ism phases , the wnm is the least understood ; yet it seems to hold the key for constraining ism models and the formation of cnm clouds .
the theoretical and numerical models of the ism take into account heating and cooling processes and try to reproduce the observed physical properties of the cnm and the wnm .
these properties include temperature distribution , column density probability density functions ( pdfs ) , the gas fraction in each phase , and even profiles of emission and absorption spectra .
@xcite , and similarly @xcite , predict that most of the wnm , which in their model envelopes the cnm , should be in thermal equilibrium and comprise only about 4% of the total column density .
however , observers have traditionally found a much higher fraction for the wnm .
in addition , a significant fraction of the wnm could be thermally unstable . using a statistically meaningful sample and
the arecibo radio telescope , heiles & troland ( 2003 ; ht03 ) found that 48% of the wnm in their survey was in the thermally unstable phase , with kinetic temperatures in the range of 500 @xmath11 5000 k. the ht03 results question the meaning of a two - phase neutral medium and have triggered substantial numerical studies .
numerical models of the ism which include turbulence easily find thermally unstable gas and lead to the conclusion that dynamical processes overshadow the microscopic ones in determining gas temperature .
macroscopical processes are hard to calculate theoretically because they depend on many forms of energy input ( ht03 ) . as a result ,
the predicted column density fraction of the thermally unstable gas varies wildly ; from very large values strongly dependent on the sn rate , as in @xcite , to @xmath12% in @xcite .
[ tab : obs ] [ cols="<,^,^,^,^,^,^,^ " , ] 0.1 in @xmath13 @xmath14 n(hi)@xmath15 was derived assuming @xmath16=100 k and fwhm=3 km s@xmath17 .
+ @xmath13 @xmath18 the total hi column density was provided by ht03 and mohan et al .
+ 0.1 in as mentioned in the previous section , we detected three wide absorption features in the direction of j0022 + 002 , p0347 + 05 and 3c273 , with a velocity fwhm of 7 - 8 km s@xmath17 , and corresponding t@xmath19 of 1100 - 1400 k. for two of these sources , our best - fit solution resulted in @xmath20 400 k. these temperatures lie above the temperature range theoretically allowed for the cnm ( wolfire et al 2003 ) and therefore signify the presence of thermally unstable wnm gas with a temperature @xmath21 k @xcite .
an alternative explanation could be that these broad absorption features represent a blending of several narrow cnm absorption components .
our current velocity resolution of 1.6 km s@xmath17 does not allow us to dismiss this possibility .
future evla observations with the widar correlator will yeild a higher velocity resolution and will allow us to address this question .
the majority of detected absorption lines ( 8 out of 11 ) in this study have a velocity fwhm @xmath22 km s@xmath17 and the corresponding t@xmath23 500 k. while the 3-@xmath24 sensitivity of our observations was good enough to detect a typical wnm ( @xmath25 @xmath26 , ht03 ) in absorption with t@xmath273000 k and a fwhm @xmath28 km s@xmath17 , only two broad wnm components were detected .
both of these detections have a fwhm @xmath29 km s@xmath17 and t@xmath301000 k. this could be an indication that the wnm with a temperature @xmath31 k is not as abundant as indicated previously by indirect observations ( e.g. ht03 ) .
in addition , as we clearly demonstrated in the case of p0347 , the detection of weak hi absorption lines can decrease the number of thermally unstable wnm components .
for the five sources with evla hi absorption detections , previous emission / absorption surveys ( ht03 and @xcite ) found seven thermally unstable wnm components .
our best - fit solutions resulted in five thermally unstable components ( three seen in emission and two in absorption ) .
if applied globally , this would decrease the fraction of thermally unstable wnm components from ht03 s @xmath32 to @xmath33 .
we note , however , that our conclusion is based on a small number statistics and that larger and more sensitive emission / absorption surveys are required to determine the fraction of the thermally unstable wnm in a statistically significant manner . in several cases ( e.g. 3c286 and 3c263.1 ) we detected absorption lines which have a low optical depth and a narrow line - width which result in a low cnm column density n(hi)@xmath34 @xmath26 .
this is significantly lower than the column density of typical cnm clouds , n@xmath35
@xmath26 @xcite , and is close to the lowest theoretical expectations for the cnm @xcite .
in addition , in these directions the cnm comprises only @xmath364% of the total hi column density .
sources without detected hi absorption also suggest a very low cnm fraction ( table [ tab : nodetect ] ) .
this suggests , as also pointed out by stanimirovic , heiles , & kanekar ( 2007 ) , that the low column density cnm population may be quite abundant in the ism .
while such small column densities and cnm fractions are hard to explain by thermal equilibrium ism models , they may be illuminating the tail - end of the cnm population produced by dynamic and turbulent ism processes .
in order to access the feasibility of sensitive hi absorption observations with the evla , we observed a sample of 12 continuum sources , primarily along the lines of sight which have indicated the existence of the thermally - unstable wnm ( t@xmath37 5000 k ) in previous emission / absorption surveys @xcite .
the results from this pilot study are as follows .
@xmath38 for all , except three sources in our sample , we could achieve the expected optical depth sensitivity of @xmath39 .
the remaining three sources were resolved with the evla ; yielding a much lower peak flux density , hence an optical depth sensitivity @xmath40 times higher than the expected value .
@xmath38 out of 12 sources observed with the evla , we detected hi absorption in the direction of 5 sources ; a detection rate of @xmath042 % for our experiment .
seven new absorption features , previously undetected in the emission / absorption surveys of heiles & troland ( 2003 ) and @xcite , were detected in the direction of these 5 sources .
@xmath38 in the case of several sources e.g. 3c273 , p0347 + 05 and j0022 + 002 , we detected wide hi absorption lines ( fwhm @xmath41 km s@xmath17 ) resulting in t@xmath42 k. these lines likely trace the thermally unstable wnm seen in absorption .
@xmath38 several detected absorption features have a peak optical depth of only a @xmath0 few @xmath43 and trace the tail - end of the cnm population .
as we demonstrated , these weak hi absorption lines play an important role in estimating temperature of the wnm and can decrease the indirectly suggested fraction of the thermally unstable wnm .
the national radio astronomy observatory ( nrao ) is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
the arecibo observatory is part of the national astronomy and ionosphere center , which is operated by cornell university under a cooperative agreement with the national science foundation .
we thank greatly lou nigra ( uw - madison ) for kindly obtaining the arecibo hi emission spectrum for one of our sources .
ss acknowledges support by the research corporation for science advancement . | we present 21-cm absorption measurements towards 12 radio continuum sources with previously identified thermally - unstable warm neutral medium ( wnm ) .
these observations were obtained with the expanded very large array ( evla ) and were complemented with the hi emission spectra obtained with the arecibo observatory .
out of 12 sources , hi absorption was detected along 5 lines of sight ( seven new absorption features in total ) , resulting in a detection rate of @xmath0 42 % .
while our observations are sensitive to the wnm with a spin temperature @xmath1 k , we detected only two wide absorption lines with @xmath2 k. these temperatures lie above the range allowed for the cold neutral medium ( cnm ) by the thermal equilbrium models and signify the thermally unstable wnm .
several absorption features have an optical depth of only a few @xmath3 . while this is close or lower than what is theoretically expected for the cnm
, we show that these weak lines are important for constraining the fraction of the thermally unstable wnm .
our observations demonstrate that , for the first time , high bandpass stability can be achieved with the vla , allowing detection of absorption lines with a peak optical depth of @xmath4 . |
recently , brownian ratchets ( motors ) have attracted considered attention simulated by research on molecular motors @xcite . a brownian ratchet , which appeared in feynman s famous textbook for the first time as a thermal ratchet @xcite , is a machine which can rectify thermal fluctuation and produce a directed current .
these subjects are motivated by the challenge to understand molecular motors@xcite , nanoscale friction@xcite , surface smoothening @xcite , coupled josephson junctions @xcite , optical ratchets and directed motion of laser cooled atoms @xcite , and mass separation and trapping schemes at the microscale@xcite .
the brownian ratchet are spatially asymmetric but periodic structure in which transport of brownian particles is induced by some nonequilibrium processes @xcite , such as external modulation of the underling potential or a nonequilibrium chemical reaction coupled to a change of the potential , or contact with reservoirs at different temperatures .
the most intensively studied quantity has been on the velocity of the transport particle . however , another important quantity is efficiency of energy conversion characteristic the operation of the system when the transported particle does work .
how efficiency can brownian ratchets work ?
this question is important not only for the construction of theory of molecular motors @xcite but also for foundation of non - equilibrium statistical physics . like carnot cycle ,
brownian heat engine can extract work from the temperature difference of the heat baths , where brownian working material operates as a transducer of thermal energy into work .
the study of the energetics of brownian ratchets is relevant for several reasons @xcite .
firstly , highly efficient motors are desirable in order to decrease the energy consumption or to decrease the heat dissipation . secondly , brownian motors are related to fundamental problems of thermodynamics and statistical mechanics , such as the maxwell demon and the trade - off between entropy and information .
thirdly , many models proposed in the literature are based on nonequilibrium fluctuations without specifying their source . on the other hand ,
the study of the energetics of such models required a more precise formulation , since one has to determine the physical nature of the external agent and verify that the motor is consistent with the second law of thermodynamics .
recently , sekimoto has been unambiguously defined the concept of the heat at mesoscopic scales in terms of langevin equation @xcite .
he refers to the formalism providing this definition as stochastic energetics .
the essential point of this formalism is that the heat transferred to the system is nothing but the microscopic work done by both friction force and the random force in the langevin equation .
stochastic energetics has been applied to the study of thermodynamic processes .
derenyi and astumian @xcite have studied the efficiency of one - dimensional thermally driven brownian engines .
they identify and compare the three basic setups characterized by the type of the connection between the brownian particle and the two reservoirs : ( 1 ) simultaneous @xcite , ( 2 ) alternating in time @xcite , and ( 3)position dependent @xcite .
parrondo and cisneros @xcite has reviewed the literature the energetics of brownian motors , distinguishing between forced ratchet , chemical motors - driven out of equilibrium by difference of chemical potential , and thermal motors - driven by temperature differences . in this paper we give a detailed study of the last motors - thermal motors .
energetics of the thermal motors - driven by temperature differences are investigated by some previous works @xcite .
asfaw @xciteet al . have explored the performance of the motors under various conditions of practical interest such as maximum power and optimized efficiency .
they found that the efficiency can approach the carnot efficiency under the quasistatic limit .
the same results are also obtained in matsuo and sasa s work @xcite by stochastic energetics theory .
the previous works are limited to case of heat flow via potential .
the present work extends the study to case of heat flow via both potential and kinetics energy .
the motor can work as a heat engine or a refrigerator under various conditions .
the efficiency of heat engine ( cop of refrigerator ) is different from the results of previous works and can never approach the carnot efficiency ( cop ) .
the heat flow via potential is reversible when the engine works at quasistatic limit .
the heat flow via the kinetic is always irreversible in nature .
the model ( shown in fig .
1 ) consists of brownian particles moving in a sawtooth potential with an external load where the medium is alternately in contact with hot and cold heat reservoirs along the space coordinate .
let @xmath0 be the barrier height of the potential .
@xmath1 , @xmath2 are the length of the left part and the right part of the ratchet , respectively . @xmath3 is the energy needed for a forward jump , while @xmath4 is the energy required for a backward jump
. the left part of a period ratchet is at temperature @xmath5 ( hot reservoir)and the right one is at temperature @xmath6 ( cold reservoir ) .
we can assumes that the rates of both forward and backward jumps are proportional to the corresponding arrhenius factor @xcite , such that , @xmath7 is a piecewise linear and periodic potential .
@xmath8 is a potential due to the load .
the period of the potential is @xmath9 .
the temperature profiles are also shown.,width=302,height=226 ] @xmath10},\ ] ] @xmath11},\ ] ] are the number of forward and backward jumps per unit time , respectively , @xmath12 the boltzmann constant , @xmath13 a proportionality constant with dimensions of time .
if @xmath14 , the ratchet works as a two - reservoir heat engine shown in fig .
2 . , @xmath15 is power output , heat flows via potential energy is reversible , heat flows via kinetic energy is irreversible.,width=302,height=151 ] the rate of heat flow from hot reservoir to the heat engine via potential is given by @xmath16 the rate of heat flow from the engine to the cold reservoir via potential is @xmath17 the heat flow via the kinetic energy of the particle is much more complicated to determined @xcite .
whenever a particle stay at a hot segment ( temperature @xmath5 ) it absorbs @xmath18 energy on average from the hot reservoir
. it can pick up @xmath19 energy on average from the cold reservoir when the particle stay at a cold segment .
it is obvious that when a particle leaves from a hot segment to a cold segment and then returns to the hot segment , the hot reservoir will lost @xmath20 energy on average , the lost energy is released to the colder reservoir and never to the hot reservoir or to the particle s potential energy , which indicates the inherently irreversible nature of this heat flow . on the other hand ,
if a particle goes out from a hot segment to a cold segment and never returns to the hot segment , the hot reservoir will lost @xmath18 energy on average
. therefore the rate of net heat flow via kinetic energy from the hot reservoir to the cold reservoir is given by @xmath21 therefore , the rate of total heat transferred from the hot reservoir is given by @xmath22 the rate of total heat transferred to the cold reservoir is @xmath23 the power output is @xmath24 it is easy to obtain the efficiency of the heat engine @xmath25 if the heat flow via the kinetic energy is not considered , the efficiency is given by @xmath26 in order to discuss simply , we can rewritten eq .
( 5 - 10 ) in a dimensionless form , then we get @xmath27,\ ] ] @xmath28,\ ] ] @xmath29,\ ] ] @xmath30f,\ ] ] @xmath31f}{(\epsilon+\mu f+\frac{1}{2})e^{f_{0}-\mu f}-(\epsilon+\mu f+\frac{1}{2}\tau)e^{(1-\mu)f/\tau}},\ ] ] @xmath32 where @xmath33 from the above equations , one has @xmath34 , @xmath35 , @xmath36 .
since the motor acts as a heat engine , the power output can not be negative , namely , @xmath37 .
one can also obtain the value range of @xmath38 , @xmath39 , where @xmath40 therefore , it is easy to obtain @xmath41 where @xmath42 is carnot efficiency .
however , @xmath43 attains the carnot efficiency , namely , @xmath44 which indicates that the power output is zero . from eq .
( 15 ) and eq .
( 16 ) , it obvious that @xmath45 is always less than @xmath43 .
the results are given by fig .
3 - 10 . ,
@xmath46 and power output @xmath47 vs the load @xmath38 at @xmath48,@xmath49,@xmath50 . ,
width=302,height=226 ] figure 3 shows that the heat flow @xmath51 , @xmath46 and power output @xmath47 as a function of the load @xmath38 .
when @xmath52 , namely , the engine runs without a load , @xmath51 is equal to @xmath46 , which indicates that the heat that absorbs from the hot reservoir releases to the cold reservoir entirely and no power output is obtained .
when @xmath38 increases , @xmath51 and @xmath46 decreases . when @xmath53 , no power output is obtained ( @xmath47=0 ) .
when @xmath54 , no currents occur , the ratchet ca nt give any power output .
so the power output @xmath47 has a maximum value at certain value of @xmath38 which depends on @xmath55 , @xmath56 and @xmath57 .
, @xmath58 and @xmath45 vs the load @xmath38 at @xmath48,@xmath49,@xmath50.,width=302,height=226 ] figure 4 shows the variation of the efficiency @xmath42 , @xmath58 , @xmath45 with the load @xmath38 .
if the heat flow via kinetic energy is ignored , the efficiency @xmath58 increases with the load @xmath38 , it approaches the carnot efficiency @xmath42 at quasistatic condition ( @xmath38=@xmath59 ) .
the result is also presented in asfaw s work @xcite .
when the heat flow via kinetic energy are considered , the curve of the efficiency @xmath45 is observed to be bell - shaped , a feature of resonance .
the efficiency @xmath45 is always less than @xmath58 and never approaches carnot efficiency @xmath42 .
it is obvious that the heat flow via kinetic energy is always irreversible and energy leakage is inevitable , so the efficiency is less than @xmath58 and ca nt approach carnot efficiency .
vs the load @xmath38 ( @xmath60 , @xmath49 , @xmath61 ) ; ( b)dimensionless heat flow @xmath62 vs barrier height @xmath56 for different values of @xmath63 at @xmath60 , @xmath61.,title="fig:",width=302,height=226 ] vs the load @xmath38 ( @xmath60 , @xmath49 , @xmath61 ) ; ( b)dimensionless heat flow @xmath62 vs barrier height @xmath56 for different values of @xmath63 at @xmath60 , @xmath61.,title="fig:",width=302,height=226 ] figure 5a shows the heat flow @xmath62 out of hot reservoir via kinetic energy as a function of the load @xmath38 .
when @xmath64 , @xmath62 is positive .
when @xmath65 , @xmath62 is negative .
no heat flow occurs at @xmath38=@xmath59 .
it is found that the heat flow via kinetic is dependant on the current of the ratchet .
figure 5b shows the heat flow out of the hot reservoir via kinetic energy as a function of barrier height @xmath56 .
the curve is bell - shaped .
therefore , there is an optimized value of @xmath56 at which @xmath62 takes its maximum value . vs the load @xmath38 for different values of @xmath66 at @xmath48 , @xmath67.,width=302,height=226 ] figure 6 shows the power output @xmath47 as a function of the load @xmath38 for different values of @xmath66 . from the figure
, we can see that the power output has a maximum value at certain value of @xmath38 .
the maximum load @xmath59 of the engine changes with the parameter @xmath57 of asymmetry in potential . in other word
, the structure of the potential determines the maximum load capability of the engine . vs the load @xmath38 for different values of @xmath66 at @xmath48 , @xmath67.,width=302,height=226 ] figure 7 shows the variation of the efficiency @xmath45 with the load @xmath38 for different values of @xmath68 .
the maximum value of @xmath45 increases with @xmath57 .
however , the maximum value of @xmath45 ca nt approach the carnot efficiency . vs barrier height @xmath56 for different values of @xmath38=0.1 , 0.2 , 0.5 at @xmath48 , @xmath50.,width=302,height=226 ] figure 8 shows the power output @xmath47 as a function of the barrier height @xmath56 for different values of the load @xmath69 .
when @xmath70 , the effect of ratchet disappears , the power output tends to zero .
when @xmath71 so that the particle ca nt pass the barrier , the power output @xmath47 goes to zero , too .
the power output @xmath47 has a maximum value at certain value of @xmath56 which is dependant on @xmath55 , @xmath57 and @xmath38 . on the other hand , the minimum height of the barrier for working as a heat engine increases with the load @xmath38 .
, @xmath58 and @xmath45 vs the barrier height @xmath56 at @xmath48 , @xmath72 , @xmath50.,width=302,height=226 ] figure 9 shows the efficiency @xmath42 , @xmath58 and @xmath45 as a function of the barrier height .
the efficiency @xmath58 without the heat flow via kinetic energy approaches the carnot efficiency at @xmath56=@xmath73 , at which no power output occurs .
the efficiency @xmath45 is a peaked function of the barrier height @xmath56 which is dependant on values of @xmath55 , @xmath38 and @xmath57 .
, @xmath46 and power output @xmath47 vs @xmath55 at @xmath74 , @xmath67 , @xmath61.,width=302,height=226 ] figure 10 shows plot of @xmath47 , @xmath51 and @xmath45 versus @xmath55 . from the figure
, we can see that @xmath47 , @xmath51 and @xmath45 change very slowly at small @xmath55 and they decreases drastically near @xmath75 .
when the load @xmath38 is less than the maximum load @xmath59 , the motor works as a heat engine .
the power output is a peaked function of the load @xmath38 and the barrier height @xmath56 .
the efficiency @xmath45 is less than the efficiency @xmath58 and can never approach the carnot efficiency @xmath42 .
the heat flow via kinetic energy causes the energy leakage unavoidably and reduces the efficiency of the heat engine .
if the load is large enough along with appropriately chosen other quantities , the motor will run in the reverse direction @xmath76 . the motor will absorb the heat from the cold reservoir and release it to the hot reservoir . under this condition , the thermal motor acts as a refrigerator shown in fig . 11 . , @xmath15 is power input , heat flow via potential energy is reversible , heat flow via kinetic energy is irreversible.,width=302,height=151 ] similarly , the rate of total heat transferred to the hot reservoir is given by @xmath77 the rate of total heat transferred from the cold reservoir to the heat engine is given by @xmath78 the power input is @xmath79 the above equations can also be rewritten in a dimensionless form with eq .
( 17 ) @xmath80,\ ] ] @xmath81,\ ] ] @xmath82f.\ ] ] for a refrigerator , @xmath83 , one can get @xmath84
. the heat flow from cold reservoir must not be negative ( @xmath85 ) , which is important feature of a refrigerator .
as for a refrigerator , its generalized cop is most important , which is given by @xmath86x}\ ] ] if the heat flow via the kinetic energy is not considered , the cop is given by @xmath87 where @xmath88 is cop of a carnot refrigerator .
when @xmath89 attains the @xmath88 , namely , @xmath44 , the power input is zero .
the results are shown in fig .
12 - 14 .
, @xmath46 and power input @xmath47 vs the load @xmath38 at @xmath90 , @xmath49 , @xmath61.,width=302,height=226 ] figure 12 shows @xmath91 , @xmath46 and @xmath47 as a function of the input force @xmath38 .
the power input @xmath47 increases with the input force @xmath38 . the heat @xmath46 that absorbs from the cold reservoir has a maximum value at certain value of input force @xmath38 . the cop @xmath91 of the refrigerator takes its maximum value at which the power input tends to zero . , @xmath89 and @xmath91 vs input force @xmath38 at @xmath92 , @xmath49 , @xmath61.,width=302,height=226 ] figure 13 gives the plot of cop @xmath88 , @xmath89 and @xmath91 versus the input force @xmath38 .
when the heat flow via kinetic energy is ignored , the cop of the refrigerator @xmath89 decreases with increasing of @xmath38 and it approaches the carnot cop at @xmath44 . when the heat flow via kinetic energy is considered , the curve of cop @xmath91 is bell - shaped and it can never approach the carnot cop @xmath88 .
the cop @xmath91 without the heat flow via kinetic energy is less than @xmath89 when @xmath93 , larger than @xmath89 when @xmath94 and equal to @xmath89 at @xmath95 .
however , @xmath91 can never approach the carnot cop @xmath88 .
vs the input force @xmath38 for different values of @xmath96 at @xmath97 , @xmath61.,width=302,height=226 ] figure 14 shows the variation of the cop @xmath91 with the input force @xmath38 for different values of the @xmath98 . from the figure , it is easy to see that the cop @xmath91 increases with the value of @xmath55 and the minimum input force @xmath59 increases with the decreasing of the value of @xmath55 . in an other word , when temperature difference is small , for example , @xmath90 , the heat leakage is small , so the cop @xmath91 is large . on the contrary ,
when the minimum input force is large , the temperature difference is large , for example , @xmath99 . when @xmath38 is so large that the motor runs reversibly , @xmath38 becomes the input force , so the motor works as a refrigerator .
the cop @xmath91 is a peaked function of input force @xmath38 which indicates the the cop can not obtain a maximum value at quasi - static limit .
the cop of refrigerator can never approach the carnot cop @xmath88 because of the heat flow via kinetic energy .
in present work , we study the energestic of a thermal motor which consists of brownian particles moving a sawtooth potential with an external load where the viscous medium is alternately in contact with hot and cold reservoirs along the space coordinate
. the thermal motor can works as both a heat engine and a refrigerator .
we make a clear distinction between the heat flow via the kinetic and the potential energy of the particle , and show that the former is always irreversible and the latter may be reversible when the engine works quasistatically . when the external load is not enough , the thermal motor can work as a heat engine .
the power output has a maximum value at certain value of the load @xmath38 which dependant on the others parameters .
if only the heat flow via potential is considered , the efficiency @xmath58 increases with the load @xmath38 and approaches the carnot efficiency @xmath42 under quasistatic condition , which agrees with the previous work .
when the heat flow via kinetic energy is also considered , the efficiency @xmath45 is less than @xmath58 and can never approach the carnot efficiency @xmath42 .
it is also found that the structure of the potential decides the maximum load capability of the heat engine .
the heat flow via potential is reversible , while the heat flow via kinetic energy is irreversible .
the heat flow via kinetic energy reduces the heat engine efficiency . when the external load is so large that the motor moves reversely , the thermal motor can work as a refrigerator .
when the heat flow via kinetic energy is ignored , the cop @xmath89 decreases with increasing of the input force @xmath38 and can approach the carnot cop @xmath88 .
when the heat flow via both potential and kinetic energy are considered , the cop @xmath91 is less than @xmath89 when @xmath93 , larger than @xmath89 when @xmath94 and equal to @xmath89 at @xmath95 .
however , @xmath91 can never approach the carnot cop @xmath88 .
the thermal motor was initially proposed in an attempt of describing molecular motors in biological systems .
the heat flow via energy is irreversible , so the efficiency ( cop ) can not approach the the carnot efficiency ( cop ) at a quasistatic condition . however , molecular motors are known to operator efficiency @xcite .
they can convert molecular scale chemical energy into macroscopic mechanical work with high efficiency in water at room temperatures , where the effect of thermal fluctuation is unavoidable .
thus , the next challenging question would be `` how to reduce the heat flow via kinetic energy ? '' 0.2 in r.
d. astumian and p. hanggi , physics today ( 2002)33 .
p. reimann , phys .
361 ( 2002 ) 57 .
r. p. feynman , r. b. leighton , m. sands , the feynman lectures on physics , vol .
i , addison - wesley , reading , 1963 .
j. maddox , nature ( london ) 368 287(1994 ) ; r. d. astumian and i. derenyi , eur .
j. 27 , 474 ( 1998 ) ; n. thomas and r. a. thornhill , j.phys .
d 31 , 253 ( 1998 ) ; c.r . doering , b. ermentrout , g. oster , biophys .
j. 69 , 2256(1995 ) ; j. krim , d.h .
solina , r. chiarello , phys . rev
66 ( 1991 ) 181 ; j.b .
sokolo2 , j. krim , a. widom , phys . rev .
b 48 ( 1993 ) 9134 ; l. daikhin , m. urbakh , phys . rev .
e 49 ( 1994 ) 1424 ; c. daly , j. krim , phys .
76 ( 1996 ) 803 ; m.r .
sorensen , k.w .
jacobsen , p. stoltze , phys .
b 53 ( 1996 ) 2101
. i. derenyi , c. lee , and a .-
barabasi , phys .
80 ( 1998 ) 1473 .
i. zapata , r. bartussek , e. sols , p. hfanggi , phys .
77 , 2292 ( 1996 ) .
faucheux , l.s .
bourdieu , p.d .
kaplan , a.j .
libchaber , phys .
lett . 74,1504 ( 1995 ) ; c. mennerat - robilliard , d. lucas , s. guibal , j. tabosa , c. jurczak , j .- y .
courtois , g. grynberg , phys .
82 , 851(1999 ) .
l. gorre - talini , j.p .
spatz , p. silberzan , chaos 8 , 650 ( 1998 ) ; i. derenyi , r.d .
astumian , phys .
e 58 , 7781 ( 1998 ) ; d. ertas , phys . rev .
80 , 1548 ( 1998 ) ; t.a.j .
duke , r.h .
austin , phys .
80 , 1552 ( 1998 ) .
j. l. mateos , phys .
84 , 258 ( 2000 ) .
j. rousselet , l. salome , a. adjari and j. prost , nature 370 ( 1994 ) 446 .
l. p. faucheux et al . ,
74 ( 1995 ) 1504
. b. q. ai , x. j. wang , g. t. liu and l. g. liu , phys .
e , 68 , 061105 ( 2003 ) ; b. q. ai , x. j. wang , g. t. liu and l. g. liu , phys .
e , 67 , 022903 ( 2003 ) ; b. q. ai , g. t. liu , h. z. xie , l.g .
liu , chaos 14(4),957 ( 2004 ) ; b. q. ai , x. j. wang , g. t. liu , h. z. xie , d. h. wen , w. chen , and l. g. liu .
j. b 37 , 523 - 526 ( 2004 ) .
f. julicher , a. ajdari and j. prost , rev .
69 , 1269 ( 1997 ) .
j. m. r. parrondo and b. j. de .
cisneros , app .
phys . a 75 , 179 ( 2002 ) .
k. sekimoto , j. phys .
japan 66 ( 1997)1234 ; k. sekimoto and s. sasa , j. phys .
japan 66 3326 ( 1997 ) ; k. sekimoto , progr .
phys . suppl .
130 , 17 ( 1998 ) .
i. derenyi and r. d. astumian , phys .
e 59 , r6219 ( 1999 ) .
p. peimann , r. bartussek , r. haussler , and p. hanggi , phys .
lett . a 215 , ( 1996)26 ; i. m. sokolov and a. blumen , j. phys .
a 30 , 3021 ( 1997 ) .
m. buttiker , z. phys .
b 68 , 161 ( 1987 ) ; r. landauer , j. stat .
53 , 233 ( 1988 ) .
m. asfaw and m. bekle , eur .
j. b 38 , 457 ( 2004 ) .
m. matsuo and s. sasa , physica a 276 , 188 ( 2000 ) .
s. velasco , j. m. m. roco , a. medina and a. calvo her nandez , j. phys .
34 1000 ( 2001 ) .
t. yanagida , t. arata , and f. oosawa , nature 316 , 366 ( 1985 ) .
a. ishijima , h. kojima , h. higuchi , y. harada , t. funatsu , and t. yanagida , biophys .
j. 70 , 383 ( 1996 ) .
uyeda , s. j. kron , and j. a. spudich , j. mol .
214 , 699 ( 1990 ) .
f. takagi and t. hondou , phys .
e 60 , 4954 ( 1999 ) .
r. yasuda , h. noji , k. kinosita , f. motojima , and m. yoshida , j. bioenerg .
29 , 207 ( 1997 ) . | we study the energestics of a thermal motor driven by temperature differences , which consists of brownian particles moving in a sawtooth potential with an external load where the viscous medium is alternately in contact with hot and cold heat reservoir along space coordinate .
the motor can work as a heat engine or a refrigerator under different conditions .
the heat flow via both potential and kinetic energy are considered .
the former is reversible when the engine works quasistatically and the latter is always irreversible .
the efficiency of the heat engine ( coefficient of performance ( cop ) of a refrigerator ) can never approach carnot efficiency ( cop ) . |
the formation history of galaxies has been extensively studied in terms of the evolution of their stellar mass @xcite .
however , it is still poorly constrained how galaxies developed their stellar systems in the early universe at @xmath11 .
while the observed stellar mass density of the galaxies shows significant and rapid growth along the time at @xmath12 , at least @xmath13 % of the present - day stellar mass in the galaxies is observed at @xmath7 @xcite , which seems also true for the massive galaxies with @xmath14 @xcite . where and how were they formed ?
the biased galaxy formation models in a universe dominated by cold dark matter suggest that galaxy formation preferentially occurs in regions with relatively high density at the larger scale @xcite .
recent results based on optical wide - field imaging reveal that high - redshift star - forming galaxies indeed show very strong clustering largely biased to the expected underlying mass distribution ( e.g. , @xcite , ) , which is also true for near - infrared ( nir)-selected massive galaxies @xcite . at @xmath15 , individual large - scale high - density regions of the galaxies have also been discovered @xcite .
many of these fields are , however , identified as an excess in the number density of rest - frame uv bright galaxies , such as ly@xmath8 emitters ( laes ) , lyman break galaxies ( lbgs ) or their analogues .
the question is how much of the stellar mass has already formed and assembled in such high density region of star - forming galaxies at @xmath16 . in other words , we would like to see whether the structure traced by the star formation activity is correlated ( or anti - correlated ) with the structure traced by the stellar mass . in this paper , we present the results of @xmath17- , @xmath18- and @xmath10-band imaging of the @xmath0 proto - cluster region in and around the ssa22 field ( @xcite , ; @xcite ) using multi - object infrared camera and spectrograph ( moircs ) @xcite equipped with the 8.2 m subaru telescope .
the wide field of view of moircs allowed us to observe such proto - cluster regions efficiently .
the proto - cluster was first discovered by @xcite as the excess of lbgs .
later wider - field and deeper narrow - band observations revealed that the overdense region of lbgs is a part of a large - scale high density structure of laes , which is extended over @xmath19 60 mpc in a comoving scale @xcite .
@xcite confirmed that the structure has a three - dimensional filamentary appearance .
@xcite also identified 35 lyman@xmath8 blobs ( labs , extended , bright ly@xmath8-emitting clouds ) , along this structure .
the large nebulae typically have ly@xmath8 luminosity @xmath20-@xmath21 ergs s@xmath22 and physical extent 30 - 150 kpc .
the most luminous and extended labs are the two giant nebulae first discovered by @xcite ( hereafter referred as lab1 and lab2 , following @xcite ) . rather than studying the differences in their distribution from only the uv - selected galaxies , observing the star forming regions in nir ,
the rest - frame optical wavelengths puts more weight in their stellar mass content , allowing us to study the mass and population of the stars associated with them .
the nature of dusty star - forming objects in such high - density structure can also be examined by the nir data .
the field observed with moircs ( hereafter referred as ssa22-m1 ) corresponds to the southern part of ssa22a @xcite and is located within the field observed by @xcite . in ssa22-m1 , there are 8 labs @xcite , 16 laes @xcite , and 17 lbgs @xcite , of which 11 are located at @xmath0 .
the sample selection for laes is overlapped with that for labs . while laes are selected by the aperture photometry of the detected sources @xcite , labs are selected by their isophotal area on the narrow - band image reduced with the optimum sky subtraction @xcite .
some labs are associated with laes if they have enough compact high - surface brightness cores . on the other hand , hayashino et al .
( 2004 ) excluded the laes detected in lab1 and lab2 from their sample as they are clearly un - isolated parts of diffuse ly@xmath8 nebulae .
two of the 16 laes in ssa22-m1 are identified with labs ( lab16 and lab31 , referred by @xcite ) .
we describe the observation and the data analysis method in section 2 . the number density and the sky distribution of distant red galaxies ( drgs , @xcite ) , the result of photometric redshift analysis of the @xmath6-selected sources , and the photometric properties of labs , lbgs , and laes
are shown in section 3 . in section 4
we discuss our results in terms of the mass assembly history in the proto - cluster .
we use the cosmological parameter values @xmath23 , @xmath24 , and @xmath25=70 km s@xmath22 mpc @xmath22 throughout this paper .
all the magnitude values are in _ ab _ system @xcite unless explicitly noted .
@xmath17- , @xmath18- , and @xmath10-band images of the field centered at ( @xmath26)@xmath27= ( 22h 17 m 33.9s , @xmath28 ) were obtained using moircs mounted on the subaru telescope on june 15 , 16 and august 14 in 2005 , and july 23 in 2006 ( ut ) .
moircs uses two 2048@xmath292048 hawaii-2 arrays with the pixel scale @xmath30 pixel@xmath22 and the field of view @xmath31 .
the images are reduced in a standard manner with the mcsred software packages ( i. tanaka et al .
2007 , in preparation ) . for the @xmath10-band flat - fielding
, we use on / off sets of domeflat images to avoid the thermal emission seen at the edge of the frames due to the foreground telescope structure .
the median - sky for each frame is subtracted before the frames are stacked .
if notable diffuse residual patterns remain after the median - sky subtraction , we further subtracted them by using the iraf imsurfit command .
we also subtract fringe patterns in each frame , if they appear , using the templates made from the images taken in the same run .
the total exposure times of the final images are 3,240 s , 3,600 s , 2,235 s for @xmath17 , @xmath18 , @xmath10 , respectively .
the effective area of the image is 21.7 arcmin@xmath1 .
the stellar image sizes are @xmath32 , @xmath33 , @xmath34 for @xmath17 , @xmath18 , @xmath10 , respectively .
the limiting magnitudes are @xmath35 , @xmath36 , and @xmath37 with 5 @xmath5 in a @xmath38 diameter aperture .
we calibrated our nir data to the ukirt photometric system @xcite .
we convert between the vega system to the ab system using @xmath39 , @xmath40 , @xmath41 .
the summary of the observations is shown in table [ tab1 ] .
for the detection and photometry , we use sextractor ( version 2.3 ) @xcite .
the objects which have more than 16 connected pixels with 1.5 @xmath5 above the background noise are selected .
the @xmath42 color is measured in a @xmath43 diameter aperture .
the sextractor mag auto value is adopted as the @xmath6-band pseudo total magnitude .
the @xmath10-band image is smoothed to be matched with the psf in @xmath17-band for @xmath42 color measurement .
the completeness in @xmath10-band is 95@xmath44 at @xmath45 and 50 @xmath44 at @xmath4 for a point source .
we also use the optical @xmath46 bands data obtained by @xcite and @xmath47-band data @xcite to estimate the photometric redshift . as the fwhm of the @xmath47-band image
is @xmath48 , we smoothed the @xmath49-band images to be matched and measured the colors in a @xmath50 aperture .
narrowband ( nb497 ; 4977 , fwhm 77 ) images used in section [ sec : lab ] are taken in @xcite .
we corrected our data for the galactic extinction adopting the average value at the field , @xmath51 @xcite .
cccc filter & integration & fwhm & limiting mag .
+ & s & arcsec & mag + @xmath17 & 3,240 & 0.51 & 24.5 + @xmath18 & 3,600 & 0.47 & 24.3 + @xmath10 & 2,235 & 0.44 & 23.9 +
in this section , we present the observed properties of the @xmath6-selected objects in ssa22-m1 . since we are studying the galaxies at @xmath0 , we need to apply some photometric selections to suppress the contamination of foreground / background objects . while full photometric redshift analysis is useful , it may suffer from misidentification especially for @xmath6-band faint objects because the optical data used here may not be deep enough for them to achieve high photometric accuracy .
simple color - cut method such as selecting drgs with @xmath52 @xcite works complementary , so that we can more completely pick - up objects possibly located at @xmath0 , if they exist .
we therefore first investigate the distribution of drgs and then that of the photometric redshift ( photo - z ) selected sources to compare with the uv - selected objects in the field .
we first focus on the distribution of drgs in our field .
drgs are supposed to be evolved , relatively massive galaxies ( @xmath53 ) at @xmath54 @xcite or dusty galaxies at @xmath55 @xcite . in the analysis of the deep multicolor data obtained at goods - n
, @xcite found that 83% of the drgs with @xmath56 have the photometric redshift @xmath11 , while more than a half of the drgs with @xmath57 have @xmath58 .
thus the faint drgs are supposed to be dominated by the galaxies at high redshift . as @xcite suggested that the global stellar mass at @xmath59 appears to be dominated by drgs , the number of drgs might be a good indicator of massive galaxies at @xmath16 .
in fact , @xcite showed that drgs dominate the high mass end of the mass function at high redshift .
we use the same criteria as in @xcite , which corresponds to @xmath60 in our photometric system .
the red @xmath42 color of drgs is due to the balmer or 4000 break of galaxies at @xmath61 , or due to the dust extinction @xcite .
the color @xmath62 corresponds to that of a galaxy at @xmath0 with the stellar population age older than 200 myr if no extinction and the solar metallicity are assumed @xcite .
@xcite reported the median age and stellar mass for drgs with @xmath63 in hdf - s are 1.7 gyr and @xmath64 .
the @xmath6-band selected drgs are supposed to be relatively more massive , more evolved objects than ultra - violet(uv)-selected galaxies like lbgs at the same epoch . in fig .
[ fig_colmag ] , we show the @xmath6 versus @xmath42 color - magnitude diagram of the @xmath6-selected objects .
the objects with @xmath65 are likely to be galactic stars .
the color of the sequence coincides with that seen in the other field @xcite , which suggests that our photometric calibration is robust .
the 2@xmath5 limiting magnitude in @xmath17-band is indicated by the solid line in fig .
[ fig_colmag ] .
the @xmath17-band image is deep enough to select drgs down to @xmath66 .
the distribution of the photometric redshifts for drgs is shown in fig .
[ fig_photozp50 ] .
the detailed photo - z analysis is described in section [ sec : photoz ] .
we find that 62 % of drgs are classified as @xmath67 .
the cumulative number of drgs in our observed field is 28 and the surface density is 1.29 arcmin@xmath68 at @xmath69 ( or @xmath70 ) , which is comparable with that in goods - n by @xcite , 1.36 @xmath71 0.26 arcmin@xmath68 to the same depth .
[ fig_drg_density ] compares the differential number counts for drgs in our observed field with those in goods - n .
our sample is dominated by drgs with @xmath72 , which implies most of them should be at high redshift .
the completeness fraction and the 95% limit , @xmath45 , are also marked .
we do not see any notable excess in the number counts in ssa22-m1 over the magnitude range .
we note that there are nine hyper extremely red objects ( heros , @xcite ) with @xmath73 in ssa22-m1 at @xmath74 .
the surface density of heros at @xmath74 is @xmath75 arcmin@xmath22 which is consistent with those in sdf , hdf - s , and goods - n within the error @xcite .
the sky distribution of drgs is shown in fig .
[ fig_geo]a , and labs @xcite , laes @xcite , and lbgs @xcite are plotted in fig .
[ fig_geo]c .
we then investigated the spatial correlation between drgs and other uv - selected objects in the field .
[ fig_box]a shows the mean surface number density of drgs as a function of a distance from labs , laes , and lbgs .
first , we note that the distribution of the drgs is weakly correlated with the sample of labs as well as lbgs .
since the lbgs are associated with 7 of 8 labs , it is not surprising that the similar correlations with drgs are seen for these two populations .
we also plotted the profile from the two distinctive giant labs , namely lab1 and lab2 and found that the drg density shows the notable excess within 1.0 arcmin ( 1.9 mpc at @xmath0 in comoving scale ) around them .
in fact , the correlation between drgs and labs is found to be dominated by the two objects ; if we remove them from the sample , the correlation seen in fig .
[ fig_box]a diminishes significantly . as mentioned in the beginning of this section ,
the redshift range of the drg criteria is broad and there may be a chance projection of interlopers at intermediate redshift .
we made a simple test to see if the density excess of drgs around lab1 and lab2 is significant by evaluating the probability of finding similar excess around the eight randomly placed objects .
we find the similar profile in 592 among the 10000 trials , which gives the probability of @xmath76 in random distribution of lab to the observed distributions drgs .
thus the association of lab1/lab2 with drgs is marginally significant ( @xmath77 ) .
we investigate the closer look for the @xmath6-selected objects near the eight labs in the field in section [ sec : lab ] .
we also find that there is no significant positive correlation between drgs and laes , while a hint of anti - correlation is seen at @xmath78 separation .
it should be noted that there is such anti - correlation among the massive drgs and less massive laes . using the @xmath79 photometric data , we estimated the photometric redshift of all the @xmath6-selected sources using the hyperz code @xcite .
the photometric redshift method performs the sed fitting with the redshift , spectral type , age , and dust extinction as free parameters .
the best fit sed with its redshift is determined from the minimum @xmath80 value .
the template spectrum we use are derived from galaxev , which is the library of evolutionary stellar population synthesis models by @xcite . at @xmath7 ,
the uncertainty of the photo - z is typically @xmath81 @xcite .
[ fig_speczphotoz ] shows the resulting photometric redshifts versus spectroscopic redshifts obtained by previous studies ( @xcite ) .
the relative errors @xmath82 in the redshift estimation are also shown in the bottom panel of the figure .
the mean of relative errors for lbgs at @xmath83 is @xmath84 and the standard deviation is 0.08 ( also see table [ tab3 ] in section 3.3 ) .
we therefore pick up the objects with @xmath85 as the candidate galaxies at @xmath0 .
we found 29 objects of which 9 are classified as drgs and 6 are classified as lbgs .
[ fig_geo]b shows a sky plot of the @xmath6-selected sources , similar to fig .
[ fig_geo]a , but for those with @xmath85 and @xmath74 .
four lbgs associated with labs are not appeared since they are fainter than the limit in @xmath6-band .
cross correlation with labs , laes , and lbgs for this sample is also shown in fig .
[ fig_box]b .
we find that the association between these @xmath6-selected sources at @xmath86 and labs , especially lab1 and lab2 , seems significant .
in contrast to drgs , labs excluding the two giant blobs show the association with the photo - z selected objects . in this section
, we investigate the nir counterparts of the individual labs .
we find that the 88% of the labs have plausible @xmath10-band candidates .
@xmath10-band images in @xmath87-side box for 8 labs are shown in fig .
[ fig_lab1lab2 ] and fig .
[ fig_lab7to31 ] .
the green lines indicate the @xmath88-band detection isophoto contours @xcite and @xmath89-band sources are indicated with the red contours in the figures .
lbgs in the field @xcite are also marked .
of 11 lbgs at @xmath0 , 8 are associated with labs . of these , ssa22a - c11 ( steidel et al .
2003 , associated with lab1 ) , m14 ( lab2 ) , m4 ( lab7 ) , c12 ( lab20 ) , d3 ( lab30 ) , and c4 ( lab31 ) are detected in the @xmath10-band , while c6 ( lab7 ) and c15 ( lab8 ) are not .
photometric properties of @xmath6-selected objects around labs are summarized in table [ tab2 ] and the evaluated photometric redshift and physical properties of the candidates of @xmath0 galaxies are presented in table [ tab3 ] . the stellar age , dust extinction , absolute magnitude , and stellar mass
are derived from the sed fitting with galaxev @xcite by assuming @xmath0 .
we assume the @xcite initial mass function ( imf ) , the @xcite extinction law , and a solar metallicity .
the characteristic time scale is set at @xmath90 gyr .
the star formation timescale , the color excess , and the age are varied as free parameters and the best fit sed is determined from the minimum reduced @xmath80 value .
the mass - to - light ratio @xmath91 and the total absolute magnitude ( @xmath92 ) derived from galaxev are used for calculating the stellar mass .
the errors of the stellar mass indicate the confidence level of 68 % . for lab1
, there are five @xmath6-selected sources within the ly@xmath8 nebula . lab1-#1 and # 2 are likely to be foreground objects as their photometric redshifts are @xmath93 .
@xcite already reported that there is the @xmath10-band counterpart candidate with extremely red @xmath94 color near the center of the ly@xmath8 nebula ( lab1-#3 in fig .
[ fig_lab1lab2 ] ) .
we find that lab1-#3 is classified as a drg and lab1-#3 and # 4 have @xmath85 . while their spectroscopic redshifts are still unknown , it is very likely that # 4 associates the lab as it is located right at the hole of ly@xmath8 nebulae ( the right panel in fig .
[ fig_lab1lab2 ] ) ; it may be an object which absorbs the ly@xmath8 emission at the redshift @xcite .
# 3 and # 4 are also detected in spitzer irac images in 3.6 - 8.0 @xmath95 m ( @xcite ; t. webb et al .
2007 , in preparation ) .
the @xmath10-band counterpart of ssa22a - c11 at @xmath96 is also detected ( # 5 ) . for lab2 , four @xmath6-selected sources are found .
lab2-#1 is classified as a drg and it can be the counterpart of ssa22a - m14 ( @xmath97 ) although it is located at @xmath98 north from the @xmath89-band position as mentioned in @xcite .
probably , we see a mature or dusty part of the object in the @xmath10-band image .
lab2-#2 is likely to be foreground objects as the photometric redshift is @xmath93 .
lab2-#4 is also classified as a drg with extremely red color , @xmath99 . in lab7 , @xcite and @xcite sampled two lbgs , ssa22a - m4 ( @xmath100 ) and c6 ( @xmath101 ) , whereas they are separated as three objects in @xmath46-band images of @xcite .
ly@xmath8 emission entirely covers the three objects . from the coordinate
we may identify the north component of this lump as lbg ssa22a - m4 , which is detected in our @xmath10-band image . on the other hand ,
other two components are not detected either in the @xmath17 nor in the @xmath10-band image .
lab8 is located at @xmath102 north of lab1 and may form one large object @xcite .
there is a lbg ssa22a - c15 ( @xmath103 ) , which is not detected in either @xmath17 nor in @xmath10-band .
lab16 is not associated with a lbg .
we find a counterpart candidate in @xmath10-band whose @xmath94 color is slightly redder than those of typical lbgs at @xmath0 .
the photometric redshift of the object is @xmath104 .
lab20 , 30 , and 31 are associated with the lbgs ssa22a - c12 ( @xmath105 ) , d3 ( @xmath106 ) , and c4 ( @xmath107 ) , respectively .
the @xmath10-band counterparts are detected for all of them .
of the 16 laes , only two are detected in @xmath10-band and they are both associated with the labs ( lab16 and lab31 , see above ) . from the upper limit of their @xmath6-band flux , for the rest 14 laes , we estimate that their stellar masses are lower than @xmath108 m@xmath109 , assuming the spectrum of very young ( 10@xmath110-yr old ) objects .
we here discuss whether massive and/or mature galaxies have already formed in the cluster or proto - cluster which was characterized by the overdensity of star - forming galaxies .
there is the evidence that at least some relatively massive young galaxies ( @xmath111 ) have been formed in the proto - cluster .
[ fig_box]b shows that the @xmath6-selected photo - z sample exhibits an excess within a radius of 0.1 - 0.7 arcminutes from the labs .
this indeed suggests that labs are related to the formation of massive galaxies @xcite .
we also found that drg density shows a notable excess around lab1 and lab2 .
how unique are lab1 and lab2 among the 35 labs in the same structure ? they are the largest and the most luminous in ly@xmath8 emission .
moreover , @xcite suggested that lab1 and lab2 are located at the intersection of the filamentary structure of laes .
our results support the picture that lab1 and lab2 sit near the center of the proto - cluster , where the growth of the galaxy structure occurs most preferentially and we are witnessing a stage where a significant amount of the stars is being formed in such regions .
we also investigated the stellar mass of the @xmath6-selected objects , which are expected to be associated with labs , by assuming they are located at @xmath0 as described in section [ sec : photoz ] .
the result is also tabulated in table [ tab3 ] .
although the uncertainty of the stellar mass derived from the sed fitting is large , it is shown that the stellar mass of the lab counterparts ranges from @xmath112 m@xmath109 to @xmath113 m@xmath109 .
[ fig_lya]a shows the relation between ly@xmath8 luminosity and the integrated stellar mass of the @xmath6-selected objects which are associated with each lab , i.e. , the objects in table [ tab3 ] .
surface brightness of ly@xmath8 vs. the stellar mass is also shown in fig . [ fig_lya]b .
these figures suggest that the more massive galaxies are seen in the brighter labs which have the higher surface brightness in ly@xmath8 .
this result implies the origin of the ly@xmath8 emission of labs may be related with their previous star - formation history .
since the local or intermediate - redshift clusters are dominated by the passively - evolving old galaxies , it is worth constraining how many such passive galaxies formed at the higher redshift exist , if any , in the proto - cluster . in fig .
[ fig_colmag ] we plotted the expected color - magnitude relation for the model galaxies with @xmath114 to @xmath115 at @xmath116 of the coma metallicity - sequence model @xcite .
it is clearly seen that few objects have color and magnitude for such passive galaxies formed at @xmath117 .
this result seems to be consistent with that in @xcite , which reported that the bright end of the red sequence in proto - clusters around radio galaxies appeared at @xmath118 , whereas it was not seen at @xmath119 .
@xcite reported that the volume density of lbgs at @xmath120 in the ssa22a / b fields , with the comoving volume of @xmath121 mpc@xmath122 in total , was 6 times higher than the average .
they also evaluated the overdensity of laes which are typically 2 mag fainter than the lbgs in uv continuum to @xmath123 in the ssa22a field and found a similar value .
@xcite revisited the issue by using the deeper and wider - area data of laes .
they confirmed the overdensity of @xmath124 around ssa22a in @xmath125 mpc@xmath122 and found the overdensity of @xmath126 even at the @xmath127 mpc@xmath122 volume .
we try to constrain the overdensity of such population of galaxies in the proto - cluster from the observed surface number density of drgs in ssa22-m1 . to obtain the number density of drgs in a general field at @xmath83
, we adopt the luminosity function with @xmath128 mpc@xmath129 mag@xmath22 , @xmath130 , and @xmath131 for drgs at @xmath132 @xcite .
our limiting @xmath6-band magnitude corresponds to @xmath133 and the expected number density of drgs above the luminosity is @xmath134 mpc@xmath129 .
the volume of the proto - cluster sampled in ssa22-m1 ( @xmath135 @xmath136 11 mpc @xmath29 7 mpc in comoving scale ) is @xmath137 mpc@xmath122 if we assume the redshift range of @xmath138 , which is similar to those studied for the excess density of lbgs or laes .
consequently , the expected average number of drgs in the volume of ssa22-m1 with @xmath139 is obtained to be @xmath140 .
we found at least three drgs in vicinity of the two most luminous labs ( table [ tab2 ] ) . as a result of the number density of drgs
, it is shown that approximately 5 drgs at the redshift provides the similar overdensity as seen for lbgs or laes ( @xmath141 times the average ) .
although the discussion is limited by the large uncertainty due to the small number of the objects , an overdensity similar to those of lbgs / laes is allowed if the number of the foreground / background drgs in the field is less than @xmath142 of that in goods - n ( 29.5 drgs are expected ) .
we presented the results of our deep near - infrared imaging observations of the @xmath0 proto - cluster in the ssa22a field taken by moircs mounted on subaru telescope .
we observed 21.7 arcmin@xmath143 field to the depths of @xmath2 , @xmath3 and @xmath4 with 5@xmath5 .
our observed field covers the area where the surface number density of the laes is highest .
we investigated whether massive and/or mature galaxies have already formed in the proto - cluster which is characterized by the overdensity of star - forming galaxies .
we examined the distribution of the @xmath6-selected galaxies by using the simple color cut for drgs as well as the photo - z selection .
the surface number density of drgs with @xmath144 in the field was 1.3 arcmin@xmath145 at @xmath146 .
while it was not likely that the density of drgs have the similar excess as seen for lbgs at the volume , we found significant evidence that more than a few galaxies with the stellar mass @xmath147 m@xmath109 , exist in vicinity of , or might be associated with , the labs .
we also investigated all the @xmath10-band counterparts which have the consistent photometric redshift not only for the labs but also for the other objects at @xmath0 in the field .
we found that 88% of the labs have the plausible @xmath10-band candidates .
the sum of the stellar mass of the galaxies possibly associated with labs correlated with the luminosity and surface brightness of the labs , which implied that the origin of ly@xmath8 emission must be closely related with the massive galaxy formation phenomena .
the most luminous labs , lab1 and lab2 , had the @xmath6-selected counterparts with @xmath148 m@xmath109 .
in addition , the marginal density excess of drgs and the photo - z selected objects was found around the most luminous labs , lab1 and lab2 .
our results suggest that labs are related to the formation of massive galaxies and a certain amount of evolved galaxies with @xmath149 are being formed in the central part of the high density region of star - forming galaxies at @xmath0 .
we are witnessing a stage when significant amount of the stars are being formed in the central part of the growing large - scale proto - cluster at @xmath0 .
we thank the staff of the subaru telescope for their assistance with the development and the observation of moircs .
this study is based on data collected at subaru telescope , which is operated by the national astronomical observatory of japan .
this research is supported in part by the grant - in - aid for scientific research 11554005 and 14340059 of the ministry of education , science , culture , and sports in japan . the image reduction and analysis facility ( iraf ) used in this paper
is distributed by the national optical astronomy observatories , u.s.a .
, which are operated by the association of universities for research in astronomy , inc .
, under cooperative agreement with the national science foundation .
abraham , r. g. , 2004 , , 127 , 2455 arnouts , s. , 2007 , , submitted ( astro - ph/0705.2438 ) adelberger , k. l. , steidel , c. c. , giavalisco , m. , dickinson , m. , pettini , m. , & kellogg , m. 1998 , , 505 , 18 basu - zych , a. & scharf , c. 2004 , , 615 , l85 benson , a. j. , frenk , c. s. , baugh , c. m. , cole , s. , & lacey , c. g. 2001 , , 327 , 1041 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bolzonella , m. , miralles , j .-
m . , & pell , r. 2000 , , 363 , 476 bower , r. g. , benson , a. j. , malbon , r. , helly , j. c. , frenk , c. s. , baugh , c. m. , cole , s. , & lacey , c. g. 2006 , , 370 , 645 bruzual , g. & charlot , s. 2003 , , 344 , 1000 caputi , k. i. , mclure , r. j. , dunlop , j. s. , cirasuolo , m. , & schael , a. m. 2006 , , 366 , 609 cen , r. & ostriker , j. p. 2000 , , 538 , 83 calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. , & storchi - bergmann , t. 2000 , , 533 , 682 cirasuolo , m. , et al .
2007,,380,585 cowie , lennox l. ; songaila , antoinette ; hu , ester m. ; cohen , j. g. 1996 , , 112 , 839 chabrier , g. 2003 , , 586 , l133 conselice , c. j. , et al .
2007 , , 660 , l55 daddi , e. , rttgering , h. j. a. , labb , i. , rudnick , g. , franx , m. , moorwood , a. f. m. , rix , h. w. , van der werf , p. p. , & van dokkum , p. g. 2003 , , 588 , 50 doherty , m. , bunker , a. j. , ellis , r. s. , & mccarthy , p. j. 2005 , , 361 , 525 fontana , a. , et al .
2006 , , 459 , 745 frster schreiber , n. m. , et al .
2004 , , 616 , 40 foucaud , s. , et al .
2007 , , in press astro - ph/0606386 francis , p. j. , 2001 , , 554 , 1001 franx , m. et al .
2003 , , 587 , 79 fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , schneider , d. p. 1996
, , 111 , 1748 geach , j. e. , smail , i. , chapman , s. c. , alexander , d. m. , blain , a. w. , stott , j. p. , & ivison , r. j. 2007 , , 655 , l9 grazian , a. , 2006 , , 453 , 507 hayashino , t. , et al .
2004 , , 128 , 2073 ichikawa , t. , et al .
2006 , in proc .
of spie , vol .
6269 , 38 ichikawa , t. , et al .
2007 , , in press ( astro - ph/0701820 ) i m , m. , yamada , t. , tanaka , i. , & kajisawa , m. 2002 , , 578 , l19 kajisawa , m. & yamada , t. 2005 , , 618 , 91 kajisawa , m. , & yamada , t. 2006 , , 650 , 12 kajisawa , m. , et al .
2006 , , 58 , 951 kauffmann , g. , colberg , j. m. , diaferio , a. , & white , s. d. m. 1999 , , 307 , 529 kodama , t. , & arimoto , n. 1997 , 320 , 41 kodama , t. , tanaka , i. , kajisawa , m. , kurk , j. , venemans , b. , de breuck , c. , vernet , j. l , & lidman , c. 2007 , , 377 , 1717 labb , i. et al .
2005 , , 624 , l81 laird , e. s. , nandra , k. , hobbs , a. , steidel , c. c. 2006 , , 373 , 217 lane , k. p. , et al .
2007 , , 379 , l25 maihara , t. , et al .
2001 , , 53 , 25 marchesini , d. , 2007 , , 656 , 42 matsuda , y. , et al .
2004 , , 128 , 569 matsuda , y. , et al . 2005 , , 634 , l125 matsuda , y. , yamada , t. , hayashino , t. , yamauchi , r. , & nakamura , y. 2006 , , 640 , l123 matsuda , y. , iono , d. , ohta , k. , yamada , t. , kawabe , r. , hayashino , t. , peck , a. b. , & petitpas , g. r. 2007 , , 667 , 667 oke & gunn 1983 , , 277 , 713 ohyama , y. , taniguchi , y. , kawabata , k. s. , shioya , y. , murayama , t. , nagao , t. , takata , t. , iye , m. , yoshida , m. 2003 , apj , 591 , l9 ouchi , m. , 2001 , , 558 , l83 ouchi , m. , 2005 , , 635 , l117 papovich , c. , dickinson , m. , & ferguson , h. c. 2001 , , 559 , 620 quadri , r. et al .
2007 , , 654 , 138 quadri , r. et al .
2007 , , 134 , 1103 reddy , n. a. , erb , d. k. , steidel , c. c. , shapley , a. e. , adelberger , k. l. , & pettini , m. 2005 , , 633 , 748 saracco , p. , 2004
, , 420 , 125 schlegel , d.j . , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 shapley , a. e. , steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , & pettini , m. 2001 , , 562 , 95 steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , pettini , m. , & kellogg , m. 1998 , , 492 , 428 steidel , c. c. , adelberger , k. l. , giavalisco , m. , dickinson , m. , & pettini , m. 1999 , , 519 , 1 steidel , c. c. , 2000 , 532 , 170 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003a , arxiv : astro - ph/0305378 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003b , , 592 , 728 songaila , a. , cowie , l. l. , hu , e. m. , & gardner , j. p. 1994
, , 94 , 461 suzuki , r. , et al .
2007 , submitted to swinbank , a. m. , smail , ian , chapman , s. c. , blain , a. w. , ivison , r. j. , & keel , w. c. 2004 , , 617 , 64 tokunaga , a. t. , simons , d. a. , & vacca , w. d. 2002 , psap , 114 , 180 totani , t. , yoshii , y. , iwamuro , f. , maihara , t. , & motohara , k. 2001 , , 558 , l87 verma , a. , lehnert , m. d. , frster schreiber , n. m. , bremer , m. n. , & douglas , l. 2007 , , 377 , 1024 webb , t. m. a. , et al .
2006 , , 636 , l17 lab1 & c11 ( # 5 ) & 22:17:25.7 & 0:12:34.6 & 23.18 @xmath71 0.16 & 1.23 @xmath71 0.42 & 1.31 @xmath71 0.56 + & # 1 & 22:17:26.1 & 0:12:46.7 & 22.23 @xmath71 0.10 & 1.08 @xmath71 0.28 & 4.08 @xmath71 0.51 + & # 2 & 22:17:25.8 & 0:12:43.8 & 22.75 @xmath71 0.13 & 0.33 @xmath71 0.26 & 1.34 @xmath71 0.39 + & # 3 & 22:17:26.0 & 0:12:36.4 & 23.07 @xmath71 0.15 & 1.48 @xmath71 0.45 & 2.77 @xmath71 0.51 + & # 4 & 22:17:26.1 & 0:12:32.3 & 22.38 @xmath71 0.11 & 1.31 @xmath71 0.33 & 2.20 @xmath71 0.30 + lab2 & m14 ( # 1 ) & 22:17:39.1 & 0:13:30.6 & 23.30 @xmath71 0.16 & 1.60 @xmath71 0.47 & 2.60 @xmath71 0.64 + & # 2 & 22:17:39.1 & 0:13:26.4 & 22.26 @xmath71 0.10 & 0.55 @xmath71 0.25 & 1.13 @xmath71 0.30 + & # 3 & 22:17:38.9 & 0:13:24.0 & 19.98 @xmath71 0.03 & 0.50 @xmath71 0.07 & 1.85 @xmath71 0.03 + & # 4 & 22:17:39.3 & 0:13:22.1 & 22.63 @xmath71 0.11 & 1.95 @xmath71 0.41 & 4.32 @xmath71 0.80 + lab7 & m4 & 22:17:41.0 & 0:11:27.8 & 23.78 @xmath71 0.19 & 0.23 @xmath71 0.34 & 0.90 @xmath71 1.02 + & c6 & - & - & - & - & - + lab8 & c15 & - & - & - & - & - + lab16 & & 22:17:24.9 & 0:11:17.5 & 23.29 @xmath71 0.16 & 0.91 @xmath71 0.36 & 1.78 @xmath71 0.59 + lab20 & c12 & 22:17:35.3 & 0:12:47.4 & 24.12 @xmath71 0.23 & 0.70 @xmath71 0.44 & 0.49 @xmath71 1.12 + lab30 & d3 & 22:17:32.5 & 0:11:32.9 & 22.88 @xmath71 0.13 & 0.73 @xmath71 0.31 & 1.08 @xmath71 0.41 + lab31 & c4 & 22:17:39.0 & 0:11:26.1 & 23.39 @xmath71 0.16 & 1.14 @xmath71 0.41 & 1.29 @xmath71 0.73 + lab1 & c11 & 3.109 & 2.83 & 1.80 & 0.08 & -22.32 & 2.5 @xmath155 + & # 3 & - & 2.85 & 1.80 & 0.20 & -22.59 & 7.2 @xmath156 + & # 4 & - & 2.61 & 1.61 & 0.18 & -23.19 & 10.9 @xmath157 + lab2 & m14 & 3.091 & 3.19 & 1.80 & 0.22 & -22.64 & 8.1 @xmath158 + & # 4 & - & 2.61 & 0.20 & 0.40 & -23.05 & 10.5 @xmath159 + lab7 & m4 & 3.093 & 2.70 & 0.57 & 0.06 & -21.75 & 0.7 @xmath160 + lab16 & - & - & 2.70 & 0.06 & 0.20 & -22.50 & 1.6 @xmath161 + lab20 & c12 & 3.118 & 2.85 & 0.57 & 0.00 & -21.43 & 0.4 @xmath162 + lab30 & d3 & 3.086 & 3.31 & 1.61 & 0.08 & -22.74 & 3.4 @xmath163 + lab31 & c4 & 3.076 & 3.19 & 1.80 & 0.00 & -22.04 & 2.0 @xmath164 + | we present the results of deep near - infrared imaging observations of the @xmath0 proto - cluster region in the ssa22a field taken by moircs mounted on the subaru telescope .
we observed a 21.7 arcmin@xmath1 field to the depths of @xmath2 , @xmath3 , and @xmath4 ( 5@xmath5 ) .
we examine the distribution of the @xmath6-selected galaxies at @xmath7 by using the simple color cut for distant red galaxies ( drgs ) as well as the photometric - redshift selection technique . the marginal density excess of drgs and the photo - z selected objects are found around the two most luminous ly@xmath8 blobs ( labs ) .
we investigate the correlation between the @xmath6-selected objects and the labs , and find that several galaxies with stellar mass @xmath9 exist in vicinity of labs , especially around the two most luminous ones .
we also find that 7 of the 8 labs in the field have plausible @xmath10-band counterparts and the sum of the stellar mass possibly associated with labs correlates with the luminosity and surface brightness of them , which implies that the origin of ly@xmath8 emission may be closely correlated with their stellar mass or their previous star formation phenomena . |
the highly correlated f electrons in rare earth materials have challenged the reliability of the results obtained by conventional density functional theory ( dft ) .
it is generally believed that regular dft in its lda / gga formulation , without any modification , is insufficient to correctly deal with the f electrons ; hence our understanding of the physical properties of these systems are often based on a many body model hamiltonian or modified lda such as lda+u @xcite , lda+sic @xcite , and lda+(open core ) @xcite approaches .
strange et.al @xcite , by removing the spurious self - interaction of f electrons reported a systematic investigation of the rare earth elements and their sulfides .
they confirmed the pre - claimed idea that there are two types of f electrons in these materials : localized core - like f electrons that determine the valency , and delocalized band - like f electrons that are formed through hybridization with the s - d bands and which participate in bonding .
the main feature of their calculation was based on an f band splitting into two sub - bands one occupied , and the other unoccupied . in a recent publication we investigated the valency of rare earths in rin@xmath0 ; ( r = sm , eu , gd ) and rsn@xmath0 ; ( r = tm , yb , lu ) and calculated the electric field gradient using dft without spin polarization @xcite .
we found out that in regular dft as all f electrons are treated as band electrons , consequently their density of states are all located in a single peak at the fermi level which is not consistent with reality .
we finally overcame this drawback by applying open core approach .
the aim of the present work is to investigate the effect of spin polarization on structural properties and the gradient of electric field of the same samples .
our objective is to investigate the borderline of lda ability , that without any modification further than spin polarization can obtain reliable results for some compounds with highly correlated f electrons .
our calculations confirm the fact that in most cases spin polarization , without using lda+u , lda+sic , or even open core modification can treat the highly correlated f electrons properly .
it means that at low temperature , where the system is magnetically ordered , spin polarization , by providing more flexibility among f electrons and opening degrees of freedom for spin - splitting , partly compensate the necessary coulomb correlation .
the organization of this paper is as follow . in section
ii we present the calculational details .
section iii deals with the results and discussions .
finally , we conclude this work with a summary in section iv .
we utilize the full potential linearized augmented plane wave ( fp - lapw ) method within density functional theory . in this method
wavefunctions , charge density , and potential are expanded in spherical harmonics within nonoverlapping muffin - tin spheres and in plane waves in the remaining interstitial region of the unit cell . as in fp - lapw method
no shape approximation is introduced for potential and charge density , then it is an appropriate procedure to obtain reliable results for the electric field gradient ( efg ) with high sensitivity to unisotropic charge density distribution around the nucleus .
the efg tensor is defined as the second derivative of the coulomb potential with respect to the cartesian coordinate at the nucleus , which is written as a traceless tensor .
the coulomb potential which is one of the crucial quantities of an accurate ( full - potential ) band structure calculation is given in lapw as lattice harmonics expansion inside the atomic spheres . in this representation , the required derivatives can be determined straightforwardly , and the largest component of efg tensor can be obtained directly from the l=2 , m=0 component of the potential expansion inside the muffin tin spheres@xcite : @xmath1 it is assumed that the main axis of the efg tensor points towards z direction .
the radial potential coefficient @xmath2 near the nucleus ( @xmath3 ) for a given charge density is obtained numerically by solving poisson s equation using a method proposed by weinert @xcite .
the calculations were performed using the wien97 code developed by blaha and coworkers @xcite , which has been successfully applied to a wide range of systems . in this code core and valence states separated by their energy
are treated differently .
core states are treated within multiconfiguration relativistic dirac - fock approach while valence states are only scalar relativistic with an exception of @xmath4 for which the spin - orbit coupling in a second order variational steps were included during self - consistent calculations of valence state . in the later case
the number of symmetry was reduced from 48 to 16 by assuming z - direction as a preferentional direction of the coupled spin to the orbit .
it was shown that by including the spin - orbit interaction the results does not change substantially , and as the incorporation of this coupling is very time consuming then it was neglected throughout the calculation . for better treatment of the core electrons we decomposed them into high lying semicore and deeper true core states .
as the electronic charge of semicore states are not completely confined inside the respective muffin - tin spheres , therefore we treated them as band states in the full ( nonspherical ) potential .
for these states in addition to the usual lapw basis we included local orbitals to increase the flexibility of the basis set @xcite , and being able to use a sufficiently fine k sampling .
this procedure results the more accurate calculation of efg , which is extremely sensitive to the basis set @xcite .
the true core states were treated atomic like in the self - consistent muffin - tin potential . in our calculations
electronic states with an energy of at least 6 ry below the fermi energy were considered as true core states . for the muffin - tin radius of rare earth elements as well as sn and in a value of 2.65 a.u .
were chosen throughout while for the cd impurity manipulated into the rsn@xmath0/in@xmath0 compounds a radius of 2.5 a.u . was used .
the maximum angular momentum quantum number l as a cutoff for expanding the kohn - sham wavefunctions in terms of lattice harmonics inside the muffin - tin spheres was confined to @xmath5 .
the potential and charge density was expanded inside the atomic spheres in crystal harmonics up to l=6 .
the wavefunctions in the interstitial region were expanded in plane waves with a cutoff of @xmath6 where @xmath7 is the smallest muffin - tin radius in the unit cell .
the periodic charge density and potential were fourier expanded up to @xmath8 .
we adjusted the value of mixing parameter of charge density to 0.1 and used broyden s scheme .
we have taken a mesh of 165 special k points to implement the valence state integration in the irreducible wedge of the brillioun zone ( bz ) that correspond to the grids @xmath9 in the scheme of monkhorst - pack @xcite .
the mesh of k points , charge density and potential expansion cutoff and finally muffin - tin radii were allowed to vary for ensuring the convergence .
the exchange - correlation energy was calculated within the local density approximation ( lda ) based on the perdew - zunger parameterization @xcite , with and without including advanced generalized gradient approximation ( gga ) using perdew - burke - ernzerhof parameterization @xcite , with and without considering spin polarization . for spin
polarized calculations , a ferromagnetic moment was imposed on r at the start of the selfconsistency cycle . as for some compounds
this is not the true magnetic order @xcite , such an assumtion should be considered as an extra possible source of error . at the end of selfconsistency cycle ,
@xmath10 and @xmath11 came out to be nonmagnet , all others remained ferromagnetic . in order to calculate the electric field gradient at cd site as a representation of an impurity atom in each compound all sides of the unit cell were doubled to construct a supercell with 16 atoms @xcite .
further increase of the sides of supercell did not change the results significantly . to keep the same accuracy with the previous primitive unit cell
the mesh of k points in supercell was reduced to 75 corresponding to the @xmath12 grids .
the energy for each orbital in @xmath4 , and in @xmath10 is linearized by selecting the most probable energy calculated for each density of state over the valence and semicore energy windows .
the linearizations obtained for @xmath4 and @xmath10 then are used as an initial input for the @xmath13 , @xmath14 , and the @xmath15 , @xmath11 , respectively and we allow the program to search automatically for the best - linearized energies of cd orbitals and check if they are perfect .
by calculating the total energy of a primitive unit cell as a function of its volume and fitting the data with the birch equation of state @xcite we obtained the lattice parameters , bulk moduli , and the pressure derivative of the bulk moduli for rsn@xmath0 ( r = sm , eu , gd ) and rin@xmath0 ( r = tm , yb , lu ) by lda , gga , gga+sp , gga + open core ( valence 2 , 2.5 and 3 ) approaches .
these results along with the experimental data are listed in tables [ table1 ] and [ table2 ] .
.[table1]the comparison of the measured lattice parameters in @xmath16 at t=300k with the theoretical values calculated at t=0k within lda , gga without spin polarization ( gga ) , gga with spin polarization ( gga+sp ) , and gga+open core ( valence 2 , 2.5 , 3 ) . [ cols="^,^,^,^,^,^,^",options="header " , ] at the end to analyze the up and down contributions to vzz on cd , sn and in in the listed compound we concentrate , as an example , on sn site in @xmath4 .
the spin up and down dos and their corresponding @xmath17 and @xmath18 on sn cites are shown in fig2 .
first of all , it is seen that spin polarization has splited the up and down f peaks ( fig2.a , and b ) .
this feature is compatible with reality . in nature
there are two types of f electrons , localized and nonlocalized , and as a result of that the f band should split into two sub - bands . in regular gga as the strong correlation between the f electrons are not treated properly , then all of them are located in a single peak at @xmath19 .
hence it is not unjustified to expect that magnetism push the system in the right direction .
how the f - peak splitting affects @xmath17 , @xmath18 , @xmath20 , and consequently efg ?
fig2c and d show the changes of @xmath17 and @xmath18 with energy for up and down spin separately .
below -4ev the two set of curves coincide . in the region
[ -4 , 0 ] ev @xmath21 is slightly larger than @xmath22 .
this difference is more visible at around -1 ev , which is the location of @xmath23 peak .
one can conclude that the sn @xmath24-electrons in the xy - plane hybridizes with the f - electrons of the neighboring rare earths ( which are all 4 in the xy - plane ) . in fig.3 the anisotropy function @xmath25
is given for two spin directions ( call it @xmath26 and @xmath27 ) .
at fermi level , @xmath28 is slightly larger than @xmath29 .
exactly the same was seen for @xmath30 in table [ table3 ] , which demonstrates once more that the p - anisotropy really reflects the behavior of @xmath30 . as we already expected from @xmath17 and @xmath31 curves ,
no difference is observed between the two curves below -4 ev . in the region
[ -4 , 0 ] @xmath26 is larger than @xmath27 and at around -1 ev @xmath24 anisotropy shows a clear peak .
this is due to the @xmath23 peak located in this area .
we employed the fp - lapw method to calculate the structural properties of rsn@xmath0 ( r = sm , eu , gd ) and rin@xmath0 ( r = tm , yb , lu ) by lda , gga , gga+sp , and gga+open core ( valence 2 , 2.5 , and 3 ) approaches . for eu and yb compounds open core and for other compounds gga+sp results are in better agreement with experiment .
we also calculated the efg at sn , in , and impurity cd sites in rsn@xmath0 ( r = sm , eu , gd ) and rin@xmath0 ( r = tm , yb , lu ) compounds . here also , the results obtained by gga+sp are in better agreement with experiment .
the reasonable agreement of the results calculated by gga+sp with experiment confirm the fact that even for such highly correlated systems the spin splitting introduced by magnetism can compensate the strong coulomb correlation .
our approach has the advantage of treating the localized and itinerant electronic states on the same common footing .
99 p. novak , f. boucher , gressier , p. blaha , and k. schwarz , phys .
b * 63 * , 235114 ( 2001 )
price , b.r .
cooper , say - peng lim , i. avgin , phys .
b * 61 * , 9867 ( 2000 ) .
anisimov , f. aryasetiawan , and a.i .
lichtenstein , j. phys .
: condens .
matter * 9 * , 767 ( 1997 ) .
a. svane , w.m .
temmerman , z. szotek , l. petit , p. strange , and h. winter , phy .
b * 62 * , 13394 ( 2000 ) .
strange , a. svane , w.m .
temmerman , z. szotek , nature , * 399 * , 756 ( 1999 ) .
m. richter , j. phys .
phys . * 31 * , 1017 ( 1998 ) .
s. jalali asadabadi , s. cotteneer , h. akbarzadeh , r. saki , and m. rots , submitted to prb * * , ( 2002 ) .
p. blaha , k. schwarz , and p. h. dederichs , phys .
b * 37 * , 2792 ( 1988 ) .
m. weinert , j. math .
* 22 * , 2433 ( 1981 ) .
p. blaha , k. schwarz , and j. luitz , wien97 , a full potential linearized augmented plane wave package for calculating crystal properties ( karlheinz schwarz , tech .
wien , vienna 1999 ) .
isbn 3 - 9501031 - 0 - 4 .
updated version of p. blaha , k. schwarz , p. sorantin , and s.b .
trickey , comp .
. commun . * 59 * , 399 ( 1990 ) . in this work ,
version 97.10 was used .
d.j . singh , phys .
b * 43 * , 6388 ( 1991 ) .
bominaar , j. guillin , a. sawayn , and a.x .
traufuxen , phy .
b * 39 * , 72 ( 1989 ) .
monkhurst , j.d .
pack , phys .
b * 13 * , 5188 ( 1976 ) .
perdew , a. zuanger , phys .
b * 23 * , 5048 ( 1981 ) .
perdew , s. burke , and m. ernzerhof , phys .
lett . * 77 * , 3856 ( 1996 ) .
sanchez , j.m .
friedt , g.k .
shenoy , a. percheron , and j.c .
achard , j. phys .
c : solid state phys .
* 9 * , ( 1976 ) .
f. birch , j. geophys .
83 * , 1257 ( 1978 ) .
a. delin , l. fast , and brje johanson , phys .
b * 8 * , 58 ( 1998 ) .
l. petit , a. svane , w. m. temmerman , and z. szotek , phys .
b * 63 * , 165107 ( 2001 ) . g.p .
schwartz , and d.a .
shirley , hyperfine interaction * 3 * , 67 ( 1977 ) .
s. lany , p. blaha , j. hamann , v. ostheimer , h. wolf , and t. wichert , phys .
b * 62 * , 2259 ( 2000 ) . | we investigated the effect of spin polarization on the structural properties and gradient of electric field ( efg ) on sn , in , and cd impurity in rsn@xmath0 ( r = sm , eu , gd ) and rin@xmath0 ( r = tm , yb , lu ) compounds .
the calculations were performed self - consistently using the scalar - relativistic full potential linearized augmented plane wave method . the local density approximations ( lda ) and generalized gradient approximation without spin polarization ( gga ) and with spin polarization ( gga+sp ) to density functional theory were applied .
in addition to that we performed some calculations within open core treatment ( gga+open core ) .
it is clearly seen that gga+sp is successful in predicting the larger lattice parameter and the dramatic drop of efg for r=(eu , yb ) relative to other rare earth compounds .
this is an indication that spin splitting generated by spin polarization without any modification , is capable of treating properly the highly correlated f electrons in these systems . |
the hybrid monte carlo algorithm has been already explored in recent years .
it can be improved further , as iterative solvers become more efficient and non - local reversible integrators can be constructed .
the multi - boson algorithm is new and allows itself for further improvement .
it is a more complex algorithm which has more degrees of freedom for improvement .
one issue is the optimality of the polynomial .
the range of applications is also broader .
we have mentioned in @xcite that one flavor qcd becomes possible with this algorithm .
it has a straightforward application to the staggered fermions . in finite density qcd simulations ,
the sign problem is a long standing problem . with the multi - boson algorithm
is possible to approximate the phase of the quark determinant for small chemical potential @xcite . as illustration we computed in the hermitian approximation ritz values of @xmath79 by lanczos algorithm for @xmath80 for one quenched @xmath81 blocked configuration at @xmath82 and @xmath83 . in fig .
1 we show in the complex plane ritz values together with adaptive roots of @xmath18 . we compare them with the roots of the chebyshev polynomial . comparison between ritz and chebyshev polynomials is given in fig .
2 . for @xmath84
we have repeated the computation and the result is shown in fig .
3 . in both cases
the ritz polynomial performs better than the chebyshev one .
it is exactly zero at the first ritz value .
i would like to thank professor k. schilling and his staff for the warm atmosphere that sorrounded me in hlrz jlich , where this work was written .
i thank also j. sexton and h. hoeber for helpful suggestions on this work .
finally , i thank ph . de frocrand and a. galli for discussions on the global heatbath algorithm .
the gauge field was taken form large scale quenched simulation of taro collaboration . | we estimate theoretically the cost of the multi - boson method in the non - hermitian approximation . it is shown that it is proportional to @xmath0 . for a global update of the scalar fields
the cost decreases by a factor @xmath1 with a @xmath2 overhead .
epsf -1.8 cm there is an increasing interest in lattice qcd community in better algorithms for dynamical fermions @xcite . in @xcite
the kramers algorithm is considered .
this is a variant of the hybrid monte carlo ( hmc ) algorithm @xcite , where the equations of motion have a stochastic part .
however , the new algorithm proposed recently by lscher @xcite has become attractive for it brings new views in full qcd simulations .
the way it is implemented makes it suffer from the critical slowing down , which is mainly caused by the local heatbath update of the bosonic fields @xcite .
we want in lattice qcd to estimate the determinant of the quark matrix . for two
degenerate quarks it can be written as @xmath3 , where @xmath4 is the wilson matrix with @xmath5 let @xmath6 be an order @xmath7 polynomial with roots , @xmath8 , @xmath9 such that @xmath10 let @xmath11 be the error of the polynomial approximation , which is defined as @xmath12 lscher s original proposal uses the hermitian quark matrix @xmath13 and a real polynomial with complex conjugate roots . as we have proposed in @xcite , a non - hermitian approximation is expected to work better than the hermitian one .
the method consists in the following equalities : @xmath14 @xmath15 @xmath16 [ d{\phi}_k ] e^{-{\phi}_k^{\dag } ( w - z_k)^{\dag } ( w - z_k ) { \phi}_k } \ ] ] where @xmath17 is the rank of @xmath4 .
we use as @xmath18 the chebyshev polynomials defined in the complex plane @xcite which have certain optimal properties .
as opposed to hmc , this method introduces a local effective quark action on both gauge and scalar fields @xmath19 .
naturally , this allows a local monte carlo ( mc ) update of these fields .
the most important question is which algorithm is cheaper .
we answer this question by theoretical arguments and propose a global heatbath update for the scalar fields .
we analyse the volume ( @xmath20 ) and quark mass ( @xmath1 ) dependence of the cost , which we denote by @xmath21 ( denoting by @xmath20 both the volume and the rank of the matrix should not cause any ambiguity : they are proportional ) . clearly , each mc sweep has a cost proportional to the volume of the lattice , the number of the bosonic fields and the autocorrelation time @xmath22 .
we suppose that the gauge and scalar fields are updated locally by the heatbath algorithm .
the cost of the algorithm will scale like @xmath23 random walk arguments allow us to assume that @xmath24 where @xmath25 is the correlation length of the operator @xmath26 .
then it is clear that @xmath27^{-1}||_2\ ] ] where @xmath28 denotes the 2-norm of a matrix . to this end
we need explicitly the roots of the chebyshev polynomial which are given by @xcite @xmath29 where @xmath30 is the center of the spectrum and @xmath31 is the focal distance of the ellipse that encloses the spectrum . in the asymptotic regime , as @xmath32 , the roots approach a dense set of points , the ellipse that passes through the origin
. clearly we obtain @xmath33 which is independent of @xmath34 . assuming that the smallest singular value of @xmath4 behaves like @xmath35 we estimate the autocorrelation time to scale like @xmath36 on the other hand the dynamics of the gauge fields is coupled to that of the bosonic fields .
this can be seen if we look at the step size of one gauge field update .
we use a slightly different argument of @xcite and write @xmath37 we see that for wilson fermions the bosonic part of the action is quadratic on gauge fields @xmath38 and the variance of the corresponding distribution is proportional to @xmath39 .
this shows that the step size of a gauge field update is proportional to @xmath40 .
as result , the autocorrelation time will be proportional to @xmath7 .
then the total cost of the simulation will scale like @xmath41 to see how the number of boson fields scales with the volume and the quark mass , we consider the error @xmath42 of the approximation : @xmath43 - 1\ ] ] or in the eigenvalue basis @xmath44 - 1\ ] ] let @xmath45 be an uniform upper bound for @xmath46 ( i.e. for each @xmath47 ) .
then it can be easily shown that @xmath48 so that in the asymptotic regime ( @xmath7 large and @xmath45 small ) we obtain @xmath49 for chebyshev polynomials and small @xmath1 we have @xcite @xmath50 where @xmath51 is an @xmath52 real constant . in this way
we obtain @xmath53 keeping the approximation error constant this means that @xmath7 will scale like @xmath54 so that the total cost scales like @xmath55 the cost of the hmc algorithm scales at best like @xcite @xmath56 this shows that the lscher s algorithm scales better with the volume then hmc , whereas the opposite can be said for the scaling with the quak mass .
the simulations of dynamical fermions for an @xmath57 gauge theory with the multi - boson algorithm and kramers algorithm show that they perform comparably , the latter being a bit faster @xcite , a fact that supports our argument .
however , one can try to reduce the autocorrelation time , so that the algorithm can compare favourably to hmc .
this can be achieved by performing a global heatbath on the scalar fields .
consider a global heatbath update of the bosonic fields in the form @xmath58 this step is very costly because the inversion is not necessary well conditioned .
instead , we propose a well conditioned inversion to take place : we use as a polynomial preconditioner the chebyshev polynomials @xmath18 of the multi - bosonic method and write the above global update as @xmath59 where @xmath60^{-1}\ ] ] the factor @xmath61 , has its inverse in one of the factors of @xmath62 , so that we do not need to compute it .
we have to invert instead the better conditioned matrix @xmath63 .
this computation has to take place anyway for the exact version of the multi - boson algorithm proposed in @xcite . if @xmath34 is the number of iterative steps for the above inversion to converge , its cost will scale like @xmath64 it remains to see how @xmath34 scales with the volume and the quark mass .
the minimum eigenvalue of the matrix @xmath63 is given by @xmath65 we had @xmath66 so that we obtain @xmath67 and the number of iterations @xmath34 scales like @xmath68 as the quark mass is fixed and @xmath7 grows , the matrix @xmath63 is well conditioned . in any case
an optimal iterative solver requires a minimum number of steps to converge that grows like @xmath2 .
this can be seen for example by modeling our well conditioned problem as an inversion of the quark matrix in one dimension ( i.e. the quantum mechanics of a fermion particle ) .
this problem can be solved by divide and conquer : by even - odd splitting the original lattice we obtain two decoupled sublattices , which of them can be split similarly in two sublattices and so on ( the first step as we know can be used in higher dimensions too ) .
clearly , the number of steps needed to arrive at one - site sublattices is @xmath69 .
this way , we finally get @xmath70 and the inversion cost will scale like @xmath71 the total cost scales like @xmath72 with @xmath73 , so that finally we obtain @xmath74 the cost of the algorithm with a global heatbath update of the scalar fields decreases by a factor @xmath1 . the overhead is a factor @xmath2 in the volume . since the arguments are given in the limiting case @xmath75 and @xmath32 , we expect the simulations to verify the above cost scaling in this limit
. for moderate masses and number of bosonic fields the algorithm can scale better .
we note that the above scaling analysis dos not take into account the prefactors in the cost of the inversion that can make the simulations expensive .
we stress the fact that if the inversion techniques become less costly , the global heatbath algorithm can be a real alternative to the local one . |
it is well known that a sharp contrast exists today in the quantum - mechanical treatment of few- and many - body systems between very effective and fast bound - state calculations , on the one hand , and very time - consuming few - particle scattering calculations on the other hand .
practical solutions for a discrete spectrum may incorporate many hundreds or even thousands particles with simple coulomb - like interactions in atomic or molecular physics , or up to 20 - 25 nucleons with complicated realistic @xmath1 interactions in nuclear physics , while even the solution of the four - nucleon scattering problem with realistic @xmath1-interactions , especially above the three - body breakup threshold , represent a strong challenge for modern theorists @xcite
. there are at least two reasons for such a strong contrast .
first , the few - body scattering problem includes complicated boundary conditions , especially above the three- or four - body breakup thresholds , and second , the multi - particle hamiltonian has a degenerate continuous spectrum , so that each pair or triple of particles can be in infinite number of states at the same energy .
the first problem has been solved mathematically by formulation of the faddeev
yakubovsky equations whose full solution satisfies , as has been strictly proved @xcite , all of the necessary boundary conditions . however , the price for this correct formulation is a very sophisticated form of these integral equations , whose kernels have complicated moving singularities .
therefore , in the previous four decades a lot of exact and approximate methods for solving the faddeev
yakubovsky equations were proposed . however , due to the complexity of realistic few - body scattering problems , practical solutions usually require a massively parallel implementation , so that even now exact faddeev - like scattering calculations are performed mainly on powerful supercomputers ( see e.g. the recent @xmath2-calculations @xcite ) .
the complexity of the few - body equations leads to the fact that an accurate numerical treatment for realistic few - body scattering problems remains available only to a limited number of experts .
the most effective way to treat the second key problem related to degeneracy of few - body continuous spectra of the total and channel hamiltonians is their discretization by one or another method and usage of @xmath0 normalized states as approximations for exact continuum states . nowadays , many such discretization methods exist ( see , for example , a recent comprehensive review @xcite ) .
however , it is still not clear whether these particular discretization methods give a discrete form of scattering equations , which permit a high degree of parallelism in a numerical solution .
the last point is crucially important for further progress in few - body scattering calculations because even powerful supercomputers can not give essential acceleration of the calculations if the solution method does not permit an effective parallelization for all parts of the algorithm .
quite recently , a new computational technique has been introduced based on the general - purpose graphics processing unit ( gpgpu ) .
this technique utilizes a graphics processing unit ( gpu ) which has been initially designed to carry out computations for computer graphics .
nowadays , gpus are specialized to perform ultra - fast general purpose computations and they can replace a supercomputer realization in many particular cases . also , special extensions of standard programming languages are developed to use gpu facilities in tedious scientific calculations ( see e.g. @xcite ) . this technique has been actively pursued and successfully used in quantum chemistry @xcite , in lattice qcd calculations @xcite , monte - carlo simulations etc .
recently , _ ab - initio _ nuclear structure gpu - calculations @xcite have been performed as well as gpu treatments of faddeev equations for quantum trimer systems @xcite .
it is of great interest to apply such gpu - techniques to realistic few - body scattering calculations .
this would open new possibilities for accurate few - body studies in general and could make them more accessible to a wider number of researchers .
however , the gpu realization requires an appropriate and specific formulation for scattering problems because this realization is most effective for the algorithms with a high degree of parallelism and minimal interdependence between data processing in parallel threads .
the present authors suggested in previous years the wave - packet continuum discretization technique @xcite which has been tested carefully for the model @xmath1 interactions and found to be very efficient .
one of the important features of the above discrete approach is that the resulting discretized form of the scattering equations is well suited for such a parallel realization .
below we show that such a massively parallel implementation of the whole solution can be made on a standard pc with a modern graphics processor that can perform all of the calculations using many thousands of parallel threads . in the present paper
we have also made a generalization of the wave - packet approach to few - body equations with fully realistic @xmath1 interaction which is not a trivial problem and requires a new determination for the multi - channel resolvent in an analytical form .
so we included a section with this description in the present paper .
the paper is organized as follows . in sect .
ii we summarize the main features of the wp approach and describe how it is used to solve three - body scattering problem . in sect .
iii the case of the coupled - channel two - body input interaction is discussed and formulas for the three - body channel resolvent are given .
the results for the @xmath3 elastic scattering problem are represented in sect .
section v is dedicated to a description of our first gpu tests for the problem in question and a comparison of the corresponding computational efficiency of the cpu and gpu realizations on the same pc .
the main results of the paper are summarized in the conclusion .
the present work discusses the solution few- and many - body scattering problems in atomic , molecular , nuclear and hadronic physics .
here we discuss in detail all of the steps needed to implement our new approach and we illustrate the whole technique using a non - trivial example the solution of the faddeev equations for @xmath4 scattering below and above the three - body breakup threshold with a realistic @xmath1 interaction . in our approach , we change all of the steps used in the conventional procedure for solving the faddeev equations in momentum space .
\(i ) the first step is to replace the conventional form of the faddeev integral equation , e.g. for a transition operator @xmath5 describing elastic @xmath3 scattering @xcite , @xmath6 with the half - shell equivalent form @xmath7 here @xmath8 is the two - body interaction , @xmath9 is the two - body @xmath10-matrix , @xmath11 is the resolvent of the free three - body hamiltonian , @xmath12 , @xmath13 is the particle permutation operator and @xmath14^{-1}$ ] is the resolvent of the channel hamiltonian @xmath15 where @xmath16 is the two - body @xmath1 sub - hamiltonian and @xmath17 is sub - hamiltonian describing free motion of the third nucleon relative to the @xmath1 subsystem .
the index @xmath18 is the jacobi - set index of the initial state .
one of the main purposes for such a replacement is to change the required two - body input : instead of fully off - shell two - body @xmath10-matrices at many energies , we suggest employing two - body interactions @xmath8 in combination with the channel resolvent @xmath19 .
however , in such a replacement , one has to evaluate additionally the channel resolvent operator @xmath19 .
fortunately , in the wave - packet approach , the finite - dimensional approximation for this operator is calculated easily in a closed analytical form @xcite .
moreover , the whole energy dependence appears now in the channel resolvent operator ( rather than in the off - shell @xmath10-matrix , as in the conventional formulation ) , which is calculated explicitly .
so that , with such a replacement , we can find a solution of few - body scattering equations _ at many energies almost with the same computational effort that is needed for a single energy_. another important advantage of our approach is a new treatment of three - body breakup .
contrary to the conventional approach , we treat three- or many - body breakup processes as particular cases of inelastic excitations ( into states of the discretized continuum ) @xcite .
such a treatment strongly facilitates breakup calculations .
\(ii ) the second step is to project the integral kernels of the reformulated faddeev equations and the solution onto a special orthogonal basis of the stationary wave packets ( wps ) , which corresponds to a formulation of the scattering problem on a momentum lattice .
such basis is very appropriate for constructing normalized analogs of continuum states for the channel hamiltonian @xmath20 .
it follows that the solution of the three - body scattering problem is described in the terms of asymptotic channel states , in contrast to conventional approach which employs free three - body states ( the plane waves ) .
such a projection of the three - body scattering equations onto a three - body wp basis results in matrix equations which allow us to circumvent the main difficulties that arise in the conventional solution of the initial singular integral equation .
firstly , the use of a finite matrix for the permutation operator in a discrete wp basis eliminates the need for the very numerous multi - dimensional interpolations of a given solution into the `` rotated '' jacobi set during iterations .
further , all singularities of the faddeev kernel ( in the form @xmath21 ) are isolated now in the channel resolvent @xmath22 and thus can be easily smoothed and averaged when using the wp representation @xcite . at last , the resulting matrix equations can be solved directly at real energies without any contour rotations or deformations onto complex plane , which are often employed in the solution of singular integral equations .
the resulting matrix equation ( of high dimension ) obtained in the wp approach is solved by simple iterations when they converge or otherwise by applying an additional pade - approximant summation .
the computational scheme turns out to be very efficient and thus the whole calculation can be performed even on a standard pc .
following these steps , in our previous papers , @xcite , we studied elastic scattering and breakup cross sections in a @xmath2 system with a central @xmath1 potential .
but it was still unclear if the advantages of the above computational scheme remain valid for realistic @xmath1 interactions including tensor , spin - orbit etc .
components and in the particularly when the number of contributing spin - orbital partial channels is large . to investigate this question
we will apply our approach to a three - nucleon system interacting with a realistic @xmath1 interaction , including a tensor component ( the nijmegen @xmath1 potential @xcite ) , at energies below and above three - body breakup threshold .
\(iii ) to further extend the complexity of scattering problems that can be treated accurately , it is desirable to develop a highly parallel algorithm for the solution of the resulting matrix equations of large dimension . _ the third step _ is to parallelize the algorithm to adapt it for computations by a gpu .
such a gpu realization is shown in the present paper to make the solution of resulting matrix equations ( derived from multi - channel system of integral faddeev equations ) extremely fast even on a standard pc .
here we briefly describe our approach based on a continuum discretization using the stationary wave packets .
we illustrate this using the example of the faddeev equation ( [ ueq ] ) for the transition operator @xmath5 for @xmath3 scattering ( further details see in @xcite ) . to construct the three - body wp basis functions , we start from the two - body case and introduce partitions of the continua of two free sub - hamiltonians , @xmath23 and @xmath17 , onto non - overlapping intervals @xmath24\}_{i=1}^m$ ] and @xmath25\}_{j=1}^n$ ] respectively .
these sub - hamiltonians describe the free motion of particles 2 and 3 with relative momentum @xmath26 and the free motion of particle 1 with momentum @xmath27 relative to the center of mass of the pair ( 23 ) , respectively .
thus the free stationary wave packets @xmath28 and @xmath29 are built as integrals of free solutions @xmath30 and @xmath31 over the discretization bins : @xmath32 where @xmath33 and @xmath34 are normalization factors and weight functions respectively @xcite . here and below we denote the functions and values corresponding to @xmath27-variable with additional bar mark to distinguish them from the functions corresponding to the @xmath26-variable . when constructing the three - body wp basis one should take into account spin and angular parts of the basis functions .
we use the following quantum numbers for the subsystems and the whole three - body system according to the @xmath35-coupling scheme : @xmath36 where @xmath37 and @xmath38 are @xmath1 quantum numbers : @xmath39 is an orbital momentum , @xmath40 is a spin and @xmath41 is a total angular momentum of the subsystem ( the interaction potential depends on the value of @xmath38 ) .
the other quantum numbers are the following : @xmath42 is an orbital momentum and @xmath43 is a total momentum of the third nucleon , where @xmath44 is its spin .
finally , @xmath45 is a total angular momentum of the three - body system , @xmath46 is the total isospin and @xmath47 is parity , all of them are conserved .
let s also note that the pair isospin @xmath10 can be defined by values of @xmath39 and @xmath40 , because the sum @xmath48 must be odd .
the free wp states should be defined for each partial wave @xmath39 and @xmath49 and further they are multiplied by appropriate spin - angular states .
thus the three - body basis function can be written as : latexmath:[\[\label{xij } spin - angular state of the @xmath1 pair , @xmath51 is a spin - angular state of the third nucleon , and @xmath52 is a set of three - body quantum numbers .
the state ( [ xij ] ) is the wp analog of the exact plane wave state in three - body continuum @xmath53 for the three - body free hamiltonian @xmath54 .
the free stationary wave packets defined in eq .
( [ pq ] ) with unit weights are step - like functions in the momentum representation @xcite while the three - body free wp basis functions are constant inside the cells of the lattice built by a convolution of two one - dimensional cells @xmath55 and @xmath56 .
we refer to the free wp basis as _ a lattice _ basis .
we denote the two - dimensional bins ( i.e. the lattice cells ) by @xmath57 . in the case of a single - channel two - body input interaction (
e.g. the central one ) , we have demonstrated @xcite that it is possible to define _ scattering wps _ corresponding to the exact scattering wave functions @xmath58 of the sub - hamiltonian @xmath59 : @xmath60 are partition intervals and @xmath61 and @xmath62 are a normalization factor and a weight function . to use the states ( [ z_k ] ) practically
, one can approximate them with the pseudostates of the sub - hamiltonian @xmath59 in some @xmath0 basis @xcite .
also it has been shown that the free wp basis is very appropriate to approximate scattering states because the respective functions have a very long - range behavior in configuration space .
so we can calculate the eigenstates ( the bound and pseudostates ) of the sub - hamiltonian @xmath59 matrix in the two - body wp - basis @xmath63 via a diagonalization procedure . as a result one
gets the eigenstates of the @xmath59 sub - hamiltonian expanded in the free wp basis ( for each partial wave @xmath39 ) : @xmath64 for the case of a central interaction , the three - body quantum numbers for the channel hamiltonian @xmath20 are the same as those for the three - body free hamiltonian @xmath12 , so that total three - body wp states corresponding to the channel hamiltonian @xmath20 ( three - body scattering wave packets swp ) are built as direct products of the two - body wps with the same spin - angular quantum numbers @xmath65 : @xmath66 the main advantage of the basis constructed from swp ( [ 3swp ] ) is that one gets _
an explicit analytical _ and even _ diagonal form _ for the matrix of three - body channel resolvent @xmath19 @xcite .
having the wp basis for the channel - hamiltonian ( [ 3swp ] ) at our disposal , it is possible to project all of the constituents of the integral equation ( [ ueq ] ) and find its finite - dimensional , i.e. matrix , analog . as an important result of the projecting of the channel operator @xmath67 onto the swp - basis , one gets the main part of the faddeev kernel matrix in a convenient analytical form , with a completely analytical energy dependence , in sharp contrast to a conventional approach with a fully off - shell @xmath10-matrix in a numerical form .
the permutation operator matrix @xmath68 in the three - body swp basis @xmath69 can be expressed through the overlap matrix @xmath70 in the free wp basis @xmath71 using the rotation matrices @xmath72 from the expansion ( [ exp_z ] ) : @xmath73 a matrix element of the operator @xmath74 in the free wp basis is equal to the overlap between basis functions defined in different jacobi sets .
such a matrix element can be calculated by integration with the basis functions over the momentum lattice cells : @xmath75 where the prime on the lattice cell @xmath76 indicates that the cell belongs to the other jacobi set while @xmath77 is the kernel of particle permutation operator in momentum space .
this kernel , as is well known @xcite , is proportional to the product of a dirac delta and a heaviside theta function .
however , due to integration in the eq .
( [ perm ] ) , these singularities get averaged over the momentum lattice cells and , as a result , the elements of the permutation operator matrix in the wp basis are finite and regular . the matrix element in the eq .
( [ perm ] ) can be calculated using a double numerical integration .
the practical technique of such calculation is described in the appendix to the present paper .
having evaluated the matrix of permutation operator @xmath74 , the calculation of the kernel @xmath21 matrix becomes fast and straightforward . as a result of projecting the integral equation ( [ ueq ] ) onto the three - body swp basis ,
one gets its matrix analog ( for the each set of three - body quantum numbers @xmath78 ) : @xmath79 here @xmath80 , @xmath81 and @xmath82 are matrices of the permutation operator , the pair interaction and the channel resolvent respectively defined in the swp basis .
the on - shell elastic amplitude for the @xmath3 scattering in the wp representation is defined by a diagonal matrix element of the @xmath83-matrix @xcite : @xmath84 where @xmath85 is the nucleon mass , @xmath86 is the initial two - body momentum and @xmath87 is the swp basis state corresponding to the initial scattering state . here
@xmath88 is the bound state of the pair ( the deuteron , in our case ) , the index @xmath89 denotes the bin @xmath90 , including the on - shell momentum @xmath86 , and @xmath91 is the momentum width of this bin . it should be noted here that in our discrete wp approach _ the three - body breakup is treated as a particular case of an inelastic scattering _
@xcite ( defined by the transitions to the specific two - body discretized continuum states ) , so that the breakup amplitude can be defined in terms of _ the same matrix _
@xmath83 determined from eq .
( [ m_pvg ] ) . for the case of the tensor components of the @xmath1 interaction ( or other coupled - channel two - body interactions @xmath8 ) , the generalization of the whole formalism is straightforward .
however , it is necessary to take into account some new aspects related to the discretized spectrum of the two - body _ multichannel _ hamiltonian @xmath59 to build the correct approximation for the discretized resolvent @xmath19 .
the three - body swp states corresponding to the channel hamiltonian @xmath20 can be defined similarly to the one - channel two - body interaction case , i.e. as direct products of two - body wp states for the @xmath17 and @xmath59 sub - hamiltonians ( jointly with the bound - state ) multiplied by spin - angular functions of the system .
however , here the possible spin - angular couplings in the @xmath92 subsystem due to the tensor component in a pairwise interaction @xmath8 should be taken into account .
recently , the present authors have developed a convenient approach for solving multichannel scattering problems via a straightforward diagonalization of the multichannel hamiltonian in a wp basis the discrete spectral - shift ( dss ) formalism @xcite . in this approach ,
the multichannel hamiltonian pseudostates have been shown to correspond to scattering wave functions defined in the so called eigenchannel representation ( er ) @xcite for which the multichannel @xmath93-matrix is diagonal .
let s introduce two types of scattering states of the sub - hamiltonian @xmath59 : the scattering states @xmath94 ( which include radial and orbital parts ) corresponding to the initial wave with a definite orbital momentum @xmath39 and the scattering states defined in the er @xmath95 , where @xmath96 is an eigenchannel index . in case of a tensor @xmath1 interaction , the scattering states @xmath95
are linear combinations of states @xmath94 , e.g. the coupled pairs @xmath97 , @xmath98 etc .
the main advantage of the er formalism is that one gets the following spectral expansion for the resolvent of the pair sub - hamiltonian @xmath99^{-1}$ ] ( see e.g. @xcite ) : @xmath100 which is _ diagonal _ in the eigenchannel index @xmath96 . in eq .
( [ 2g1 ] ) , @xmath101 is an energy of the bound state @xmath102 and @xmath85 is the nucleon mass .
thus , it is convenient to construct multi - channel two - body swp s as integrals of exact scattering wave functions of the sub - hamiltonian @xmath59 defined in the er : @xmath103 where @xmath104 $ ] are new partition intervals whose parameters might be different for different @xmath105 .
the crucial feature of the dss formalism defined in @xcite is that these multichannel swp can be constructed ( jointly with the deuteron bound state wave function ) as pseudostates of the hamiltonian @xmath59 matrix in the multichannel free wp basis @xmath106 ( where radial and angular parts of wave functions included ) . because @xmath39 is not conserved , for each value of total momentum @xmath38 , the two - channel @xmath59 sub - hamiltonian states are constructed from free wp bases @xmath107 defined for both possible values of @xmath108 and @xmath109 .
finally , we have at our disposal the multichannel swp basis functions which are related to the free wps by a simple orthogonal transformation as in the one - channel case ( [ exp_z ] ) : @xmath110 where the spin - angular parts of wave functions are taken into account as well and the multi - index @xmath111 related to the er is introduced .
below we will not detail the index @xmath112 which is a part of the multi - index @xmath113 . in this way , we construct the three - body swp basis functions for the channel hamiltonian @xmath20 as direct products of the two - body ones for the @xmath17 and @xmath59 sub - hamiltonians : @xmath114 of the channel hamiltonian @xmath20 .
hence , starting from free wp bases for each two - body sub - hamiltonian one gets a set of basis states both for the three - body free and channel hamiltonians , @xmath12 and @xmath20 respectively , which are related to each other by a simple matrix rotation .
the spectral expansion of the three - body channel resolvent @xmath19 in the er can be found straightforwardly by making a convolution of the multi - channel subresolvent @xmath115 with the resolvent @xmath116^{-1}$ ] for the free sub - hamiltonian @xmath17 of the third nucleon @xmath117 this leads to the following representation for the multi - channel three - body resolvent operator @xmath19 via the scattering eigenfunctions of the @xmath1 subsystem defined in the er : @xmath118 the first term ( the bound - continuum part ) in eq .
( [ g11 ] ) is a spectral sum over three - body states corresponding to the free motion of the third nucleon relative to the deuteron .
the second term ( the continuum - continuum part ) in eq .
( [ g11 ] ) for the channel resolvent includes channel three - body states with @xmath1 pair interacting in the continuum ( in the er ) and its imaginary part is defined by a discontinuity across the three - body cut on the riemann energy surface . by projecting the channel resolvent operator onto the three - body swp basis defined in eq .
( [ si ] ) , one can find analytical formulas for matrix elements of operator @xmath19 in such a basis .
the respective matrix is diagonal in all wave - packet and spin indices : @xmath119 here the diagonal matrix elements @xmath120 depend , in general , on the spectral partition parameters ( i.e. @xmath121 and @xmath122 values ) and total energy @xmath123 only .
these matrix elements do not depend explicitly on the interaction potential @xmath8 .
the matrix elements @xmath120 are defined as integrals over the respective momentum intervals for the bound - continuum part of the whole operator : @xmath124 and for the continuum - continuum part : @xmath125 if the solution of the scattering equations in the finite - dimensional wp basis converges with increasing the basis dimension , the final result turns out to be _ independent _ of the particular spectral partition parameters
. representations ( [ dg1]a ) and ( [ dg1]b ) for the channel resolvent are the basis of the wave - packet approach , since , after a straightforward analytical evaluation of the integrals ) when one uses the wp s with the weight functions @xmath126 , @xmath127 . ] in ( [ dg1]a ) and ( [ dg1]b ) , one gets explicit formulas for the three - body resolvent and thus a drastic simplification of the solution of a general three - body scattering problem .
these analytical expressions can be used directly to solve the matrix faddeev equation in eq .
( [ m_pvg ] ) .
in this section the effectiveness of the new approach in few - body calculations will be illustrated by solving the faddeev equations for @xmath3 scattering with the realistic nijmegen @xmath1 potential @xcite .
the @xmath93-matrix of elastic @xmath128 scattering is usually parameterized by the eigen phase shifts and mixing angles in the total channel spin @xmath129 representation ( here @xmath40 is the @xmath1 pair spin and @xmath130 is a spin of the third nucleon ) .
unitary transformation between amplitudes defined in the representation of the total angular momentum of the third nucleon @xmath131 and the @xmath132-representation is given by the matrix @xcite : @xmath133 in the figs .
[ phases]-[ay ] the results of our fully discrete calculations for elastic @xmath3 scattering with the nijm i @xmath1 interaction performed within the wp approach are compared with the results of conventional faddeev calculations of the bochum
krakow group @xcite . in the example
, we restricted ourselves to the total isospin value @xmath134 and took into account all of the pairwise @xmath1 channels with two - body total angular momentum @xmath135 ( this gives up to 54 spin - angular channels ) . in fig .
[ phases ] the results for the lowest even partial phase shifts @xmath136 of elastic @xmath3 scattering both below and above a three - body breakup threshold are shown . in the example
, we employed wp bases with dimensions @xmath137 along two jacobi coordinates .
this gives a matrix system with the dimension @xmath138 which has been handled easily on an ordinary pc in contrast to typical conventional calculations for similar faddeev system which require supercomputer facilities ( see e.g. @xcite ) .
it should be especially emphasized that the calculation of the phase shifts at 100 different energy values displayed in fig .
[ phases ] takes in our approach only about twice as much time as the calculation for a single energy because for all energies we employ the same permutation matrix @xmath68 which is calculated only once . in fig .
[ cross13 ] our results for the differential cross section for elastic @xmath3 scattering at 13 mev are compared to the results of conventional faddeev calculations . ]
@xcite . while in fig .
[ ay ] the same comparison is given for the neutron vector analyzing powers @xmath139 for the elastic @xmath3 scattering at 35 mev . here the wp basis with dimension @xmath140 has been used and the partial waves with the total angular momentum up to @xmath141 have been taken into account . it is clear from all of the above illustrative examples that one can achieve almost perfect agreement between our results and the results of the standard faddeev calculations @xcite performed on a powerful supercomputer .
the present calculations were performed on the serial pc with an intel i7 - 3770k ( 3.50ghz ) processor with 32 gb of ram .
the real cpu computational time ( which includes the permutation matrix evaluation ) , i.e. the total calculation which starts from two - body interaction potential and ends with partial three - body amplitudes for a 54-channel calculation with the basis dimension @xmath142 for a single value of the total angular momentum @xmath143 and parity takes about 7 minutes .
although the whole calculation for many partial waves will take a longer time , there is the possibility within the present approach to accelerate the whole calculation and thus solve more complicated scattering problems already using an ordinary pc .
so , keeping in mind our general aim to simplify and accelerate drastically realistic many - body scattering calculations in nuclear , atomic , molecular etc .
studies by using a discrete matrix reduction of the integral scattering equations in the wp scheme , we propose to employ the ultra - fast gpgpu technique to further optimize the solution of the resulting matrix equations .
in this section , we demonstrate a high efficiency of using a graphics processor in numerical solution of the above matrix equation for the three - body scattering problem . first we describe the overall numerical scheme for solving the three - body scattering problem in the wp formalism , paying particular attention to those aspects that are important for an efficient parallelization .
the use of a fixed matrix for the permutation operator completely eliminates the necessity of the numerous and time consuming interpolations of a current solution in the iterations .
these computations take the majority of the computing time in the standard integral approach .
contrary to this , in our approach , the main computational effort ( in case of cpu - calculations ) are spent on a calculation of the permutation matrix @xmath70 in the lattice wp basis .
however , the matrix @xmath144 is independent of energy , and therefore , being calculated once , it can be used to solve the scattering problem at many energies , as well as for various two - particle input interactions . while in the standard approach , the whole calculation must be repeated for each energy and for each type of two - body interaction .
the main difficulty which is met in the practical solution of the matrix equation ( [ m_pvg ] ) is its large dimension .
so , it is impossible even to store the entire matrix of the kernel @xmath145 in a ram of a computer .
however , one can effectively employ the fact that the matrix @xmath146 can be represented as a product of four matrices : a very sparse matrix @xmath70 ( only about 1 % of its matrix elements are nonzero @xcite ) , a diagonal matrix @xmath19 and two block matrices .
so that , it is sufficient to store only factors of the matrix @xmath147 instead of all its matrix elements ( only the nonzero elements of the sparse matrix are stored ) . in this way , the required memory can be reduced by about 100 times .
this eliminates the need for an external memory in the process of iteration , which in turn reduces a computational time by an order of magnitude even for a conventional cpu realization .
so , our overall numerical scheme should be quite suitable for parallelization and implementation on the multiprocessor systems , particularly on a gpu .
the optimized algorithm for the solution of the realistic @xmath3 scattering problem consists of the following main steps : + 1 .
construction of the three - body swp basis including preparation of two - body bases ( via diagonalization of the pairwise @xmath1 sub - hamiltonian matrix in the free wp basis ) ; the calculation of the algebraic coefficients @xmath148 ( [ gal ] ) for the coupling of different spin - angular channels .
selection of the nonzero elements of the overlap matrix @xmath70 .
calculation of the nonzero elements of @xmath70 .
calculation of the channel resolvent @xmath19 .
solution of the resulting matrix equation by iteration using the pade - approximant technique .
+ the runtimes for the steps 1 and 4 are negligible in comparison with the total running time , so that we can leave these steps for sequential execution on the cpu . the main computational effort ( in the cpu realization )
is spent just on the calculation of the matrix elements of @xmath149-matrix ( i.e. the step 3 ) which are calculated independently of each other .
so this step is ideal for parallelization and we primarily parallelized only the corresponding part of the computer code . however , since the matrix @xmath70 is very sparse , in order to reach the high efficiency of its parallel computation , the preliminary selection of its nonzero elements should be carried out ( step 2 ) .
the execution of the 5th step iteration solution of the matrix equation can certainly be accelerated using linear algebra routines implemented on the gpu , but as its execution time takes no more than 20% of the total time for solving the whole problem we did not optimize this step in the present study .
thus , in this calculation we have only parallelized the time - consuming steps 2 and 3 for the gpu . in this section
we compare the runtimes for the cpu and gpu realizations of realistic @xmath3 elastic scattering calculations . for our gpu calculations we used an ordinary nvidia gtx-670 video card , which is not specialized for general - purpose computing .
first we compare the cpu and gpu realizations of the above algorithm for case of the @xmath93-wave mt i - iii @xmath1 interaction .
this example clearly demonstrates the advantage of the gpu acceleration .
[ time_gpu ] shows the dependence of the cpu- and gpu - computing times for the above steps 2 , 3 , 5 and the complete solution for dimensions @xmath150 and @xmath151 of the wp bases chosen along both jacobi variables .
we take for simplicity @xmath152 in all of our tests so that , the total dimension of the matrix kernel is equal to @xmath153 .
it is clear from fig .
[ time_gpu ] that the complete solution of @xmath3 scattering problem in our approach for a basis of the large dimension is quite enough .
] @xmath154 takes only ca . 7 sec on a serial pc using gpu !
this ultra - fast solution handles a huge matrix @xmath155 including the calculation of the ca .
@xmath156 millions nonzero elements of matrix @xmath70 where each matrix element is reduced to a integral of a rather complicated algebraic function .
the integrals are calculated numerically with a 48-grid - point gaussian quadrature . from fig .
[ time_gpu ] it is seen that all of these 256 million integrals can be computed on gpu in just 2.3 seconds compared to 255 seconds on the cpu .
this demonstrates the real very high speed of gpu computations for the case discussed here . in fig .
[ accel_gpu ] we present the dependence of the gpu acceleration ratio @xmath157 on the dimension of the basis @xmath152 for the solution of the @xmath93-wave faddeev problem .
the total acceleration for the complete solution varies from 10 to 50 times depending on the basis dimension , while the time for calculating the nonzero elements of the matrix @xmath70 ( the step 3 ) which takes the main part of the cpu - computing time is reduced by a factor of more than 100 .
we now turn to the case of the realistic @xmath3 scattering problem with the nijmegen @xmath1 potential . taking into account higher partial waves
leads to the system of coupled two - dimensional integral faddeev equations . now
the calculation of each matrix element of @xmath70 for partial waves with nonzero angular momenta includes several tens of double numerical integrals with some trigonometric functions and uses a large set of algebraic coefficients @xmath158 ( [ gal ] ) .
the proper parallelization for this calculation with pre - selection of the nonzero matrix elements leads again to rather fast algorithm realized on gpu .
[ time_gpu_nijm ] demonstrates the gpu acceleration ratio @xmath157 for the complete solution ( solid line ) and for the steps 2 ( dashed line ) and 3 ( dot - dashed ) as a function of the basis dimension @xmath152 for the solution of the 18-channel faddeev equations for the @xmath159 partial @xmath3 elastic amplitude .
it is evident from the results presented that the passing from cpu- to gpu - realization _ on the same pc _ allows one to obtain a significant acceleration of whole three - body calculation by an order of magnitude or higher .
it is clear also that the use of a more powerful specialized graphics processor , like the tesla k40 , would lead even to a considerably higher acceleration of calculations .
however , it should be emphasized that the total acceleration which can be achieved by using gpu depends crucially on the method used , the numerical scheme and parallelization implementation .
in conclusion , let s summarize the most important points of the proposed new general approach for the solution of few - body scattering problems .
first , we rewrite the initial faddeev equations , which include an off - shell @xmath10-matrix , in a fully equivalent form which incorporates the product of the channel resolvent and the interaction operator .
this allows us to simplify the required two - body input and also to fix all of the energy singularities in the channel resolvent operator .
we note here that the four- and more particle integral faddeev - yakubovsky equations have , in principle , some similar structure as compared to three - body faddeev equations ( surely being more complicated and having a higher dimension ) . therefore the discrete wp technique outlined in this paper can be employed for solving these general equations as well . \2 .
after that , we project out the transition and interaction operators onto a discrete wave - packet basis and employ a specific analytical form for one- and multi - channel resolvent operators .
this discrete form is extremely convenient for few - body calculations because we get a fixed matrix form ( with fully regular matrix elements ) for the scattering equations ( of lippmann schwinger , faddeev or faddeev - yakubovsky types ) .
the most important improvement in such a wp projection is the fact that we can replace the permutation operators by fixed matrices , thus completely avoiding the numerous and time consuming interpolations at every iteration step .
moreover , this matrix is energy- and interaction - independent and , being calculated once , can be used to solve various scattering problems with different potentials and at many energies .
these two steps lead to a rather effective numerical scheme which is realized on a standard pc for realistic @xmath2 scattering calculations .
\3 . at last ,
we perform a proper parallelization of the solution and eventually we employ the ultra - fast gpu technique to make very effective parallel calculations on a serial pc .
this gpu realization for realistic scattering problems reduces the computational time by at least one order of magnitude ( while for separate parts of the numerical scheme the total acceleration could reach two orders ) .
all of the above points open a new way for doing extensive many - body scattering calculations via continuum discretization e.g. in quantum chemistry , nuclear reaction theory , solid state theory etc .. there is little doubt that a similar ( but surely more tedious ) approach can also be applied to other scattering problems , e.g. to solve the relativistic faddeev equations , bethe
salpeter equations etc .
* acknowledgments * the authors thank dr .
boreskov for discussions of problems associated with a use of the gpu .
the authors are also deeply grateful to prof .
w. polyzou for careful reading of the manuscript and valuable comments .
this work has been supported partially by the russian foundation for basic research , grants nos .
12 - 02 - 00908 and 13 - 02 - 00399 .
the kernel of the permutation operator in momentum space has the form @xcite @xmath160 where the multi - indices @xmath161 and @xmath162 include all possible spin - angular quantum numbers for the three - body states ( as they are detailed in the sec.iv , i.e. @xmath163 ) and the following notations @xmath164 are used .
the spin - angular coefficients @xmath165 can be found from the formula @xmath166 where @xmath39 and @xmath167 are two - body subsystem orbital momenta and the indices @xmath168 arise from intermediate triangle sums .
the sum in ( [ ggaga ] ) runs according to the triangle rules in partial coefficient @xmath158 which have the following explicit form @xcite : @xmath169 where the summation is done over all allowed intermediate quantum numbers .
projection of free three - body wave - functions @xmath170 ( plane waves ) onto the wp states can be found ( for the case of unit weights @xmath171 and @xmath172 ) from the formula @xcite : @xmath173 here @xmath174 , @xmath175 are the widths of momentum intervals and function @xmath176 is analog of the heaviside step - function : @xmath177 if @xmath26 belongs to the interval @xmath178 and @xmath179 otherwise . to obtain the permutation matrix elements in the lattice basis ( [ perm ] ) , one should integrate kernel ( [ p_kern ] ) over the lattice cells @xmath180 and @xmath181 .
these matrix elements are defined by the following integrals : @xmath182_{ij , i'j'}^{{\gamma}{\gamma}'}\equiv \langle x_{ij}^{\gamma}|p|x_{i'j'}^{{\gamma}'}\rangle=\sum_{kl_1l'_1}\frac{g_{{\gamma}{\gamma}'}^{kl_1l'_1 } } { \sqrt{d_i\bar{d}_j d_{i'}\bar{d}_{j'}}}\times\nonumber\\ \int_{{\mathfrak{d}}_{ij } } dpdq \int_{{\mathfrak{d}}_{i'j'}}dp'dq ' \int_{-1}^1dx \frac{q^{l_2+l'_2 + 1}(q')^{l_1+l'_1 + 1}}{(p')^{l'+1}p^{l+1}}\times\nonumber\\ \delta(p'-\pi_1)\delta(p-\pi_2)p_k(x).\qquad\end{aligned}\ ] ] using the dirac delta - functions , the integrals over @xmath26 and @xmath183 are evaluated analytically : @xmath184_{ij , i'j'}^{{\gamma}{\gamma}'}=\sum_{kl_1l'_1}\frac{g_{{\gamma}{\gamma } ' } } { \sqrt{d_i\bar{d}_j d_{i'}\bar{d}_{j'}}}\int_{{\bar{\mathfrak{d}}}_{j } } dq \int_{{\bar{\mathfrak{d}}}_{j'}}dq ' \int_{-1}^1dx\nonumber\times\\ \frac{q^{l_2+l'_2 + 1}(q')^{l_1+l'_1 + 1}}{(\pi_1)^{l'+1}(\pi_2)^{l+1}}p_k(x)\theta(\pi_1\in{\mathfrak{d}}_{i ' } ) \theta(\pi_2\in{\mathfrak{d}}_i).\qquad\end{aligned}\ ] ] further , the residual integrals can be evaluated in polar coordinates : @xmath185 then the matrix element takes the form : @xmath186_{ij , i'j'}^{{\gamma}{\gamma}'}=\sum_{kl_1l'_1}\frac{g_{{\gamma}{\gamma } ' } } { \sqrt{d_i\bar{d}_j d_{i'}\bar{d}_{j ' } } } i,\ ] ] where @xmath187 and the following notations are employed : @xmath188 the boundaries of the integral area in @xmath189-plane are defined by transformation of the rectangle in the @xmath190-plane , so that : @xmath191 evaluating the @xmath192-integral analytically , one gets the expression for the integral @xmath131 : @xmath193 where the following function is introduced : @xmath194 ^ 2 -\nonumber\\ { { \frac{1}{2}}}\left[\max\left(\frac{q_{j-1}}{\cos\phi},\frac{q'_{j'-1}}{\sin\phi } , \frac{p'_{i'-1}}{\zeta_1},\frac{p_{i-1}}{\zeta_2}\right)\right]^2.\qquad \qquad\end{aligned}\ ] ] finally , one gets the eventual formula for the permutation matrix @xmath195_{ij , i'j'}^{{\gamma}{\gamma}'}=\sum_{kl_1l'_1}\frac{g_{{\gamma}{\gamma}'}^{kl_1l'_1 } } { \sqrt{d_i\bar{d}_j d_{i'}\bar{d}_{j'}}}\int_{-1}^1dx p_k(x)\times\nonumber\\ \int_{\phi_1}^{\phi_2}d\phi f_{ij , i'j'}(\phi , x ) \frac{(\cos\phi)^{l_2+l'_2 + 1}(\sin\phi)^{l_1+l'_1 + 1}}{(\zeta_1)^{l'+1}(\zeta_2)^{l+1}}.\end{aligned}\ ] ] in practical treatments these integrals are evaluated numerically .
abs j. carbonell , a. deltuva , a.c .
fonseca , r. lazauskas , progr .
. phys . * 74 * , 55 ( 2014 ) .
faddeev , sov .
phys.- jetp * 12 * , 1014 ( 1961 ) ; o.a .
yakubovsky , sov . j. nuclear . phys . * 5 * , 937 ( 1967 ) .
h. witaa , w.glckle , phys .
c * 85 * , 064003 ( 2012 ) .
https://developer.nvidia.com/cuda-zone k.a .
wilkinson , p. sherwood , m.f .
guest , k.j .
naidoo , j. comp .
chem . * 32 * , 2313 ( 2011 ) .
clark , r. babich , k. barrose , r.c .
brower , c. rebbi , comp .
phys . com .
* 181 * , 1517 ( 2010 ) .
h. potter et al .
, proceedings of ntse-2013 ( ames , ia , usa , may 13 - 17 2013 ) , eds .
a.m. shirokov and a.i .
mazur , khabarovsk , russia , p. 263
( 2014 ) ; http://www.ntse-2013.khb.ru/proc/sosonkina.pdf .
e. yarevsky , lncs 7125 , mathematical modeling and computational science , eds .
a. gheorghe , j. bua and m. hnatic , p. 290
pomerantsev , v.i .
kukulin , o.a .
rubtsova , phys .
c * 79 * , 064602 ( 2009 ) ; _ ibid .
_ * 79 * , 034001 ( 2009 ) .
rubtsova , v.n .
pomerantsev , v.i .
kukulin , a. faessler , phys .
c * 86 * , 034004 ( 2012 ) . w. glckle , h. witaa , d.hber , h. kamada , j. golack , phys .
rep . * 274 * , 107 ( 1996 ) .
stoks , r.a.m .
klomp , c.p.f .
terheggen , j.j de swart , phys .
c * 49 * , 2950 ( 1994 ) .
rubtsova , v.i .
kukulin , v.n .
pomerantsev , a. faessler , phys .
c * 81 * , 064003 ( 2010 ) .
m. danos , w. greiner , phys .
rev . * 146 * , 708 ( 1966 ) .
rubtsova , v.i .
kukulin , v.n .
pomerantsev , physics of atomic nuclei * 77 * , 486 ( 2014 ) .
kukulin , v.n .
pomerantsev , o.a .
rubtsova , theor .
* 150 * , 403 ( 2007 ) .
s.n . bunker _
et al . _ ,
a * 113 * , 461 ( 1968 ) . | background : : : the numerical solution of few - body scattering problems with realistic interactions is a difficult problem that normally must be solved on powerful supercomputers , taking a lot of computer time .
this strongly limits the possibility of accurate treatments for many important few - particle problems in different branches of quantum physics .
purpose : : : to develop a new general highly effective approach for the practical solution of few - body scattering equations that can be implemented on a graphics processing unit . methods : : : the general approach is realized in three steps : ( i ) the reformulation of the scattering equations using a convenient analytical form for the channel resolvent operator ; ( ii ) a complete few - body continuum discretization and projection of all operators and wave functions onto a @xmath0 basis constructed from stationary wave packets and ( iii ) the ultra - fast solution of the resulting matrix equations using graphics processor . results : : : the whole approach is illustrated by a calculation of the neutron - deuteron elastic scattering cross section below and above the three - body breakup threshold with a realistic @xmath1 potential which is performed on a standard pc using a graphics processor with an extremely short runtime .
conclusions : : : the general technique proposed in this paper opens a new way for a fast practical solution of quantum few - body scattering problems both in non - relativistic and relativistic formulations in hadronic , nuclear and atomic physics . |
the phenix experiment at the relativistic heavy ion collider has made measurements of several event - by - event fluctuation observables .
these measurements include the following : fluctuations in the net charge , which are sensitive to the charge distribution within the collision volume ; fluctuations in the mean transverse momentum , which are sensitive to critical fluctuations in the temperature of the system and may be large in the presence of a phase transition from hadronic matter to a quark - gluon plasma ( qgp ) ; and fluctuations in charged particle multiplicity , which appear to exhibit non - monotonic behavior as a function of centrality at sps energies .
however , each of these fluctuation observables are also sensitive to known processes such as charge conservation , resonance decays , elliptic flow , and hard processes .
therefore the contribution of these known processes must be thoroughly investigated before reliably interpreting the results of the measurement .
details about the phenix experimental configuration can be found elsewhere @xcite .
all of the measurements described here utilized the phenix central arm detectors .
the phenix acceptance with a maximum acceptance of @xmath0 in pseudorapidity and @xmath1 in azimuthal angle is considered small for event - by - event measurements .
however , the event - by - event multiplicities are high enough in rhic heavy ion collisions that phenix has a competitive sensitivity for the detection of fluctuation signals . for example , a detailed examination of the phenix sensitivity to temperature fluctuations derived from the measurement of mean @xmath2 fluctuations is described in @xcite .
a promising signature for the presence of a qgp is the observation of event - by - event fluctuations in the net charge .
it has been hypothesized that fluctuations in the net charge would be significantly reduced in a qgp scenario , where the fractional electric charges of the quarks are more evenly spread throughout the qgp volume than the unit electric charges of the hadrons in a hadronic gas volume @xcite .
phenix has measured net charge fluctuations in au+au collisions at @xmath3 = 130 and 200 gev @xcite .
the measurements have been made as a function of centrality , pseudorapidity range , and azimuthal angle range .
the fluctuations are quoted using the variable @xmath4 where @xmath5 is the net charge , @xmath6 is the total number of charged particles , and @xmath7 represents the variance of the net charge .
the expectations for the value of @xmath8 are 1.0 for random particle emission , 0.75 for a resonance gas , 0.83 for a quark coalescence scenario , and 0.25 for a qgp in small acceptance ranges @xcite .
@xmath8 as a function of azimuthal angle range for @xmath3 = 130 gev ( band ) and @xmath3 = 200 gev ( points ) au+au 0 - 10% central collisions .
the width of the band represents the statistical error from the 130 gev measurement . ]
@xmath8 as a function of azimuthal angle range for @xmath3 = 130 gev ( band ) and @xmath3 = 200 gev ( points ) au+au 0 - 10% central collisions .
the width of the band represents the statistical error from the 130 gev measurement . ]
shown in fig .
[ fig : qflucvscent ] are the phenix measurements as a function of centrality .
the small difference in the two datasets is due to the effect of tighter track cuts that effectively reduce the acceptance in the @xmath3 = 200 gev sample . for the 10% most central collisions with @xmath0 , @xmath9 mev / c , and @xmath10 , @xmath11 for @xmath3 = 130 gev and @xmath12 ( phenix preliminary ) for @xmath3 = 200 gev au+au collisions .
the value of @xmath8 must be adjusted for charge conservation by a factor of @xmath13 , where @xmath14 is the fraction of the charged particles produced within the acceptance range .
@xmath8 is also expected to be reduced by the decay of neutral resonances .
the reduction due to resonance decays should decrease with increasing geometrical acceptance since more particle pairs from larger opening angles will be included in the measurement .
[ fig : qflucvsphi ] shows the measurement of @xmath8 as a function of azimuthal angle range for the 10% most central collisions . the expected decrease in @xmath8 as the azimuthal range is increased
is observed .
the decrease is larger than the expectation from charge conservation alone by more than 3 standard deviations .
however , the decrease is consistent with results from rqmd simulations @xcite in the phenix acceptance , thus emphasizing the importance of the resonance decay contribution . as a result ,
phenix measurements of charge fluctuations are consistent with the expectations of statistically independent particle emission .
event - by - event fluctuations in the mean transverse momentum , @xmath15 , may be sensitive to temperature fluctuations near the qcd critical point @xcite .
phenix has measured the magnitude of these fluctuations in au+au collisions at @xmath3 = 130 gev and 200 gev as a function of centrality and @xmath2 @xcite .
phenix measurements are quoted in the variable @xmath16 in percent , where @xmath17 .
this observable can be related to other fluctuation variables currently in use with some reasonable assumptions @xcite . here , the baseline , random , distribution is constructed using mixed events .
it is important that the number of tracks , @xmath18 , for the mixed events are sampled directly from the data distribution since the standard deviation of the distribution of the event - by - event mean @xmath19 depends on @xmath18 .
also , no two tracks from the same data event are allowed to populate the same mixed event .
@xmath20 as a function of centrality is shown in fig .
[ fig : ptflucvscentmm ] .
a significant non - random fluctuation is observed for all centralities .
mean @xmath19 fluctuations as a function of @xmath2 range ( @xmath21 ) for @xmath3 = 200 gev au+au collisions .
the curve is the result of a pythia - based model that estimates the relative contribution due to jet production as a function of centrality . ]
mean @xmath19 fluctuations as a function of @xmath2 range ( @xmath21 ) for @xmath3 = 200 gev au+au collisions .
the curve is the result of a pythia - based model that estimates the relative contribution due to jet production as a function of centrality . ]
mean @xmath19 fluctuations as a function of @xmath2 range ( @xmath21 ) for @xmath3 = 200 gev au+au collisions .
the curves are the results of hijing 1.35 with jet suppression into the phenix acceptance . ]
mean @xmath19 fluctuations as a function of @xmath2 range ( @xmath21 ) for @xmath3 = 200 gev au+au collisions .
the curves are the results of hijing 1.35 with jet suppression into the phenix acceptance . ] since a positive signal is observed , contributions to the signal from known sources must first be understood .
contributions from hbt , resonance decays , and elliptic flow have been studied using simulations and are found to be very small or negligible .
another possible source of the signal is from hard scattering processes , which are expected to produce event - by - event correlations in @xmath19 , especially at higher @xmath19 . as shown in fig .
[ fig : ptflucvsptmm ] , the majority of the contribution to the fluctuations occur at high @xmath19 . there is a 243% increase in the signal when the maximum of the @xmath19 range over which the average @xmath19 is calculated increases from 1.0 to 2.0 gev / c .
this very large increase can not be attributed to statistical fluctuations , since the number of particles increases by less than 15% over that range . in order to investigate the contributions due to jet production , a data - driven two - component model consisting of 1 ) a simulation of soft processes by reproducing the inclusive @xmath19 and @xmath18 distributions as a function of centrality , and 2 ) a simulation of hard scattering processes by embedding pythia jet events at a given rate per produced particle from step 1 ( while conserving the @xmath18 distribution ) .
this rate is the only free parameter in the simulation . after adjusting the rate
so that the fluctuations match the data in a reference centrality bin , the simulation is run as a function of centrality within the phenix detector acceptance with a ) the pythia event embedding rate kept constant and b ) with the pythia event embedding rate scaled by the measurement of the nuclear modification factor , @xmath22 .
the results of the simulation superimposed with the phenix data are shown in fig .
[ fig : ptflucvscentmm ] and [ fig : ptflucvsptmm ] .
the simulation with the @xmath22-scaled rate agrees with the centrality - dependent ( and also the @xmath19-dependent ) trends remarkably well . within this simulation , the decrease of fluctuations in the most peripheral collisions
is attributed to the fact that the signal begins to compete with the magnitude of the statistical fluctuations , while most of the decrease of fluctuations in the most central collisions may be attributed to the onset of jet suppression . using the rate determined from the phenix data
, the simulation also reproduces the magnitude of fluctuations observed by star @xcite in 130 gev au+au collisions @xcite .
comparisons to the hijing model are shown in figs .
[ fig : ptflucvscenthij ] and [ fig : ptflucvspthij ] .
although hijing qualitatively reproduces the @xmath19-dependent trend , it does not reproduce the centrality dependence .
the reason for this is not yet understood .
one way to directly compare the fluctuation measurements from all of the experiments is to estimate the magnitude of any residual event - by - event temperature fluctuations using the prescription outlined in @xcite .
the most central collision fluctuation values correspond to maximum temperature fluctuations of 1.8% in phenix , 1.7% in star , 1.3% in ceres , and 0.7% in na49 .
disagreement in the ceres and na49 measurements may be due to differences in the pseudorapidity region over which the measurement was made . in all cases ,
the residual temperature fluctuations are small and do not significantly increase between sps and rhic energies .
interest in the topic of charged particle multiplicity fluctuations has been recently revived by the observation of non - monotonic behavior in the scaled variance as a function of system size at sps energies @xcite .
the scaled variance is defined as @xmath23 , where @xmath24 represents the variance of the multiplicity distribution in a given centrality bin , and @xmath25 is the mean of the distribution .
the scaled variance of a poisson distribution is 1.0 , independent of n. phenix has studied the behavior of charged particle multiplicity fluctuations as a function of centrality in @xmath3 = 62 gev and 200 gev au+au collisions in order to investigate if this behavior persists at rhic energies .
negatively charged particle multiplicity fluctuations in terms of the scaled variance as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions .
the error bars include statistical and systematic errors . ]
negatively charged particle multiplicity fluctuations in terms of the scaled variance as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions .
the error bars include statistical and systematic errors . ]
negatively charged particle multiplicity fluctuations in terms of the scaled variance as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions .
the error bars include statistical and systematic errors . ]
negatively charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions . ]
negatively charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions . ]
negatively charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of centrality for @xmath3 = 62 and 200 gev au+au collisions . ]
results from a hijing 1.35 simulation filtered through the phenix acceptance for inclusive charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit as a function of @xmath26 for @xmath3 = 200 gev au+au collisions . shown are results of the simulation with the default jet energy loss activated , without energy loss , and with jet production deactivated . ]
results from a hijing 1.35 simulation filtered through the phenix acceptance for inclusive charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit as a function of @xmath26 for @xmath3 = 200 gev au+au collisions . shown are results of the simulation with the default jet energy loss activated , without energy loss , and with jet production deactivated . ] inclusive charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of @xmath2 over the range 200 mev / c @xmath27 for @xmath3 = 62 and 200 gev au+au 25 - 30% central collisions . ] inclusive charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of @xmath2 over the range 200 mev / c @xmath27 for @xmath3 = 62 and 200 gev au+au 25 - 30% central collisions . ] inclusive charged particle multiplicity fluctuations in terms of the k parameter from a negative binomial distribution fit to the data as a function of @xmath2 over the range 200 mev / c @xmath27 for @xmath3 = 62 and 200 gev au+au 25 - 30% central collisions . ]
charged particle multiplicity fluctuations in terms of the scaled variance as a function of centrality are shown in fig .
[ fig : mfvscentall ] for inclusive charged particles , fig .
[ fig : mfvscentpos ] for positive particles , and fig .
[ fig : mfvscentneg ] for negative particles .
contributions from geometry fluctuations are included in these measurements . while the 62 gev fluctuations are relatively flat as a function of centrality , the 200 gev data exhibit a structure that peaks in mid - central collisions .
the behavior of the fluctuations at both energies are qualitatively different than that observed at the sps , where the scaled variance increases from central to peripheral collisions over this range .
the qualitative features of the data at both of the rhic energies are the same for inclusive charge , positively charged , and negatively charged particle distributions .
charge selection at rhic energies appears only to divide the value of @xmath28 roughly by a factor of two .
it is well established that charged particle multiplicity fluctuation distributions in elementary and heavy ion collisions are well described by negative binomial distributions ( nbd ) @xcite .
the nbd of an integer @xmath29 is defined by @xmath30 where @xmath31 is normalized for @xmath32 , @xmath33 .
the nbd contains an additional parameter , @xmath34 , when compared to a poisson distribution .
the nbd becomes a poisson distribution in the limit @xmath35 - the larger the value of @xmath34 , the more poisson - like is the distribution .
the variance and the mean of the nbd is related to @xmath34 by @xmath36 , where @xmath37 is the mean and @xmath38 is the standard deviation of the distribution .
the phenix multiplicity distributions are well described by nbd fits for all centralities .
the extracted values of the @xmath34 parameter are shown in fig .
[ fig : mfkvscentall]-[fig : mfkvscentneg ] as a function of centrality for inclusive charged , positively charged , and negatively charged particles . except for the most central collisions , the values of @xmath34 for the 62 and 200 gev data are consistent with each other for all charge selections .
there is also little difference in @xmath34 for the different charge selections and the @xmath34 parameter sharply increases in the most central collisions for both collision energies .
some of the qualitative features seen in the data can be studied by examining results from a preliminary hijing simulation within the phenix acceptance .
[ fig : mfvscenthij ] shows the scaled variance according to hijing for the following conditions : a ) jet energy loss activated , b ) jet energy loss deactivated , and c ) jet production deactivated . according to this simulation
, the activation of jet production , which produces events that are correlated in multiplicity , may explain the rise in the fluctuations seen in the 200 gev data when going from central to mid - central collisions .
however , hijing over - predicts the magnitude of the scaled variance . since the jet production cross section is much lower in 62 gev collisions , their contribution should be greatly reduced , thus the absence of significant jet contributions may explain why the fluctuations are roughly independent of centrality . shown in fig .
[ fig : mfkvscenthij ] are the nbd @xmath34 parameter results from the hijing simulation for 200 gev au+au collisions . with jet production excluded ,
the distributions are well - described by poisson distributions - the @xmath34 parameter is very large . adding jet production introduces a non - poissonian component to the distribution and reduces @xmath34 to a value that is close to that of the data .
there is little difference when including or excluding energy loss . in order to better understand the various sources of the multiplicity fluctuations ,
the fluctuations have been measured as a function of @xmath19 range , @xmath21 .
the results in terms of the nbd @xmath34 parameter are shown for three centralities in fig .
[ fig : mfkvsptc0]-[fig : mfkvsptc5 ] .
the qualitative behavior of the fluctuations changes dramatically when comparing central to mid - central collisions .
this behavior may be consistent with what is expected from jet suppression . in the most central collisions
, jet suppression effectively dilutes the hard , non - poissonian contribution to the distribution resulting in an increasing value of @xmath34 . in non - central collisions ,
the dilution of the non - poissonian hard component is reduced and @xmath34 decreases .
this could also explain the rapid increase in @xmath34 as a function of centrality in the most central collisions .
more detailed simulations are being studied to better understand the measured trends and to better estimate contributions from hard processes and geometry fluctuations .
phenix has measured fluctuations of charge , mean transverse momentum , and charged particle multiplicity .
phenix has demonstrated that despite the small acceptance , sensitive and accurate measurements of fluctuation observables can still be made .
the charge fluctuation measurements are consistent with statistically independent particle emission .
the mean transverse momentum fluctuation measurements demonstrate that any residual temperature fluctuations that may be present are very small , and are on the order of the fluctuations seen at sps energies .
the multiplicity fluctuation measurements demonstrate a qualitative behavior as a function of centrality that differs from measurements at sps energies .
there are interesting prospects for future fluctuation measurements from phenix using the new cu+cu datasets taken at @xmath3=19 , 62 , and 200 gev .
9 adcox k _ et al _ ( phenix collaboration ) 2003 _ nucl . instrum . meth . _
a * 499 * 469 .
et al _ ( star collaboration ) 2002 _ phys .
_ c * 66 * 024901 .
jeon s and koch v 2000 _ phys .
lett . _ * 85 * 2076 .
et al _ ( phenix collaboration ) 2002 _ phys .
lett . _ * 89 * 082301 .
nystrand j _ et al _ ( phenix collaboration ) 2003 _ nucl .
phys . _ a * 715 * 603c . bialas a 2002 _ phys .
_ b * 532 * 249 .
stephanov m _
et al _ 1999 _ phys .
_ d * 60 * 114028 .
gavin s 2004 _ phys .
lett . _ * 94 * 162301 .
et al _ ( phenix collaboration ) 2004 _ phys .
* 93 * 092301 .
adams j _ et al _
( star collaboration ) 2003 _ preprint _ nucl - ex/0308033 .
mitchell j 2004 _ j. phys .
_ g * 30 * s819 .
korus r 2001 /it phys .
c * 64 * 054908 .
rybczynski m _
et al _ ( na49 collaboration ) 2004 _ preprint _ nucl - ex/0409009 .
et al _ ( e-802 collaboration ) 1995 _ phys .
_ c * 52 * 2663 . | the phenix experiment at the relativistic heavy ion collider has made measurements of event - by - event fluctuations in the net charge , the mean transverse momentum , and the charged particle multiplicity as a function of collision energy , centrality , and transverse momentum in heavy ion collisions .
the results of these measurements will be reviewed and discussed . . |
nonlinear wave equations ( nlwe ) lie at the heart of many fields in physics including hydrodynamics , classical integrable systems , field theories , etc .
let us give a prototype class of nlwe for a real scalar field , @xmath0 in @xmath1 dimensional space - time : [ nlwe ] - + ^(n ) = f(),^(n)=_i=1^n , where the real function , @xmath2 , defining the theory is given for example as @xmath3 for the `` canonical '' @xmath4 model , and @xmath5 for the sine - gordon ( sg ) model .
a particularly interesting class of solutions of nlwe is the class of nonsingular ones exhibiting spatial localization .
such spatially localized solutions have finite energy and can correspond to static particle - like objects or to various traveling waves . in field theory
localized static solutions have been quite intensively studied in a great number of models in various space - time dimensions , see e.g. the recent book of sutcliffe and manton @xcite .
spatially localized solutions with a non - trivial time dependence ( i.e. not simply in uniform motion ) of nlwe are much harder to find .
in fact the simple qualitative argument stating that _ `` anything that can radiate does radiate '' _ , indicates that the time evolution of well localized cauchy data leads _ generically _ to either a _ stable static _ solution plus radiation fields , or the originally localized fields disperse completely due to the continuous loss of energy through radiation . since derrick - type scaling arguments exclude the existence of localized static solutions in scalar field theories given by eq .
( [ nlwe ] ) in more than two spatial dimensions , for @xmath6 one does not expect to find localized states at all at the end of a time evolution . simplifying somewhat the above
, one could say that in the absence of static localized solutions , localized initial data can not stay forever localized .
in fact for localized initial data there is a time scale which is the crossing time , @xmath7 , ( the time it takes for a wave propagating with characteristic speed to cross the localized region ) and a priori one would expect the rapid dispersion of initially localized cauchy data within a few units of @xmath7 .
one of the rare examples of a time - periodic solution , staying localized forever is the famous `` breather '' in the @xmath8 dimensional sine - gordon ( sg ) model . in the @xmath8 dimensional `` canonical '' ( with a double well potential ) @xmath4 theory the pioneering work , based on perturbation theory by dashen , hasslacher and neveu @xcite indicated the possible existence of breather - like solutions .
a completely independent numerical study by kudryavtsev @xcite has also indicated that suitable initial data evolve into breather - type states .
these results stimulated a number of investigations about the possible existence of non - radiative solutions in the @xmath8 dimensional @xmath4 model .
after a long history segur and kruskal @xcite and vuillermot @xcite have finally established that in spite of the above mentioned perturbative and numerical indications time - periodic spatially localized finite energy solutions ( breathers ) do not exist .
even if genuine localized breathers in the @xmath8 dimensional @xmath4 model are absent , in view of the perturbative and numerical evidence for the existence of long living states `` close '' to genuine breathers , it is a natural question how to describe them .
boyd has made a detailed study of time periodic solutions which are only _ weakly localized _ in space ( i.e. the field @xmath0 possesses a slowly decreasing oscillatory tail ) but as long as the amplitude of the oscillatory `` wings '' are small they still have a well defined core .
boyd has dubbed such solutions `` nanopterons '' ( small wings ) , we refer to his book for a detailed review @xcite .
quite interestingly bogoluvskii and makhankov have found numerical evidence for the existence of spatially localized breather - type states in the spherically symmetric sector of @xmath4 theory in @xmath9 dimensions @xcite .
these breather - like objects observed during the time evolution of some initial data are called nowadays `` oscillons '' .
most of the observed oscillon states are unstable having only a _ finite lifetime_. they lose their energy by radiating it ( slowly ) to infinity .
more recent investigations started by gleiser @xcite have revealed that oscillons do form in a fairly large class of scalar theories in various spatial dimensions via the collapse of field configurations ( initial data ) that interpolate between two vacuum states of a double well potential .
such a spherically symmetric configuration corresponds to a bubble , where the interpolating region is the bubble wall that separates the two vacuum states at some characteristic radius .
these works have led to a renewed interest in the subject . it has been found that oscillons have extremely long lifetimes which is already quite remarkable and makes them of quite some interest .
these long living oscillon states seem to occur in a rather generic way in various field theories involving scalar fields in even higher dimensional space - times @xcite and according to @xcite oscillons are also present in the nonabelian su(2 ) bosonic sector of the standard model of electroweak interactions at least for certain values of the pertinent couplings
. such oscillons might have important effects on the inflationary scenario @xcite as they could form in large numbers retaining a considerable amount of energy .
in a recent study @xcite of a @xmath8 dimensional scalar theory on an expanding background exhibited very long oscillon lifetimes , while in ref .
@xcite , @xcite extremely long living oscillons have been exhibited in a @xmath10 dimensional sine - gordon model .
the sophisticated numerical simulations by honda and choptuik of the cauchy problem for spherically symmetric configurations in the @xmath4 theory in @xmath9 dimensions @xcite , @xcite , have revealed some interesting new features of the oscillons .
in particular in ref . @xcite it has been found that by a suitable fine - tuning of the initial data the lifetime of the oscillons could be increased seemingly indefinitely , and it has been conjectured that actually an infinitely long lived , i.e. non - radiative , spatially localized solution exists .
furthermore the existence of such a solution would even provide the explanation of the `` raison dtre '' and of the observed genericness of long lived oscillons .
the eventual existence of a non - radiative breather in this simple and `` generic '' @xmath4 model in @xmath9 dimensions would be clearly of quite some importance .
it should be noted , that in quite a few spatially discrete models , localized time - periodic `` discrete breathers '' have been shown to exist and they are being intensively studied @xcite .
one of the motivation of our work has been to clarify if non - radiative breathers indeed exist and to find them directly by studying time - periodic solutions of the nlwe .
our numerical results led us to conclude that no localized ( with finite energy ) time - periodic solutions exist in the @xmath4 model . on the other hand
this study has led us to understand the oscillon phenomenon better and we present a simple but quantitatively correct scenario explaining some important properties of the oscillons ( such as their existence and their long lifetimes ) .
our scenario is based on the existence of a special class of time periodic solutions which are weakly localized in space . such solutions ( which have infinite energy )
will be referred to as quasi - breathers ( qb ) . in the present paper
we make a detailed study of oscillons in the ( already much studied ) @xmath4 model in @xmath9 dimensions . using a previously developed and well tested time evolution code where space is compactified , thereby avoiding the problem of artificial boundaries @xcite , we compute some long - time evolutions of gaussian - type initial data .
we observe that long living ( @xmath11 in natural units ) oscillon states are formed from generic initial data .
these oscillons radiate slowly their energy and for short ( as compared to their total lifetime ) timescales they can be characterized by a typical frequency , @xmath12 .
this frequency increases slowly during the lifetime of the oscillon and when @xmath12 reaches a critical value @xmath13 there is a rapid decay . by fine - tuning the initial data one can achieve that the oscillon state instead of rapidly decaying evolves into a near time - periodic state , whose frequency is nearly constant in time and @xmath14 .
the existence of such near - periodic states ( referred to as resonant oscillons ) has been already reported in ref .
@xcite , and we also observe an increase of the lifetime of these states without any apparent limit by fine - tuning the parameters of the initial data to more and more significant digits .
there are , however , also some discrepancies .
we find that clearly distinct near - periodic states for various values of the pulsation frequency @xmath14 exist .
according to our results there is little doubt that for any value of @xmath12 in this range a corresponding near - periodic oscillon state exists .
our data clearly shows that the near - periodic states also radiate , although very weakly .
the radiation becomes weaker and weaker as @xmath15 .
on the other hand we have implemented a multi domain spectral method in order to find directly stationary , time periodic solutions of the nlwe and compare them to the long living oscillon states obtained from the time evolution .
this makes it possible to attack the problem of finding directly the putative time periodic breather(s ) .
( one could easily generalize our method for the quasi - periodic case . )
we find that there is a large family of time periodic solutions which are only _ weakly localized _ in space , in that they have a well defined core , and an oscillatory tail decreasing as @xmath16 .
we single out a special family among them by minimizing the amplitude of their oscillatory tail which definition comes close to minimizing the energy density of the oscillatory tail .
these solutions are the closest to a breather and for that reason we call them quasi - breathers .
they seem to exist for any frequency , @xmath17 , although in this paper we exhibit qbs only with @xmath18 .
the amplitude of the oscillatory tail of the qbs becomes arbitrarily small as the frequency approaches the continuum threshold defined by the mass of the field , @xmath15 .
our numerical evidence speaks clearly against the possible existence of a truly localized , breather - like solution periodic in time , for the frequency range @xmath19 contrary to the claims of ref . @xcite . in view of these conflicting numerical findings
it is now highly desirable to try to find an analytical proof or disproof of the existence of a localized non - radiative solution to settle this issue .
we do not expect the situation being qualitatively different from the @xmath8 dimensional case , and although the proofs of refs .
@xcite are not applicable for the @xmath9 dimensional case , we see no reason that their negative conclusion would be altered .
more importantly , we believe to have made a step towards understanding the mechanism behind the existence of such long - living oscillonic states without the need to invoke genuine breather solutions , which even if they would exist would be clearly non - generic , while the qbs seem to be generic .
the total energy of the quasi - breathers is divergent , due to the lack of sufficient ( exponential ) spatial localization , hence they are not of direct physical relevance .
nevertheless a careful numerical study shows that the _ oscillons _ produced during the time evolution of some suitable cauchy data are quantitatively very well described by the _ quasi - breathers_. by comparing the fourier decomposition of an oscillon state at some instant , @xmath20 characterized by a frequency , @xmath21 , obtained during the time evolution with that of the corresponding qb , we have obtained convincing evidence that the localized part of the oscillon corresponds to the core of the qb of frequency @xmath21 .
what is more , the oscillatory tail of the qb describes very well the standing wave part of the oscillon .
our oscillon scenario is based on this analysis and leads us to propose that any oscillon contains the core and a significant part of the oscillatory tail of the corresponding qb .
the time evolution of an oscillon can be approximatively described as an adiabatic evolution through a sequence of qbs with a slowly changing frequency @xmath21 .
this oscillon scenario is based on the existence and of the genericness of the qbs .
this paper is organized as follows . in sec .
[ s : evol ] we study the time evolution of localized , gaussian - type initial data in @xmath4 theory and investigate some important aspects of the oscillon solutions . in sec . [ s : mode ] we present an infinite system of coupled ordinary differential equations ( ode s ) obtained by the fourier - mode decomposition of the nlwe eq .
( [ e : evol ] ) and discuss some of its properties .
section [ s : numerics ] is devoted to the description of the spectral methods used to solve this system .
in particular , we carefully deal with the asymptotic behavior of the fourier - modes .
various convergence tests are exhibited .
the quasi - breathers are discussed in sec .
[ s : results ] , the results on our oscillon scenario are discussed in sec .
[ s : compar ] and conclusions are drawn in sec .
[ s : conclu ] .
we consider the following @xmath4 theory in @xmath22 dimensions whose action can be written as [ action ] s = dt d^3 x , where @xmath0 is a real scalar field , @xmath23 , @xmath24 and @xmath25 . in this paper
we shall restrict ourselves to spherically symmetric field configurations , when the corresponding nlwe is given by [ e : evol ] -_,tt + = (^2 - 1),=^2_r+_r .
the energy corresponding to the action ( [ action ] ) can be written as e = 4 _
0^ drr^2 , where @xmath26 denotes the energy density [ e : density ] = (_t ) ^2 + (_r ) ^2 + (^2 - 1)^2 .
it is easy to see that the finiteness of the total energy is guaranteed by @xmath27 as @xmath28 .
we briefly outline here the main ideas for the implementation of our evolution code to solve numerically the cauchy problem for eq .
( [ e : evol ] ) . assuming @xmath29 as @xmath28 we introduce the new field , @xmath30 as @xmath31 then the nlwe eq .
( [ e : evol ] ) takes the form : @xmath32 the next step is to compactify in the spacelike directions by a suitable coordinate transformation of @xmath33 .
this way we can guarantee that our computational grid , associated with a finite - difference scheme , covers the entire physical spacetime , at least in principle .
specifically , a new radial coordinate , @xmath34 , is introduced in the following way : @xmath35 where @xmath36 is an arbitrary positive constant . in the new radial coordinate , @xmath34 ,
the entire minkowski spacetime is covered by the coordinate domain @xmath37 while spacelike infinity is represented by the ` hypersurface ' @xmath38 .
the @xmath39 ` lines ' represent world - lines of ` static observers ' , i.e. integral curves of the vector field @xmath40 . in the compactified representation
the field equation , ( [ e : fe2 ] ) , reads as @xmath41 where @xmath42 is given by @xmath43 we remark that the spacelike compactification used here is a simplified variant of the conformal transformation used in @xcite .
there instead of the @xmath44 hypersurfaces the initial data are specified on hyperboloids .
furthermore minkowski spacetime is compactified mapping null infinity to finite coordinate values .
since in the present case the scalar field , @xmath0 , is massive , i.e. never reaches null infinity the hyperboloidal compactification is not essential . in order to obtain a system of first order equations
we introduce the independent variables @xmath45 and @xmath46
. then equation ( [ e : fe3 ] ) can be rewritten as @xmath47 which together with the integrability condition @xmath48 and the defining equation @xmath49 form a strongly hyperbolic system of first order differential equations for the three variables @xmath30 , @xmath50 and @xmath51 ( see e.g. @xcite ) .
the initial value problem for such a first order system is known to be well - posed @xcite .
note that the relation @xmath52 is preserved by the evolution equations , and therefore it corresponds to a constraint equation . in order to solve the initial value problem for equation ( [ e : fe4 ] )
we discretize the independent variables @xmath20 and @xmath34 .
a simple uniform grid with steps @xmath53 and @xmath54 is introduced .
spatial derivatives are calculated by symmetric fourth order stencils .
time integration is done using the ` method of lines ' in a fourth order runge - kutta scheme , following the recipes proposed by gustafsson _
et al _ @xcite .
a dissipative term proportional to the sixth derivative of the field is added in order to stabilize the evolution . since this dissipative term
is also chosen to be proportional to @xmath55 , it does not reduce the order of the applied numerical method , in other words , its influence is decreased by the increase of the used resolution .
the numerical methods and the actual numerical code we use for calculating time evolution in this paper are also based on those developed in @xcite .
a few non - physical grid points are introduced for both negative radii @xmath56 and for the region `` beyond infinity '' @xmath57 .
instead of calculating the time evolution of the @xmath56 points , the symmetry property of @xmath0 about the origin @xmath58 is used to set the function values at each time step .
similarly , @xmath0 , being a massive field , decays exponentially towards infinity , consequently all the field values @xmath30 , @xmath50 and @xmath51 are set to zero for @xmath59 during the entire evolution .
this takes care of the spacelike infinity @xmath60 in equation ( [ e : fe4 ] ) .
therefore it is possible to use symmetric stencils exclusively when calculating spatial derivatives .
the grid point at the origin @xmath58 needs special treatment since the last term as it stands on the right hand side of equation ( [ e : fe4 ] ) is apparently singular .
however , this term , when evaluated in terms of the original ( non - singular ) field value @xmath0 have zero limit value at @xmath58 . although compactifying in spatial direction restricts the coordinate @xmath34 to a finite domain , grid points in our numerical representation get separated by larger and larger physical distances as we approach @xmath38 .
this far region , where shells of outgoing radiation can not be represented properly , moves out to higher and higher physical radii as @xmath54 decreases .
numerical simulations with increasing number of grid points demonstrate that wave packets of outgoing massive fields get absorbed in the transitional region without getting reflected back into the inner domain . in this way
our numerical simulations still give a good description of the field behavior precisely in the central region for very long time periods .
the simple but physically non - uniform grid together with the help of the dissipation term appears to absorb outgoing radiation in a similar way to the explicit adiabatic dumping term method applied by gleiser and sornborger @xcite .
furthermore , because of the very low inward light velocity in the asymptotic region @xmath61 the field behavior in the central region is correctly given for a long time period even after the appearance of numerical errors at @xmath61 .
simulating time evolution of oscillons up to their typical maximal lifetime of @xmath62 using spatial resolution of @xmath63 points took usually a week on personal computers .
however , because of the need of several runs when fine - tuning parameters in the initial data , we mostly used typical resolutions of @xmath64 spatial points .
the parameter @xmath36 in the coordinate transformation ( [ e : itr2 ] ) was set to @xmath65 in our simulations in order to concentrate approximately the same number of grid points to the central oscillon region and to the far away region where the massive fields form high frequency expanding shells .
a courant factor of @xmath66 turned out to be appropriate to obtain stable simulations with our choice of @xmath36 .
the convergence tests confirmed that our code does provide a fourth order representation of the selected evolution equations .
moreover , we monitored the energy conservation and the preservation of the constraint equation @xmath52 .
most importantly , we compared the field values which can be deduced by making use of the green s function and by the adaptation of our particular numerical code to the case of massive linear klein - gordon fields @xcite .
the coincidence between the values in the central region provided by these two independent methods for long time evolutions ( @xmath67 measured in mass units ) made it apparent that the phenomena described below should be considered as true physical properties of the investigated non - linear field configurations .
following refs .
@xcite we start with the following gaussian - type initial data : @xmath68 with @xmath69 and @xmath70 being the field values at the center @xmath71 and at infinity @xmath72 while @xmath73 is the characteristic size of the bubble at which the field values interpolates between @xmath69 and @xmath70 . by fixing @xmath74 as in @xcite but varying @xmath73 and @xmath69 , eq . ( [ e : ff ] )
provides a two - parameter family of smooth and suitably localized initial data . for a large open subset of the possible initial parameters @xmath69 and @xmath73 , after a short transitional period
the field evolves into a long living localized nearly periodic state , named _
oscillon _ by copeland , gleiser and mller .
although these configurations live much longer than the dynamical time scale expected from the linearized version of the problem ( i.e. light crossing time ) , their lifetime is clearly not infinite .
the energy of these oscillating states is slowly but definitely decreasing in time , and after a certain time period they quickly disintegrate . for the time dependence of the energy in a compact region see fig . 3 of @xcite .
we illustrate on fig .
[ f : nontune1 ] two such oscillon states with rather different lifetimes .
the upper and lower envelope of the oscillations of the field @xmath75 is shown for the evolution of gaussian initial data of the type ( [ e : ff ] ) with two different sets of initial parameters .
it is important to note that the field @xmath0 oscillates with a non - constant period @xmath76 . for generic oscillons
the period @xmath77 shows a decreasing tendency towards @xmath78 .
, width=453 ] the parameter dependence of the lifetime is illustrated on figs .
6 , 7 and 8 of @xcite .
we note that since the final decaying period is relatively short , furthermore its time dependence is almost the same for each particular choice of initial data , the lifetime plots are quite insensitive of the precise definition of how one measures the lifetime of a given configuration . in our calculations
the lifetime @xmath79 was defined by observing when the value of the oscillating field at the center @xmath71 falls ( and remains ) below a certain prescribed value ( e.g. @xmath80 ) .
already copeland , gleiser and mller have noticed that a delicate fine structure appears in the lifetime plot .
this peculiar dependence on the precise value of the initial parameter @xmath73 has motivated the detailed investigation of honda and choptuik ( see figs . 4 and 5 of ref .
the calculations have shown that fixing the value @xmath69 the lifetime increases without any apparent upper limit when the parameter @xmath73 approaches some element of a large set of discrete resonance values @xmath81 .
for example , in the case @xmath82 honda and choptuik have found @xmath83 peaks on the lifetime plot between @xmath84 .
these fine - tuned oscillons at a later stage during their evolution develop into a state which is very close to a periodic ( non - radiating ) one .
[ f : csx4 ] we show the field value at the center for typical oscillons close to a chosen peak .
upper envelope of the field value for different oscillon states close to the @xmath85th peak of the lifetime curve belonging to @xmath82 .
these states are characterized by initial value @xmath73 which is near @xmath86 .
the _ near - periodic state _ is approximately at @xmath87 for the fine - tuned oscillons in the lower plot . ,
width=453 ] the seemingly stable almost periodic stages of the fine - tuned oscillon configurations will be referred as _ near - periodic states_. the closer the initial parameters are to the critical value the longer the lifetime of the near - periodic state becomes . although the period @xmath77 of the underlying high frequency oscillations remains constant to a very good approximation during a chosen near - periodic state , different near - periodic states , corresponding to various peaks on the lifetime curve , oscillate with clearly distinct periods in the range @xmath88 .
generic oscillon states and even the not near - periodic initial part of fine - tuned oscillons pulsate with longer period , in the range @xmath76 .
generic oscillons with initial data between two neighboring resonance values @xmath89 may have already quite a long lifetime without developing into a near - periodic state .
plotting the field value at the center @xmath71 one can see a low frequency modulation of the amplitude ( called shape mode by honda and choptuik ) of the high frequency basic oscillating mode ( see upper plot of fig .
[ f : csx4 ] ) .
oscillons between the next two resonance values @xmath90 are distinguished from those in the previous interval by the fact that they possess exactly one more or one less peak associated with these low frequency oscillations on the envelope of the field value @xmath91 before they disperse .
longer living _ supercritical _ states arise when one closely approaches a critical value from one of the possible two directions .
then the last low frequency modulation peak gets shifted out to a later and later time as one goes towards the resonance , making room for a near - periodic state between the last two modulations .
close to the resonance , but on the other side of it , the same long living near - periodic state appears , now called _
, with the only difference that at the end the field disperses without forming a last low frequency modulation peak .
it was shown in ref .
@xcite that close to a resonance the oscillon lifetime @xmath79 obeys a scaling law @xmath92 where the scaling exponent @xmath93 has specific values for each resonance , while the constant @xmath94 is also different for subcritical and supercritical states .
although the lifetime appears to increase without any limit by fine - tuning the initial parameter , in practice it is very difficult to achieve very long lifetimes because one can not represent numbers very close to the resonance value due to the limitation implied by the applied machine precision .
achieving longer lifetime for the near - periodic state is possible by using high precision arithmetics , although then there is a considerable increase of the computation time which limits the applicability of this approach . using `` long double '' variables on an sgi machine , or applying the software ``
doubledouble''@xcite on a personal computer , we could calculate with twice as many significant digits than standard double precision computer variables can represent ( i.e. @xmath95 instead of @xmath96 ) . this way we could improve the fine - tuning and double the observed lifetime of near - periodic states . on fig.[f : lt ] we plot the scaling law of the lifetime for oscillons near three different resonances .
the oscillon lifetime , @xmath79 , versus @xmath97 is shown for states close to three different resonances .
the lower lines plotted with a given line type represent subcritical , @xmath98 , states while the upper lines with the same line type correspond to supercritical , @xmath99 , solutions .
, width=453 ] instead of choosing three resonances with @xmath82 , we calculated states close to the _ first peak _ on the lifetime curves corresponding to three different values of @xmath69 .
the reason for this was that the normal , slowly but evidently decaying , oscillon state is the shortest near the first peak ( i.e. no modulation on the contour curve ) , thereby we could concentrate computational resources on the near - periodic state . in order to clarify what we mean
by fine - tuning the parameter @xmath73 to @xmath96 ( or @xmath95 ) digits and what is the actual error of the quantities , we present a table on the precise location of the first peak for @xmath82 when performing fine - tuning with five different numerical resolutions . for each spatial resolution
we could achieve approximately the same lifetime of approximately @xmath100 when the parameter @xmath73 approximated a resolution dependent value to @xmath96 digits .
the convergence of the data indicates that the actual position of the peak is at @xmath101 with an error of @xmath102 . .
the position @xmath103 of the first peak for @xmath82 using various resolutions .
the number of spatial grid points used for a specific fine - tuning is @xmath104 .
the error is estimated as @xmath105 .
the convergence factor is defined as @xmath106 . [ cols="^,^,^,^,^",options="header " , ] our numerical simulations clearly show that the different resonance peaks on the lifetime plot correspond to different near - periodic states . in fact , a one - parameter family of distinct near - periodic states appears to exist .
this statement is in marked contrast with the claim of honda and choptuik in @xcite , where in sec .
iii - a they claim that the oscillation period is almost the same for all oscillons and is roughly @xmath107 , which corresponds to a pulsation frequency of @xmath108 . from our numerical analysis
we can clearly see that the period of the oscillation depends on the resonance considered . on fig .
[ f : max ] the upper envelope of the oscillating field value at the origin is shown as function of time for oscillon states close to the same three resonances as on fig .
[ f : lt ] .
upper envelopes of the oscillations of the field value at @xmath71 for three pairs of sub and supercritical oscillon states close to the first resonances corresponding to three given @xmath69 initial parameters .
the fine - tuning was performed using @xmath95 digit arithmetics .
, width=453 ] on fig .
[ f : fr ] the time dependence of the frequency at the origin is shown for the same three states during the time period where the oscillation is almost periodic ( i.e. during the near - periodic states ) . the pulsation frequency , @xmath12 in function of the time for the three pairs of oscillons shown on the previous figure during their corresponding near - periodic states . , width=453 ] the frequency @xmath12 of the oscillations along with its time dependence at some radius @xmath109 has been determined from our numerical results by minimizing the following integral for the oscillation period @xmath110 @xmath111 ^ 2dt\ ] ] using some suitably chosen integration interval determined by @xmath112 .
this procedure , applying polynomial interpolation , yields significantly more precise frequency values than the direct use of the fast fourier transform method , especially when the time step by which our evolution code outputs data is not extremely short .
another advantage of our procedure is it being much less sensitive to the particular choice of the sampling interval ( in this case @xmath113 ) than the fft algorithm .
the relatively small value of @xmath112 ( for example @xmath114 ) makes it possible to monitor relatively sharp changes in the time dependence of the frequency .
all three near - periodic states show a low frequency change of @xmath12 with a decaying amplitude , with the lowest frequency modulation belonging to the state with the highest @xmath12 .
the amplitudes on fig .
[ f : max ] would show a similarly decaying slight modulation if we would plot them individually in time intervals where the oscillations are almost time - periodic .
these observations suggest that near - periodic states of frequency @xmath12 also contain a superposition of states with frequencies @xmath115 where @xmath116 . in the next sections we shall see that at least the core part of an oscillon of frequency @xmath12 can be extremely well described by a weakly localized quasi - breather of the same frequency .
we emphasize , that the three chosen near - periodic states are typical , i.e. the near - periodic states of all fine - tuned oscillons ( including all peaks belonging to @xmath82 ) are qualitatively and quantitatively very similar to them . actually , the frequency of all calculated near - periodic states fell into the interval @xmath117 $ ] spanned by the frequency of the two chosen first peaks belonging to @xmath118 and @xmath119 . on the third plot of fig . [
f : fr ] we can also see a slow but steady decrease of the frequency @xmath12 .
for the other two states , with frequencies closer to @xmath120 no such behavior is apparent in the time interval we could simulate .
this slow decrease is also manifested in the energy of the configuration . on fig .
[ f : en ] we plot the time dependence of the energy contained inside spheres with three subsequent radii . the time dependence of the energy contained in spheres of radii @xmath121 , @xmath122 and @xmath123 for the @xmath118 first peak .
the thickness of the curve indicates that the high frequency oscillation is still not negligible at these high radii .
the different endings of the curves correspond to sub and supercritical states .
, width=453 ] the decrease of the energy indicates that the near - periodic state slowly loses its energy , consequently it can not be exactly time - periodic .
looking at the energy in the sphere at @xmath122 , we can see that the decrease of energy from @xmath124 to @xmath125 is @xmath126
. taking into account that the total energy is @xmath127 , we can give a naive estimate on the lifetime by calculating when the energy would decrease into its half value at this rate , getting @xmath128 .
we expect that all near - periodic states radiate , but this radiation is becoming considerably weaker as the frequency get closer to @xmath120 .
this expectation is confirmed by the direct analysis of periodic solutions in the next sections . on fig .
[ f : en ] the curve corresponding to the largest radius ( @xmath123 ) indicates how slowly the energy moves outwards because of the massive character of the field .
the different endings of the curves belonging to sub and supercritical states illustrate the two kinds of decay mechanism of near - periodic states . in the first , subcritical mechanism ,
the field quickly disperses , while energy moves essentially outwards only . in the second supercritical way
, the energy of the field first collapses to a smaller region near the origin and then disperses to infinity .
this resembles to the behavior of some unstable spherical shell , although no shell structure is visible on the density plots , being the highest energy density always at the center .
the instability of the near - periodic states with two distinct kinds of decay mechanisms gives a qualitative explanation on why it is possible to reach long lifetimes by fine - tuning the initial parameters .
since during the time evolution the field @xmath0 becomes approximately time - periodic for any longer living oscillon state , it is natural to look at the fourier decomposition of the results provided by our evolution code .
since the fast fourier transform algorithm is very sensitive to the size of the time step with which our evolution code writes out data , we use an alternative direct method which turns out to be significantly more precise in determining the basis frequency and the amplitude of the higher modes . as a first step
we determine the oscillation period by locating two subsequent maxima , at instants @xmath129 and @xmath130 , of the field at the origin @xmath71 .
since @xmath129 and @xmath130 fall in general between two consequent time slices written out by our evolution code , we approximate their position by fitting second order polynomials on the data .
it is apparent from our results that for near - periodic states these maxima correspond to two consecutive moments of time symmetry not only at the center but also in a large region around the center to a very good approximation . after determining the oscillation frequency @xmath131
, we obtain the @xmath132-th fourier mode at a radius @xmath33 by calculating the following integral using the output @xmath133 of the evolution code @xmath134 we note that care must be taken for the first and last incomplete time steps when evaluating the integral ( [ fint ] ) . in the case of exact periodicity and time symmetry
the imaginary part of this integral would be zero for all @xmath132 . in order to check the deviation from time symmetry for various radii at the moments @xmath129 and @xmath130 we also evaluate the imaginary part of the integral for each @xmath132 and verify whether it is small compared to the real part . as an example , on fig .
[ f : nolog2 ] we give the fourier decomposition of the @xmath119 , @xmath135 near - periodic state presented on fig .
. radial dependence of fourier modes obtained by decomposition of a near - periodic state with frequency @xmath136 .
, width=453 ] the integral ( [ fint ] ) was calculated between two subsequent maximums of the field at the center at @xmath137 and @xmath138 , yielding an oscillation frequency of @xmath136 . for the decomposition of a near - periodic state with a lower frequency
see fig . 9 of @xcite , where the frequency is @xmath139 ( page 107 . of @xcite ) .
from the two plots one can already see the general tendencies . as the basis frequency increases towards the upper limit @xmath120 the oscillon becomes wider , although with a decreasing amplitude .
the influence of the higher modes ( the relative amplitude ratio ) is also getting smaller when the frequency grows .
this section is concerned with a direct search for time - periodic solutions of the nlwe eq .
( [ e : evol ] ) by means of a fourier - mode decomposition of the scalar field @xmath0 .
a simple parameter counting already indicates that an infinite number of parameters would be necessary in order to obtain an exponentially localized breather , so even if some do exist , it seems to be difficult to produce them directly
. therefore we look for weakly localized solutions of eq .
( [ e : evol ] ) , and by analyzing their behavior in function of the free parameters we expect to obtain a clear indication about the existence of truly localized breathers . since we seek periodic solutions of eqs .
( [ e : evol ] ) we make a fourier mode decomposition of the form [ e : period ] (t , r ) = -1 + = -1+_0(r ) + _ n=1^_n(r ) ( n t ) .
we have assumed here that by a suitable choice of the time origin the solutions are time - symmetric . inserting this form in eq .
( [ e : evol ] ) gives rise to a system of coupled elliptic equations [ e : system_bis ] [ -2 ] _ 0 & = & _ 0 ^ 2 (_0 - 3 ) + (_0 - 1 ) _ m=1^_m^2 + _ m , p , q = 1^ _ m _ p _ q _ m , p q + [- _ n^2 ] _ n & = & 3_0 (_0 - 2)_n + (_0 - 1 ) _
m , p = 1^_m _ p _ n , m p + & & + _ m , p , q = 1^ _ m _ p _ q _ n , m p q where @xmath140 .
in fact we have already regrouped together all the linear ( first order ) terms of the corresponding fourier mode , @xmath141 to the left hand site of eqs .
( [ e : system_bis ] ) .
we will make a systematic search for localized solutions of the system ( [ e : system_bis ] ) .
we shall not assume any a priori connection with the oscillons presented in the previous section for the possible frequencies , @xmath12 . as we shall see below the eventual existence of a localized solution
can only be possible if system ( [ e : system_bis ] ) exhibits some very special properties . in order to look for regular solutions of eqs .
( [ e : system_bis ] ) we analyze the various asymptotic behaviors of the homogeneous solutions of the corresponding operator on the left hand side of eqs .
( [ e : system_bis ] ) .
* if @xmath142 then there is one regular homogeneous solution , which falls off exponentially : @xmath143/r$ ] .
this function clearly decays sufficiently fast at infinity so that the energy would always be convergent .
* @xmath144 is a degenerate case when the operator is the usual laplacian .
the non - singular ( decaying ) homogeneous solution is simply @xmath145 which , however , does not decay sufficiently fast for the energy to be bounded . *
if @xmath146 , both homogeneous solutions tend to zero at infinity , @xmath147/r$ ] and @xmath148/r$ ] . as for the previous case
, these functions do not decrease fast enough to make the energy convergent .
apart from that , both functions are regular at infinity and this implies that there is no uniqueness of the solution which is only defined up to a given phase ( for more details , see sec .
[ ss : raccord_inf ] ) .
observe that no matter what the value of @xmath12 is , there always exists an @xmath149 , such that _ all modes _ , @xmath150 will be of the third type ( i.e. with @xmath146 ) .
it is now easy to understand why in the generic case , one can not expect to find exponentially localized ( with finite energy ) solutions of eqs .
( [ e : system_bis ] ) . to ensure localization one has to suppress all slowly decaying oscillatory modes which would normally require an infinite set of freely tunable parameters
. the problem of finding a localized solution can be seen as a matching problem between the set of modes regular in a neighborhood of the origin , @xmath151 and the set of modes with fast ( exponential ) decay for @xmath28 , @xmath152 . in this case
each mode @xmath153 regular at the origin , has a single freely tunable parameter whereas none of the modes with fast fall off , @xmath154 for @xmath155 have any free parameters and therefore a sort of a miracle is needed that the two sets could be matched .
this counting implies that while one can expect to find time - periodic solutions eqs .
( [ e : system_bis ] ) ( even a whole family ) , these solutions have generically oscillatory tails for large values of @xmath33 .
this clearly reflects the argument _ `` anything that can radiate does radiate '' _ transposed to the stationary case . on the other hand , it is not excluded that a localized solution might exist for very particular values of @xmath12 , for which the oscillatory tails are absent .
we solve the system ( [ e : system_bis ] ) using the lorene library @xcite .
the basic features of lorene are the use of spectral methods and multi - domain decomposition .
the problem we are facing here being purely spherical , the fields are expanded on chebyshev polynomials .
the physical space is decomposed in various spherical domains . using such techniques , solving differential equations
can be reduced to , in each domain , inverting a matrix on the coefficients space .
then , a linear combination of the particular solutions with the homogeneous ones is done , in order to impose regularity at the origin , appropriate boundary conditions and continuity of the overall solution .
we refer the reader to @xcite for more details on the algorithm , in the case of a poisson equation . for this work ,
we have extended the operators presented in @xcite and included the helmholtz operators appearing in eqs .
( [ e : system_bis ] ) , i.e. @xmath156 with both signs of @xmath157 .
this is rather straightforward because , as for the laplacian , they produce two homogeneous solutions .
only the case of infinity has to be treated differently as we will see in sec .
[ ss : raccord_inf ] . as we have seen in sec .
[ ss : asymptot ] , for all value of @xmath12 , there exists a value of @xmath132 after which all the modes are dominated by homogeneous solutions of the type @xmath158 .
such functions are not easily dealt with by our solver .
indeed , in order to treat spatial infinity , lorene usually uses a simple compactification by means of the variable @xmath159 in the exterior domain . in the past , this has enabled us to impose exact boundary conditions at infinity .
when waves are present , it is well known @xcite that such a compactification is not practical . indeed ,
no matter what scheme is used , there is always a point after which the distance between computational points ( grid or collocation ) is greater than the characteristic length of change of the wave , thus causing the scheme to fail .
this issue is dealt with by imposing boundary conditions at a finite radius @xmath160 for the slowly decaying oscillatory modes .
the reader should refer to @xcite where such methods have been applied to gravitational waves . for the exponentially decaying modes in eqs .
( [ e : system_bis ] ) we decide to keep in the right hand side ( in the sources ) only those terms that are dominated by the exponentially decaying homogeneous solutions , i.e. , only those modes for which @xmath161 . as it will be seen later
, the range of interesting pulsations is @xmath162 . therefore the only two exponentially decaying modes are @xmath163 and @xmath164 . in the eqs . for @xmath163 and @xmath164 for @xmath165
, we set all the higher modes ( @xmath166 ) to zero ( including terms of the type @xmath167 ) .
so , we effectively solve for large values of the radius the following equations : [ e : sources_ext ] [ -2 ] _ 0&= & (_0 - 1 ) _ 1 ^ 2 + _ 0 ^ 2 (_0 - 3 ) + [ e : sources_ext2 ] [- _ 1 ^ 2 ] _ 1&= & 3 _ 0(_0 - 2)_1+_1 ^ 3 .
this method yields solutions for @xmath163 and @xmath164 which are correct for `` intermediately large '' values of @xmath165 where the oscillatory and slowly decaying terms induced by the nonlinearities do not dominate .
it is clear that for sufficiently large values of @xmath33 @xmath163 and @xmath164 do not decay exponentially , since their behavior will be dominated by the slowly decaying oscillatory nonlinear source terms .
we have carefully checked that changing the value of @xmath160 does not influence the oscillatory modes , therefore we can conclude that the back reaction of the @xmath163 and @xmath164 is negligible on @xmath141 , @xmath166 for @xmath168 . when @xmath142 , the helmholtz operator @xmath156 admits two homogeneous solutions of which only one tends to zero at infinity .
this situation is exactly the same as the one for the standard laplace operator and all the techniques presented in @xcite can be used .
once again , as we will be working for @xmath169 , this happens only for @xmath163 and @xmath164 and the associated sources have been given in the previous section [ ss : sources ] .
the situation is quite different when dealing with the helmholtz operator with @xmath170 .
apart from the compactification problem previously mentioned , we have to note that now there are two homogeneous solutions that are regular at infinity .
there is no reason to prefer one to the other .
this means that there is no unique solution to our problem .
indeed , one can get a solution by doing the matching with any homogeneous solution of the type @xmath171 , where @xmath172 can take any value in @xmath173 .
so to summarize , for all modes such that @xmath170 we have to match , at a finite radius @xmath160 , the solution with a homogeneous one of the type : _
n2 (r > r_lim ) = a_n (|_n| r + _ n)/r .
clearly we need additional conditions to fix the values of the phases @xmath172 . in order to do so ,
let us recall that we are mainly interested in finite energy solutions .
such solutions should not contain any oscillatory behavior in @xmath145 at infinity , i.e. _ all _ the coefficients @xmath174 of such homogeneous solutions should be zero .
one can hope to achieve that by searching , in the parameter space of the phases @xmath175 , the values that minimizes the absolute value of coefficient of the first oscillatory homogeneous solution that appears : @xmath176 .
this solution being `` quite close '' to a localized breather , for that reason we refer to it as a quasi - breather . given the non - linearity of the problem
, the location of the minimum can not be found analytically and one has to rely on a numerical search .
we make use of the multi - dimensional minimizer provided by the _ gsl _ numerical library @xcite .
the algorithm is based on the simplex algorithm of nelder and mead @xcite .
we start by setting all the phases to @xmath177 and iterate the procedure until the value of the minimum converges with a given threshold .
let us mention that the phases are searched in @xmath178 , the rest of the interval , being described by changing the sign of the amplitudes .
the simplex solver converges rapidly , the function @xmath176 being very smooth , with no local extrema .
so the final situation is the following : * for @xmath163 and @xmath164 the space is compactified and the sources in the exterior region are given by eqs . ( [ e : sources_ext ] ) and ( [ e : sources_ext2 ] ) . * for @xmath141 , @xmath179 , we only solve for @xmath180 and match the solution with a homogeneous solution @xmath181 , the @xmath172 being determined by the simplex solver by minimizing @xmath176 .
the system ( [ e : system_bis ] ) is solved by iteration but a problem arises from the fact that the trivial solution @xmath182 is a solution of the equations ( then @xmath163 is a constant whose value is either @xmath183 , @xmath184 or @xmath185 ) .
no matter what the initial guess for the different modes is , the code always converges to a trivial static solution of this type .
so , one needs to find a way to prevent this to happen .
given that the system ( [ e : system_bis ] ) is coupled , it is sufficient to impose @xmath186 , in order to avoid the trivial solution .
to do so , after each step of the iteration , we rescale @xmath164 everywhere , by a factor @xmath187 in order to impose that @xmath188 has a certain value . in the general case , after convergence of the code , this scaling parameter is different from @xmath184 meaning that the value of @xmath164 we found is no longer solution of the system ( [ e : system_bis ] ) .
however for some values of @xmath12 ( being in an interval , see sec .
[ s : results ] ) , it is possible to find , exactly one value of @xmath188 such than the scaling factor is one .
thus , it is possible , at least for some values of @xmath12 to find the appropriate value of @xmath188 such that the obtained modes are non - trivial solutions of eqs .
( [ e : system_bis ] ) . in practice , after convergence to a threshold level of typically @xmath189 we switch on the convergence toward the true value of @xmath188 .
with a technique already use , for example to get neutron stars of appropriate mass ( see sec .
ivd3 of @xcite ) , at each step , we change the value of @xmath164 at the center , in order to make @xmath187 closer to one .
the value of @xmath188 _ to which one wants to converge _ is modified according to : _
1(r=0 ) []^0.1 .
doing so , the value of @xmath164 at the center converges to the only value such that the scaling parameter @xmath187 is one , providing us with the real , non - trivial , solution of the system ( [ e : system_bis ] ) . in order to verify that the simplex solver converges to the proper minimum of @xmath176 , we show , on fig . [
f : conv_prec ] how various quantities vary when one changes the precision at which the extremum is found .
[ f : conv_prec ] presents the values of : * the phase of the second mode @xmath190 at which the minimum is found , * the values of the first modes at the origin @xmath191 , * and the value of the minimum of @xmath176 .
those three quantities are shown as a function of @xmath12 , for three different values of the precision required for the simplex solver , @xmath189 , @xmath192 and @xmath193 .
convergence of the various results with respect to the precision required when finding the minimum of @xmath176 . , title="fig:",height=245 ] convergence of the various results with respect to the precision required when finding the minimum of @xmath176
. , title="fig:",height=245 ] convergence of the various results with respect to the precision required when finding the minimum of @xmath176 . , title="fig:",height=245 ] from the values of the phase @xmath190 , it appears that a precision of @xmath189 or @xmath192 is not good enough , the curve being quite noisy .
the curve obtained with a precision of @xmath193 is smoother and we will assume that this level of precision is sufficient for the purpose of this paper and this value will be chosen for all the rest of this work .
the situation is even better for both the values of the modes at the origin and the actual value of the minimum . indeed , as can be seen on fig .
[ f : conv_prec ] , those quantities show almost not dependence on the level of precision .
this is simply related to the fact that the values of the fields depend very weakly on the values of the phases of the homogeneous solutions in the external region .
this is true for the phase of the second mode @xmath190 but even more for the higher order phases @xmath194 .
the very moderate dependence of @xmath195 on the phases is illustrated by fig .
[ f : low_dep ] where the value of the amplitude is shown , as a function of the phase @xmath190 , for @xmath196 .
all the other phases @xmath194 are fixed to the same value , each of them corresponding to one curve on fig .
[ f : low_dep ] .
it is clear that the influence of the phases of the modes @xmath6 is very weak .
the extrema on the three curves of fig .
[ f : low_dep ] are very close to the real extremum found by the simplex solver .
therefore we are quite confident that we can find the minimum of @xmath176 with a good accuracy , even if the associated values of the phases are slightly less accurate .
value of @xmath195 as a function of the phase @xmath190 , for @xmath197 .
the other phases are maintained equal and set to 1.0 ( solid line ) , 1.6 ( dashed line ) and 2.2 ( dotted line ) .
, height=245 ] in order to further check the validity and accuracy of our code with respect to the computational parameters , we present in this section various additional convergence tests .
the idea is to take a reference set of computational parameters and to change one of those at a time , in order to check that the obtained results do not change much .
the first three radial domains are chosen with boundaries @xmath198 $ ] , @xmath199 $ ] and @xmath200 $ ] .
after @xmath201 , we keep the size of the domain constant to @xmath202 ( i.e. the boundaries of the domain @xmath203 are @xmath204 $ ] ) .
we found that this is necessary to ensure that in every domain , we have enough collocation points to resolve the oscillatory homogeneous solutions .
the various computational parameters are the following : @xmath205 is the number of radial domains which relates directly to the value of @xmath160 given the setting of the domain mentioned above .
@xmath206 is the number of coefficients in each domain .
finally we will also checked the convergence of the results with respect to the finite number of modes we consider in the system ( [ e : system_bis ] ) .
our standard setting consists of @xmath207 , @xmath208 , and the use of 6 modes .
we then change one parameter at a time to verify that the obtained results are indeed meaningful .
the convergence is presented by showing the same three quantities as in sec .
[ s : phase ] : the phase @xmath190 , the values of the first modes at the origin and the amplitude @xmath176 .
the three plots of fig .
[ f : conv_nr ] show the dependence of the results when varying the number of coefficients in every domain . since for the values of @xmath206 the curves of fig .
[ f : conv_nr ] change only slightly , this is a strong sign that @xmath209 is greatly sufficient to get accurate results .
it is also supported by the fact that , after the end of the iterative scheme we find that the relative difference between the right and left hand sides of eqs .
( [ e : system_bis ] ) is smaller than @xmath102 .
same as fig .
[ f : conv_prec ] but when varying the @xmath206 in each domain .
, title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but when varying the @xmath206 in each domain .
, title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but when varying the @xmath206 in each domain . ,
title="fig:",height=245 ] a strong test is provided by the plots of fig .
[ f : conv_rlim ] .
they show that the results are independent of the value of the outer computational radius @xmath160 thus validating the matching procedure with the oscillatory homogeneous solutions .
finally fig .
[ f : conv_modes ] illustrates the fact that the results do not change substantially when increasing the number of modes .
same as fig .
[ f : conv_prec ] but changing the value of @xmath160 .
, title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but changing the value of @xmath160 .
, title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but changing the value of @xmath160 .
, title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but varying the number of modes.,title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but varying the number of modes.,title="fig:",height=245 ] same as fig .
[ f : conv_prec ] but varying the number of modes.,title="fig:",height=245 ] the last test of our code has been a comparison of the r.h.s . and l.h.s .
( [ e : evol ] ) , using the fourier expansion ( [ e : period ] ) .
we have computed the relative difference between the r.h.s . and l.h.s .
after averaging over one period .
the results for various number of modes are depicted on fig .
[ f : error ] as a function of @xmath12 .
as expected , the error decreases as the number of modes increases .
this is not surprising , given that eq .
( [ e : system_bis ] ) could be satisfied only for an infinite number of modes .
the error also increases when @xmath12 decreases .
this can be understood by recalling that the modes are more important when @xmath12 is small ( i.e. see the behavior of @xmath191 as a function of @xmath12 on figs . [
f : conv_nr ] , [ f : conv_rlim ] and [ f : conv_modes ] ) .
thus , the effect of the missing modes is more important for smaller @xmath12 .
relative difference between the l.h.s . and r.h.s .
members of eq .
( [ e : evol ] ) .
the result is averaged over one period and we present the error as a function of @xmath12 , for 5 , 6 and 7 modes .
, height=302 ] finally we would like to emphasize that the computational parameters exhibited in this section are sufficient to compute solutions in the regime of moderate frequencies ( i.e. frequencies around @xmath210 ) .
if one wishes to go to much lower frequencies , one would need to include more modes , for higher order modes will be more important , as indicated by fig .
[ f : error ] . on the other hand , as will be seen in the next section , the matching point @xmath160 must be increased when one approaches the critical value @xmath211 .
as illustrated on figures [ f : conv_nr ] , [ f : conv_rlim ] and [ f : conv_modes ] we do find time - periodic quasi - breather solutions of eqs .
( [ e : system_bis ] ) for any value of the pulsation frequency @xmath212 , and there is little doubt that such solutions exist for all frequencies in the range @xmath2130 , \sqrt{2}[$ ] .
it is also clear from the very same figures that for @xmath214 the family of solutions we consider converges pointwise to the trivial solution .
we do not expect such quasi - breather solutions of the system ( [ e : system_bis ] ) to exist for @xmath215 .
the critical value @xmath211 is expected to be related to the change of nature of the helmholtz operator for @xmath164 , @xmath120 being the value at which @xmath164 ceases to decay like an exponential . in the @xmath8 dimensional case coron @xcite
has proved that the allowed frequencies are indeed constrained by @xmath216 .
we have not attempted to prove the analogous statement for our case , as it would certainly require some sophisticated mathematical tools .
the quasi - breather solutions obtained this way are not well localized in space because of their slowly decaying @xmath16 oscillatory tail .
consequently none of these solutions has finite energy .
in fact we have selected a special class of time periodic solutions by minimizing the amplitude of their oscillatory tail .
this is , in some sense , the closest one can get to a breather and we have called those configurations `` quasi - breathers '' .
there is no special value of @xmath12 for which the amplitude of the oscillatory tail would show any tendency to become very small or going to zero .
this is illustrated on fig .
[ f : coefs ] where the logarithm of the coefficients of `` tails '' are shown for the modes @xmath217 , @xmath218 , @xmath219 and @xmath220 , as a function of @xmath12 .
the only value of @xmath12 for which the curves go to zero is the critical one for which the solution tends to the trivial one .
the many convergence tests of sec .
[ ss : convergence ] show that we can accurately compute the value of those coefficient and that they are not numerical artifacts .
coefficients of the oscillatory homogeneous solutions @xmath172 ( @xmath221 ) as a function of @xmath12 .
, height=302 ] to illustrate the behavior of the quasi - breathers we show on fig .
[ f : dens ] and [ f : lin ] the energy density @xmath26 and the one including the volume element , i.e. @xmath222 as a function of the radius , for values of @xmath12 going from @xmath223 to @xmath224 .
we can clearly see two qualitatively different behaviors : i ) the solutions have a well defined `` core '' , where the behavior is dominated by the exponential decay of the fields and so the density goes to zero
. this core is getting larger when @xmath15 .
ii ) however , inevitably at some point , the oscillatory tails start to dominate and the density reaches a plateau , ultimately causing the total ( integrated ) energy to be infinite . let us mention that if for high values of @xmath12 , the plateau is not seen , it comes solely from the fact that the value of @xmath225 used on fig .
[ f : dens ] and [ f : lin ] is not sufficient . in spite of the numerical smallness of the amplitude of this plateau , we are quite confident that its value is reasonably accurate , and also that it really corresponds to a physical effect .
indeed , its value is very stable when varying the various computational parameters
. energy density for various values of @xmath12 , near the origin ( left panel ) and in the region where the oscillatory tails dominate ( right panel ) .
@xmath160 is equal to 44 . , title="fig:",height=245 ] energy density for various values of @xmath12 , near the origin ( left panel ) and in the region where the oscillatory tails dominate ( right panel ) .
@xmath160 is equal to 44 .
, title="fig:",height=245 ] same as fig .
[ f : dens ] but with the volume element included : @xmath222 . ,
title="fig:",height=245 ] same as fig .
[ f : dens ] but with the volume element included : @xmath222 . ,
title="fig:",height=245 ] as @xmath12 increases , the value of the plateau diminishes .
this is related to the fact that the coefficients of the homogeneous solutions are decreasing functions of @xmath12 and to the quadratic nature of the energy density in @xmath0 , i.e. the value of the plateau is roughly the square of the greatest coefficient ( i.e. the one for @xmath217 ) .
as already stated , we can also confirm that the transition radius between the core and the plateau increases as one approaches @xmath226 . to be more quantitative ,
we define the transition radius @xmath227 as the first value of @xmath33 for which @xmath164 reaches the amplitude of the oscillations : _
1(r = r_trans ) = .
it is the radius at which the oscillatory behavior starts to dominate .
@xmath227 as a function of @xmath12 is shown on the left panel of fig .
[ f : trans ] .
the precise location of the transition radius can not be computed for @xmath228 , given our nominal choice for @xmath229 .
once again , this illustrates the fact that when going to higher values of @xmath12 , one needs to increase the boundary radius @xmath230 . on the right panel of fig . [
f : trans ] we show the total energy inside the transition radius , i.e. the total energy of the core of the configuration .
contrary to the transition radius , this is not a monotonic function and there is a configuration of minimum energy .
this non trivial behavior is due to two competing effects when increasing @xmath12 , first the increase of the transition radius , and second the decrease of the magnitude of the modes .
we also note that this minimum is very close to the value of the frequency which honda and choptuik @xcite claim to be the one of their conjectured breather solution ( i.e. they quoted @xmath231 ) but this may very well be just a coincidence .
it seems likely that the energy contained in the core diverges when @xmath232 . the transition radius between the exponential decay and the oscillatory tail ( left panel ) and the energy inside this radius ( right panel ) , as a function of @xmath12 .
, title="fig:",height=245 ] the transition radius between the exponential decay and the oscillatory tail ( left panel ) and the energy inside this radius ( right panel ) , as a function of @xmath12 .
, title="fig:",height=245 ] finally , on fig .
[ f : oscillons ] we show the first modes near the origin ( left panels ) and in the region of the transition radius ( right panels ) .
three different values of @xmath12 are shown , corresponding to a low value ( @xmath233 ) , the value corresponding to the minimum of energy ( @xmath234 , cf fig .
[ f : trans ] ) and one rather high frequency ( @xmath235 ) .
the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ] the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ] + the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ] the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ] + the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ] the first modes near the origin and in the transition region for @xmath236 , @xmath234 and @xmath235.,title="fig:",height=245 ]
in this section we present our oscillon scenario , based on the existence of the quasi - breathers .
in particular we present some convincing evidence , that an oscillon state of main frequency @xmath12 is _ quantitatively described _ by a quasi - breather of the same frequency , whose tail is cut off at some value of @xmath33 . in order to do this we perform fourier decomposition of various a long - lived oscillon solutions found by doing the evolution ( see subsection [ s : fourier ] for the actual method ) .
this gives us the fundamental pulsation frequency @xmath12 and the various modes .
then , those modes are compared to the ones found by directly solving the system ( [ e : system_bis ] ) for this particular value of @xmath12 .
the same kind of comparison is done in ref . @xcite as to support the existence of a periodic solution of frequency @xmath237 .
an important difference is that in ref .
@xcite the comparison is performed only for higher frequency near - periodic states in an interval of @xmath238 $ ] which is comparable to the core part of the quasi - breather and therefore its oscillatory tail is not yet apparent .
it is apparent from the results provided by our evolution code that the solutions corresponding to the time evolution data are very close to being time symmetric not only at the center but also in the intermediately far region , justifying the choice of only cosine terms in the mode decomposition ( [ e : period ] ) when constructing qbs .
first we perform the fourier decomposition of various near - periodic states obtained by fine - tuning the initial parameter @xmath73 , since these states appear to be periodic and time symmetric to a very high degree . for a specific case figures [ f : comp_139a ] and [ f : comp_139b ]
show the real part of the modes coming from the fourier transform of the evolution and the modes obtained by the direct solution of ( [ e : system_bis ] ) .
comparison between the modes coming from solving the system ( [ e : system_bis ] ) directly and the real part of the modes obtained by performing a fourier transform of the evolution for @xmath239 . in order to be able to represent on the same graph the large values in the central oscillon region along with the small tails at high radii
the absolute values are plotted logarithmically .
we note that the central value @xmath240 is positive for @xmath241 and for odd @xmath132 while it is negative for other @xmath132 .
the peaks pointing downwards on the figure correspond to places where the functions change signature .
, width=453 ] comparison of the higher modes of the two systems presented on the previous figure .
, width=453 ] the comparison is done with a near - periodic state corresponding to the first peak on the lifetime curve with initial data @xmath82 .
the fourier decomposition is performed between the two maxima at @xmath242 and @xmath243 .
the frequency calculated from the position of the maxima is @xmath244 , which is used to calculate the corresponding periodic solution by solving ( [ e : system_bis ] ) .
the agreement for the first oscillating mode @xmath245 is remarkably good even in the oscillating tail region .
the curves obtained by the two different decompositions are indistinguishable in the central region . in order to show the excellence of the agreement the radial dependence of the relative difference of the corresponding modes obtained by the two methods
is shown on fig .
[ f : compd ] .
dependence on the radius @xmath33 of the relative difference of the modes coming from the mode decomposition and the fourier transformation for @xmath244 .
the relative difference is defined as the absolute value of the ratio of the difference and the mode decomposition value .
, width=453 ] we have obtained similarly good agreement not only for @xmath244 but for all frequencies for which near - periodic states exist ( i.e. for @xmath246 ) .
this comparison between the results coming from the evolution code of sec .
[ s : evol ] and from the fourier - mode decomposition of the quasi - breathers is a very strong consistency check .
let us recall that the two codes have been developed completely independently .
figure [ f : zoom ] shows , for large radii , the first mode for which the oscillatory term appears , i.e. @xmath217 , in the case @xmath139 .
the first two graphs compare the oscillatory region of the mode @xmath217 coming from solving the system ( [ e : system_bis ] ) directly and the complex @xmath217 obtained by performing fourier transform on the result of the evolution code , for @xmath139 .
the first graph shows the real and imaginary parts separately , while the second compare the absolute values .
the third graph plots the radial dependence of the complex argument of the modes @xmath164 and @xmath217 obtained by the fourier transform .
, width=453 ] it can be seen from the first of the three plots that the real part of the fourier transform agrees very well with the decomposition of the corresponding qb , the two curves can hardly be distinguished .
however , just at those radii where the oscillatory behavior appears the imaginary part starts being comparable to the real part , breaking the time symmetry in this outer region .
the second plot shows that the absolute value of the complex @xmath217 obtained by the fourier decomposition essentially behaves like a smooth envelope curve covering the standing wave like peaks of @xmath247 obtained by the mode decomposition .
the presence of the complex argument of the fourier transform shown on the third plot indicates that the oscillating tail obtained by the evolution code is composed of outgoing waves carrying out energy from the core region .
oscillons obtained by the evolution of generic gaussian initial data are relatively long living for a large set of the possible values of the initial parameter @xmath73 . on fig .
[ f : nontune2 ] we exhibit the time evolution of the frequency of the two typical non fine - tuned states presented on fig . [ f : nontune1 ] .
time evolution of the underlying high frequency oscillation of the oscillon state of fig .
[ f : nontune1 ] is presented .
, width=453 ] although these states are clearly non - periodic , their lifetime is still very large compared to the period of the basic high frequency oscillation mode .
we have performed the fourier decomposition of these states at various moments of time and compared the results to the mode decomposition of the corresponding qb . on figs .
[ f : ntmode30a ] , [ f : ntmode30b ] , [ f : ntmode36a ] and [ f : ntmode36b ] the fourier decomposition of the two oscillon states shown on figs .
[ f : nontune1 ] and [ f : nontune2 ] is given for moments of time where the frequency is approximately @xmath248 and @xmath223 .
for comparison the corresponding modes calculated by solving ( [ e : system_bis ] ) using the given frequencies is also shown on the figures .
the modes of the periodic solution with frequency @xmath249 is compared with the fourier decomposition of two different oscillon states .
the first oscillon state with initial data @xmath82 , @xmath250 is decomposed between the two subsequent maxima after @xmath251 , where the calculated frequency is @xmath252 .
the second state with @xmath82 , @xmath253 is decomposed just after @xmath254 , where the frequency is @xmath255 .
, width=453 ] comparison of the higher modes of the two systems with frequency @xmath256 presented on the previous figure .
, width=453 ] the modes of the periodic solution with frequency @xmath223 is compared with the fourier decomposition of two different oscillon states .
the first oscillon state with initial data @xmath82 , @xmath250 is decomposed between the two subsequent maxima after @xmath257 , where the calculated frequency is @xmath258 .
the second state with @xmath82 , @xmath253 is decomposed just after @xmath259 , where the frequency is @xmath260 . ,
width=453 ] comparison of the higher modes of the two systems with frequency @xmath261 presented on the previous figure .
, width=453 ] although the agreement of the modes is not as excellent as for the near - periodic states , the curves calculated by the two methods are still remarkably close .
especially important is the similarity at the tail section of the first oscillating mode , i.e. mode @xmath185 .
this indicates that the periodic quasi - breather solutions can describe , at least qualitatively , even the generic long living oscillons .
in particular , the existence of the oscillating tail region is responsible for the slow but steady energy loss of these configurations .
we have also tested our scenario in the reverse way .
first , for a given @xmath12 ( 1.38 in the first example ) , we have solved the system ( [ e : system_bis ] ) for the modes of the quasi - breather .
in particular , we can compute @xmath262 at the moment of time symmetry , and use this as initial data for the evolution code .
the evolution of such initial data is shown on fig .
[ f : evolve ] for numerical simulation with various spatial resolutions .
on the first graph the upper and lower envelopes of the oscillations of the field at the origin are shown as functions of time for numerical simulations with various number of spatial grid points .
the initial data is generated by solving the system ( [ e : system_bis ] ) for the modes at the pulsation frequency @xmath263 .
on the second graph the time dependence of the oscillation frequency is plotted for each simulation . , width=453 ] the field ( here its value at the origin ) oscillates at the appropriate frequency for a relatively long period ( approx .
t=300 , with about 66 oscillations ) before it collapses in a sub or supercritical way very similar to the collapse of the @xmath118 states shown on fig .
[ f : max ] .
the initial state is so close to the ideal configuration that it even depends on the chosen numerical resolution whether in the final stage the configuration collapses in a subcritical or supercritical way .
strangely , the longest living state is not the one with the highest resolution , which is probably connected to the fact that the initial data still contain small numerical errors .
although the lifetimes attainable by this method are very long compared to the dynamical time scale of the linearized problem , they are still shorter than the @xmath264 achievable by fine - tuning the first peak of the @xmath118 initial data
. actually , it should not be surprising that the numerically determined initial data corresponding to a qb can not be as long - living as a configuration which is fine - tuned to 32 decimal digits . since the modes are matched to oscillating tails at large radii ,
the initial data provided by the mode decomposition of a qb is valid up to arbitrarily large @xmath33 .
however , because the slow falloff of the oscillating tails , the energy contained in balls of radius @xmath33 diverges as @xmath33 goes to infinity .
using such configuration as initial data for our evolution code does not cause serious problems , since the physical distance between grid points in our numerical representation increases with the distance from the center , and consequently the high frequency tails can not be represented numerically above a certain radius .
this provides a cutoff in the initial configuration and an effective outer boundary during the evolution .
an advantage of our conformal compactification method is that this outer boundary moves to higher and higher radii when increasing the numerical resolution . on fig .
[ f : evolve2 ] we show the initial part of the upper envelope curve for numerical runs with various spatial resolutions
. initial behavior of the maximum curve of the evolution of the @xmath263 initial data with numerical resolutions corresponding to various number of spatial grid points .
the points on the curves represent the actual maximum points of the oscillating field @xmath0 at the origin @xmath71 .
width=453 ] we emphasize that this strong resolution dependence is entirely due to the inadequate representation of the infinite energy initial data . when we used initial data with compact support or sufficiently fast falloff , such as the gaussian initial data in ( [ e : ff ] )
, then the code remained convergent up to much larger time periods ( at the order of @xmath265 ) , and the curves with various resolutions agreed to very high precision at the initial stage ( at around @xmath266 ) .
an important point is that , with this method , we can generate almost periodic oscillons at frequencies which are difficult or impossible to attain by fine - tuning initial data of the form ( [ e : ff ] ) . on fig .
[ f : evolve3 ] we show the upper envelope of the oscillations at @xmath71 in the initial stage of the evolutions when using initial data provided by the mode decomposition method with frequencies @xmath267 and @xmath261 .
initial behavior of the upper envelope of the oscillations at @xmath71 of the evolution of the @xmath267 and @xmath261 initial data for numerical simulations with various number of spatial grid points .
, width=453 ] again , as the numerical resolution is increased , the size of the oscillating tail that can be taken into account in the numerical representation of the initial data gets larger , and the evolution will remain nearly periodic for longer times .
the field can oscillate truly periodically only if the energy lost by radiation through the dynamically oscillating tail is balanced out by an incoming radiation already present at the tail section of the initial data . if the tail is cut off above some radius @xmath33
then the non - periodic region moves inwards from that radius essentially at the speed of light .
getting into the non - periodic domain , for the frequencies @xmath268 the tail amplitude appears to be large enough to cause a small but significant energy loss , a decrease in amplitude , and consequently a steady frequency increase . on figs .
[ f : maxev ] and [ f : frev ] we show the long - time behavior of the amplitude and frequency of the oscillations generated by initial data with frequencies @xmath267 and @xmath261 . time evolution of the upper envelope curve of the oscillation of @xmath0 at @xmath71 generated by three different initial data . the first two initial data is provided with the mode decomposition method , solving the system ( [ e : system_bis ] ) by choosing the frequencies @xmath267 and @xmath261 .
the third evolution corresponds to gaussian - type initial data ( [ e : ff ] ) with @xmath82 and @xmath250 . , height=302 ] time evolution of the frequency of the oscillation of @xmath0 at @xmath71 for the three kind of evolution simulations shown on the previous figure .
, width=453 ] apart from a relatively short stable initial stage , the main characteristic of the evolution is a slow but steady decrease of the amplitude accompanied by a simultaneous increase of the frequency up to a point ( approx .
@xmath269 ) where the configuration quickly decays .
this evolution is very similar to the behavior of generic ( i.e. not fine - tuned ) oscillons started from gaussian initial data of type ( [ e : ff ] ) .
a typical example of such evolution , with initial data corresponding to an @xmath73 close to the top of the lifetime curve ( but between two peaks ) with @xmath270 is also shown on figs .
[ f : maxev ] and [ f : frev ] in order to facilitate comparison .
it is also rather remarkable that the onset of the rapid decay seems to coincide with the configuration which minimizes the energy inside the core ( see fig .
[ f : trans ] ) .
the longtime evolution of a low frequency initial data provided by the mode decomposition method is very similar to the evolution of generic oscillons evolving from gaussian initial data .
an important difference though , is the smaller low frequency modulation of the envelope curve and of the time dependence of the frequency . from this
one can expect that the fourier decomposition of these states is even closer to that of the quasi - breather state with the corresponding frequency . on figures [ f : stable1 ] and [ f : stable2 ]
we give the fourier decomposition of the evolution of the @xmath249 initial data between the two maxima after @xmath271 , where the calculated frequency is @xmath272 .
fourier decomposition of the evolution of the @xmath256 initial data close to the moment of time when its frequency increases to @xmath223 .
the results are compared to the modes of the periodic solution obtained by the mode decomposition method for the @xmath261 frequency . , width=453 ] same thing as fig .
[ f : stable1 ] but for the higher order modes . ,
width=453 ] the agreement is markedly better than on figs .
[ f : ntmode36a ] and [ f : ntmode36b ] , which is most likely due to the much lower amplitude of the low frequency modulations on the envelopes of field value at the center .
these low frequency modes are excited to a higher amplitude by gaussian type of initial data , while they appear to be present to a much lower extent in the lower frequency periodic initial data ( with @xmath256 in this case ) .
we remind the reader that the existence of the peaks on the lifetime curve ( see fig . 4 of @xcite )
is closely related to these low frequency oscillations .
namely , the peaks separate domains of the initial data in @xmath73 with different number of low frequency oscillations on the envelope curve .
in this paper , we have adapted and applied spectral methods implemented in the lorene library to find time - periodic solutions of the spherically symmetric wave equation of @xmath4 theory .
our code passed numerous tests and we are quite confident that it is sufficiently precise . with our code
we find that for frequencies @xmath17 there is a whole family of standing wave - type solutions with a regular origin having a well localized core and an oscillatory tail . because of the slow decrease of the oscillatory tail the total energy of these solutions in a ball of sufficiently large radius , @xmath34 , is proportional to @xmath34 cf .
( 15,16 ) .
we have constructed a special class of solutions , called quasi - breathers , defined by minimizing the amplitude of the oscillatory tail of the solutions .
the size of the core of the qbs gets larger and larger as @xmath15 and the amplitude of the oscillatory tail is getting smaller and smaller . interestingly the energy contained in the core of the qbs exhibits a minimum for @xmath273 . using our high precision time evolution code
, we have also investigated the time evolution of gaussian initial data .
we observe that a generic oscillon state can be characterized by a slowly varying frequency , @xmath21 , increasing in time up to a critical one @xmath274 , when it decays rapidly .
we have also investigated in detail the near - periodic states described first in ref .
@xcite which can be characterized by an almost constant frequency . in contradistinction to ref .
@xcite we find such near - periodic states for any frequency @xmath275 .
moreover we find that the near - periodic states decay in time ( i.e. they can not be truly time periodic states ) cf .
( [ f : fr],[f : en ] ) .
in particular while loosing some of their energy their frequency decreases very slowly with time . by a careful comparison of the fourier modes of an oscillon state of frequency @xmath21 with that of the corresponding qb ,
we have obtained convincing evidence that the localized part of the oscillon is nothing but the core of the qb of the same frequency .
our results demonstrate that the time evolution of an oscillon state can be described to a good approximation as an adiabatic evolution through a sequence of qbs with a slowly changing frequency @xmath21 .
what is more , the oscillatory tail of the qb describes very well the standing wave part of the oscillon .
therefore any oscillon contains the core and a significant part of the oscillatory tail of the corresponding qb .
the existence of near - periodic states is closely related to the fact that the energy of the oscillon core exhibits a minimum for a critical frequency @xmath276 and for @xmath275 a single unstable mode appears which can be suppressed by fine - tuning the initial data .
this research has been supported by otka grants no .
t034337 , ts044665 , k61636 and by the pole numerique meudon - tours . g. f and i. r would like to thank the bolyai foundation for financial support .
y. hida and x. s. li and d. h. bailey , quad - double arithmetic : algorithms , implementation , and application , technical report lbnl-46996 , lawrence berkeley national laboratory ( 2000 ) , see and download at ` http://crd.lbl.gov/~dhbailey/mpdist/index.html ` | strong numerical evidence is presented for the existence of a continuous family of time - periodic solutions with `` weak '' spatial localization of the spherically symmetric non - linear klein - gordon equation in 3 + 1 dimensions .
these solutions are `` weakly '' localized in space in that they have slowly decaying oscillatory tails and can be interpreted as localized standing waves ( quasi - breathers ) . by a detailed analysis of long - lived metastable states ( oscillons ) formed during the time evolution
it is demonstrated that the oscillon states can be quantitatively described by the weakly localized quasi - breathers .
it is found that the quasi - breathers and their oscillon counterparts exist for a whole continuum of frequencies . |
the h1 and zeus experiments at hera ( desy , hamburg ) allowed the study of electron - proton and positron - proton collisions at center of mass energies of up to 920 gev . during the so called hera
i running phase ( 1994 - 2000 ) about 100 @xmath2 of data were collected per experiment , mainly coming from @xmath3 collisions .
after the collider upgrade in 2000 - 2001 , resulting in a significant increase of luminosity , about 400 @xmath2 of data per experiment were collected in the so called hera ii phase ( 2002 - 2007 ) .
moreover , spin rotators installed at the h1 and zeus interaction regions provided longitudinal electron and positron polarization . with average lepton beam polarization of about 30 - 40% and significant increase of integrated data luminosity ( especially of collected @xmath4 sample ) hera ii
has significantly increased the sensitivity of the experiments to physics beyond the sm .
new interactions between electrons and quarks involving mass scales above the center - of - mass energy can modify the deep inelastic @xmath5 scattering cross sections at high @xmath0 via virtual effects , resulting in observable deviations from the standard model predictions .
four - fermion contact interactions are an effective theory , which allows us to describe such effects in the most general way .
vector @xmath1 contact interactions considered at hera can be represented as an additional terms in the standard model lagrangian : @xmath6 where the sum runs over electron and quark helicities , and a set of couplings @xmath7 describe the helicity and flavor structure of contact interactions .
various scenarios , with different chiral structures , were considered by h1 and zeus .
limits on the model parameters were derived from the analysis of neutral current deep inelastic scattering ( nc dis ) events .
details of the analyses are described in @xcite .
in the general case ( also referred to as compositeness models ) , limits on the effective `` new physics '' mass scale @xmath8 ( compositeness scale ) are extracted assuming the relation @xmath9 .
figure [ fig : ci ] shows the results obtained by the zeus for different compositeness models , based on the analysis of 1994 - 2006 data . limits on the effective mass scale @xmath8 range form 2 up to 8 tev .
corresponding limits obtained by the h1 collaboration , based on the hera i data only , range from 1.6 to 5.5 tev @xcite . and @xmath10 data from zeus ( 1994 - 2006 ) .
horizontal bars indicate the 95% cl limits on @xmath11 ; values outside these regions are excluded .
@xmath12 are the 95% cl limits on the compositeness scale for @xmath13 . ] for the model with large extra dimensions @xcite both collaboration set limits on the effective planck mass scale @xmath14 . for negative coupling sign scales below 0.90 tev ( zeus 1994 - 2006 ) and 0.78 tev ( h1 1994 - 2000 @xcite ) are excluded on 95% cl . for positive couplings the limits are 0.88 tev and 0.82 tev respectively .
possible effects of graviton exchange on the @xmath15 distribution of nc dis events , as measured by zeus , are shown in figure [ fig : led ] . ) and negative ( @xmath16 ) couplings .
results of the experiments are normalized to the standard model expectations using cteq5d parton distributions . ] and @xmath17 leptoquarks @xcite . the indirect limits from zeus and l3 and the direct d0 limits are shown for comparison . ,
title="fig : " ] and @xmath17 leptoquarks @xcite .
the indirect limits from zeus and l3 and the direct d0 limits are shown for comparison .
, title="fig : " ] also , searches for possible quark substructure can be made by measuring the spatial distribution of the quark charge . by using the `` classical '' form factor approximation , and assuming that both electron and exchanged bosons are point - like , limits on the mean - square radius of the electroweak charge of the quark can be set . from the analysis of combined hera
i and hera ii data quark radii bigger than @xmath18 cm ( h1 ) and @xmath19 cm ( zeus ) have been excluded at 95% cl .
figure [ fig : rq ] shows the h1 data together with 95% c.l .
exclusion limits for the effective radius of the quark @xcite .
contact interactions can also be used to describe the effects of virtual leptoquark production or exchange at hera , in the limit of large leptoquark mass @xmath20 .
the zeus collaboration have used data taken between 1994 and 2006 to constrain the leptoquark yukawa coupling for different leptoquark types and masses .
limits on the ratio between mass and yukawa coupling range from 0.29 to 2.08 tev . in figure
[ fig : lq ] indirect limits from zeus are compared to the h1 exclusion limits at 95% c.l . on the yukawa coupling of @xmath21 and @xmath17 leptoquarks , as a function of their mass @xcite .
limits from lep and the tevatron are also indicated .
lepton beam polarization and high luminosity delivered at hera ii opened a new window for precise ew studies and searches for physics beyond sm .
measured nc dis cross sections at high @xmath0 are in very good agreement with sm , so only limits on deviations from sm could be set within different models .
hera running has finished , but analyses of large samples of data are still ongoing and more interesting results can be expected . | the h1 and zeus collaborations at hera have searched for signatures of physics beyond the standard model with high @xmath0 neutral current deep inelastic electron - proton and positron - proton scattering events .
no significant deviations from standard model predictions were observed .
various @xmath1 contact interaction models have been considered .
limits on the compositeness scale in general @xmath1 contact interaction models , mass to the yukawa coupling ratio for heavy leptoquarks , the effective planck mass scale in models with large extra dimensions and the effective quark charge radius are presented . |
the symmetry principle is perhaps the most important ingredient in the development of high energy physics . roughly speaking , the symmetries of the physical system lead to conservation laws , which give many important relations among physical processes .
many of the symmetries in nature are however approximate symmetries rather than exact symmetries and are also very useful in the understanding of various phenomena in high energy physics . among the broken approximate symmetries ,
the most interesting one is the spontaneous symmetry breaking ( ssb ) which seems to have played a special role in high energy physics .
many important progress has come from the understanding of the ssb .
the ssb is characterized by the fact that symmetry breaking shows up in the ground state rather than in the basic interaction .
this makes it difficult to uncover this kind of approximate symmetries .
historically , ssb was first discovered around 1960 in the study of superconductivity in the solid state physics by nambu and goldstone @xcite .
one of consequences of ssb is the presence of the massless excitation @xcite , called the nambu - goldstone boson , or just goldstone boson for short .
later , nambu @xcite applied the idea to the particle physics . in combination with @xmath1 current algebra
, ssb has been quite successful in the understanding of the chiral symmetry in the low energy phenomenology of strong interaction .
more importantly , in 1964 it was discovered by higgs @xcite and others @xcite that in the context of gauge theory , ssb has the remarkable property that it can convert the long range force in the gauge theory into a short range force .
thus it avoids both the massless goldstone bosons and the massless gauge bosons .
weinberg , @xcite and salam @xcite , then applied this ideas to construct a model of electromagnetic and weak interactions .
the significance of this model was not realized until thooft @xcite show in 1971 that it was renomalizable . since then
this model has enjoyed remarkable experimental success and now called the `` standard model of electroweak interactions '' @xcite .
undoubtedly , this will serve as benchmark for any new physics for years to come .
in this article i will give a simple introduction to the spontaneous symmetry breaking and its application to chiral symmetries in the hadronic interaction .
the emphasis is on the qualitative understanding rather than completeness and mathematical rigor .
eventhough ssb has been quite successful in explaining many interesting phenomena , its implementation in the theoretical framework is more or less put in by hand and it is not at all clear what is the origin of ssb . here
i will also discuss some non - relativistic example where the physics is more tractable in the hope that they might give some hints about the true nature of ssb .
maybe good understanding of ssb might extend its applicability to some new frontier .
the @xmath0-model has a long and interesting history .
it was originally constructed in 1960 s as a tool to study the chiral symmetry in the system with pions and nucleons@xcite .
later the spontaneous symmetry breaking and pcac ( partially conserved axial current ) were incorporated ..
eventhough this model is not quite phenomenologically correct it remains the simplest example which realizes many important aspects of broken symmetries .
even though the strong interaction is now described by @xmath3 the @xmath0-model of pions and nucleons is still useful as an effective interaction in the low energies where it is difficult to calculate directly from @xmath4 in addition , the @xmath0-model has also been used quite often as a framework to test many interesting ideas in field theory and string theory . here
we will discuss the most basic features of the @xmath0-model .
the lagrangian for @xmath5 @xmath0-model is given by @xmath6 + \frac{\mu ^{2}}{2}\left ( \sigma ^{2}+\stackrel{\rightarrow } { \pi } % ^{2}\right ) -\frac{\lambda } { 4}\left ( \sigma ^{2}+\stackrel{\rightarrow } { % \pi } ^{2}\right ) ^{2 } \label{sigma } \\ & & + \overline{n}i\gamma ^{\mu } \partial _ { \mu } n+g\overline{n}\left ( \sigma + i\gamma _ { 5}\stackrel{\rightarrow } { \tau } \cdot \stackrel{\rightarrow } { \pi } \right ) n \nonumber\end{aligned}\ ] ] where @xmath7 is the isotriplet pion fields , @xmath0 is the isosinglet field , and @xmath8 is the isodoublet nucleon field . to discuss the symmetry property ,
it is more useful to write this lagrangian as : @xmath9 ^{2 } \label{sigma2 } \\ & & + \overline{n}_{l}i\gamma ^{\mu } \partial _ { \mu } n_{l}+\overline{n}% _ { r}i\gamma ^{\mu } \partial _ { \mu } n_{r}+g(\overline{n}_{l}\sigma n_{r}+% \overline{n}_{r}\sigma ^{\dagger } n_{l } ) \nonumber\end{aligned}\ ] ] where @xmath10 this lagrangian is now clearly invariant under transformation , @xmath11 where @xmath12 are two arbitrary @xmath13unitary matrices .
thus the symmetry group is @xmath14 and representation contents under this group are @xmath15 : the nucleon mass term @xmath16transforms as @xmath17 representation and is not invariant .
one way to construct invariant nucleon mass term is to introduce another doublet of fermions with opposite parity , @xmath18
so that the term @xmath19is invariant . this will give same mass to both doublets and is usually called parity doubling . as we shall see later , another way to give mass to nucleon is by spontaneous symmetry breaking which does not require another doublet .
the general form of noether current is of the form @xmath20 where @xmath21 is the infinitesimal change of the fields under the symmetry transformations .
we have for the left - handed transformation , @xmath22 and @xmath23 + % \overline{n}_{l}\gamma _ { \mu } \frac{\tau ^{a}}{2}n_{l}\ ] ] similarly , @xmath24 and @xmath25 + % \overline{n}_{r}\gamma _ { \mu } \frac{\tau ^{a}}{2}n_{r}\ ] ] the corresponding charges are given by @xmath26 using the canonical commutation relations , we can derive @xmath27 = i\varepsilon _ { ijk}q_{l}^{c},\qquad \left [ q_{r}^{a},q_{r}^{b}\right ] = i\varepsilon _ { ijk}q_{r}^{c},\qquad \left [ q_{l}^{a},q_{l}^{b}\right ] = 0\ ] ] which is the @xmath28 algebra .
the vector and axial charges are given by @xmath29 in particular , the axial charges are @xmath30 \label{axialcurr}\ ] ] : another way to describe the symmetry of the @xmath0-model is the @xmath31 symmetry , which is isomorphic to @xmath32 locally and is characterized by @xmath33 orthogonal matrix , @xmath34 the infinitesimal transformation is @xmath35 the scalar field @xmath36 transform as 4-dimensional vector , @xmath37 the combination @xmath38is just the length of the vector @xmath39 and is clearly invariant under the rotations in 4-dimension .
if we take @xmath40 we get @xmath41 thus we see from eqs([lefttran],[righttran ] ) that the parameters @xmath42 correspond to vector transformations and @xmath43 the axial transformation
. the classical ground state is determined by minimum of the self interaction of scalars , @xmath44 the minimum of the potential is located at @xmath45 which is a 3-sphere , @xmath46 in the 4-dimensional space formed by the scalar fields .
each point on @xmath46 is invariant under @xmath47 rotations .
for example , the point @xmath48 is invariant under the rotations of the first 3 components of the vector .
then , after a point on @xmath46 is chosen to be the classical ground state , the symmetry is broken spontaneously from @xmath31 to @xmath49 note that different points on @xmath46 are related to each other by the action of those rotations which are in @xmath31 but not in @xmath50 these rotations are usually denoted by @xmath51(this is called the coset space . ) . thus we can identify 3-sphere with @xmath52 for the quantum theory , we need to expand the fields around the classical values @xmath53 here @xmath54 is usually called the vacuum expectation value ( vev ) .
then we see that @xmath55 @xmath56 thus @xmath57 are massless and @xmath8 s are massive.remark:if we had made another choice for vev , e.g . @xmath58
the physics is still the same as we will now illustrate . in this case , we write @xmath59 to get @xmath60 @xmath61 thus we still have 3 massless scalar fields . for the nucleon , if we define @xmath62 we get @xmath63 which the mass term for the new field .
it is easy to see that @xmath64 @xmath65form an unbroken @xmath66 algebra .
there are several interesting features worth noting : @xmath67 s are massless .
this is a consequence of the goldstone theorem which states that spontaneous symmetry breaking ( ssb)of a continuous symmetry will give massless particle or zero energy excitation .
this theorem will be discussed in more detail in next subsection .
@xmath68after ssb , the original multiplet @xmath69splits into massless @xmath70 and massive @xmath71 also the nucleons become massive .
thus eventhough the interaction is @xmath72 symmetric the spectrum is only @xmath73 symmetric .
this is the typical consequence of ssb . in some sense ,
the original symmetry is realized by combining the @xmath66 multiplet , e.g. @xmath8 , with the massless goldstone bosons to form the multiplets of @xmath74 this , as we will discuss later , is the basis of the low energy theorem .
@xmath75the axial current in eq([axialcurr ] ) after the ssb will have a term linear in @xmath76 field , @xmath77 which is responsible for the matrix element , @xmath78 using this matrix element in @xmath76 decay , we can identify the vev @xmath54 with the pion decay constant @xmath79 this coupling between axial current @xmath80 and @xmath81 will give rise to a massless pole . @xmath82the appearance of the cubic term @xmath83 and the mass term @xmath84 is the result of the spontaneous symmetry breaking . since these terms have dimension 3 , they are usually called soft breaking , in contrast to the dimension 4 hard breaking terms .
@xmath85 in the scalar self interaction , quartic , cubic , and quadratic terms have only 2 parameters , @xmath86 this means that these 3 terms are not independent , and there is a relation among them .
this is an example of low energy theorem for theory with spontaneous symmetry breaking .
the ssb leads to many relations which are quite different from the usual symmetry breaking .
the most distinct ones are relations among amplitudes involving goldstone bosons in low energies .
as we have mentioned before , these relations are consequence of the fact that goldstone bosons are massless and can be tagged on to other particles to form a larger multiplet . since goldstone bosons do carry energies ,
this is possible only in limit that goldstone bosons have zero energies .
consider the following processes involving the goldstone bosons in the external states .
@xmath87the tree - level contributions are coming from diagrams in fig1 , the amplitudes for these diagrams are given by , @xmath88 @xmath89\ ] ] here @xmath90 @xmath91 and @xmath92 are the usual mandelstam variables , @xmath93 . in the limit where pions have zero momenta , @xmath94 we get @xmath95 @xmath96 and @xmath97 = 0\ ] ] where we have used @xmath98 thus the amplitude vanishes in the soft pion limit , i.e. @xmath99 @xmath100 similar calculation gives @xmath101\ ] ] in the soft pion limit , @xmath102 we get @xmath103 this is the same as the limit , @xmath104 because soft pion means pion momentum much smaller than @xmath105 these are simple examples of the low energy theorem which says that physical amplitudes vanish in the limit where myomata of goldstone bosons go to zero . in examples above , the vanishing of these amplitudes results from some cancellation among different contributions . since this is a general property of the goldstone boson , there should be a better way of getting this .
it turns that one can change the variables representing the scalar fields such that goldstone bosons always enter with derivative coupling .
then the vanishing of the amplitudes involving goldstone boson is manifest .
this can be accomplished by the which we will now describe briefly(@xcite ) .
suppose we start from a lagrangian with field @xmath106 and make a transformation to a new field @xmath107 with the relation , @xmath108 where @xmath109is some power series in @xmath110if we impose the condition that @xmath111 the free lagrangian for @xmath112 will be the same as that for @xmath113 then according to a general theorem valid with rather weak restrictions on the lagrangian and @xmath114 the on - shell matrix elements calculated with @xmath112 fields and with @xmath106 fields are the same .
we will use this field redefinition to write the goldstone boson interaction in terms of derivative coupling . consider a simplified lagrangian given by @xmath115 + \frac{\mu ^{2}}{2}% \left ( \sigma ^{2}+\pi ^{2}\right ) -\frac{\lambda } { 4}\left ( \sigma ^{2}+\pi ^{2}\right ) ^{2}\ ] ] which is just the @xmath116 version of the @xmath0-model without the nucleon . as before
, the ssb will require the shift of the @xmath0-field , as in eq([vev ] ) and @xmath76 field is massless ( goldstone boson ) .
equivalently , we can use a complex field defined by @xmath117 so that the lagrangian is of the form @xmath118 the symmetry transformation is then @xmath119 @xmath120 is some constant .now we use the polar coordinates for the complex field @xmath121 to write the lagrangian in the form , @xmath122
this clearly shows that @xmath123 is massless and has only derivative couplings .
this follows from the fact that the @xmath124 ( or @xmath125 ) symmetry @xmath126 corresponds to @xmath127 which is inhomogeneous .
so @xmath123 needs to have a derivative in order to be invariant under such inhomogeneous transformation .
note that this lagrangian , due to the presence of terms like @xmath128 is not renormalizable .
but this lagrangian will be used only as an effective theory to study the low energy phenomenology while the renomalizability deals with high energy behavior . from eq([ssbv ] ) we see that the pions @xmath129 are massless .
this is a consequence of the goldstone theorem which states that spontaneous breaking of a continuous symmetry will give a massless particle or zero energy excitation .
we will first illustrate this by showing that quadratic terms in @xmath129 are absent in the tree level as a consequence of spontaneous symmetry breaking of the original chiral symmetry .. we will use the @xmath31 notation , @xmath130the invariance under the chiral transformation implies that @xmath131 differentiating eq([invv ] ) with respect to @xmath132 and then evaluating this at the minimum , we see that @xmath133 for the case @xmath134 we see that in the expansion of @xmath135 around the minimum , @xmath136 there are no terms of the form , @xmath137 with @xmath138 therefore @xmath139 are massless in the tree level .
we can extend this to more general case where the effective potential is written as @xmath140 this potential is invariant under some symmetry group @xmath141 which transforms @xmath39 as @xmath142 where @xmath143 are the parameters for the infinitesimal transformations and @xmath144 matrix for the representation where @xmath39 belongs .
the invariance under these transformation implies that @xmath145 the minimum of the potential is located at @xmath146 which satisfies the equation , @xmath147 differentiating eq ( [ minimum ] ) with respect to @xmath132 and evaluating this at the minimum , @xmath146 we get @xmath148 this means that the vector @xmath149 if non - zero , is an eigenvector of the mass matrix @xmath150 with zero eigenvalue(massless ) .
thus the number of massless goldstone bosons is just the number of independent vectors of the form , @xmath151 in other words , if @xmath152 the combination @xmath153 is the goldstone boson , up to a normalization constant .
these arguments only show that @xmath129 s are massless in the tree level .
it turns out that this property is true independent of perturbation theory and can be illustrated in case of @xmath154-model as follows .
the axial charge is of the form , @xmath155\ ] ] @xmath156since @xmath157 and @xmath158 are just the momenta conjugate to @xmath81 and @xmath159 we can derive , @xmath160 = \delta _ { ij}\sigma \left ( 0\right ) .\ ] ] between vacuum states , this yields @xmath161 |0>=\delta _ { ij}<0|\sigma \left ( 0\right ) |0>. \label{goldstone}\ ] ] for the case of ssb , we have @xmath162 note that this condition implies that the axial charges @xmath163 s do not annihilate the vacuum , @xmath164using @xmath165 we can write the lhs of eq ( [ goldstone ] ) as @xmath166|0\rangle & = & i\int d^{3}x\langle 0|\left [ a_{i}^{0}\left ( x\right ) , \pi _ { j}\right ] |0\rangle = % \stackunder{n}{i\sum } \delta ^{3}\left ( \stackrel{\rightarrow } { p_{n}}% \right ) [ \langle 0|a_{i}^{0}\left ( 0\right ) |n\rangle \nonumber \\ & & \langle n|\pi _ { j}|0\rangle e^{-ie_{n}t}-\langle n|\pi _ { j}|0\rangle \langle 0|a_{i}^{0}\left ( x\right ) |n\rangle e^{ie_{n}t}]\end{aligned}\ ] ] this has explicit dependence on @xmath91 while the right hand side , @xmath167 this is the content of the goldstone theorem . for the relativistic system ,
the energy and momentum is related by @xmath168then eq ( [ massless ] ) implies the existence of massless particle in the system .
more specifically , there are physical states @xmath169 with the property that @xmath170 and are massless from goldstone theorem .
it is convenient to choose the normalization such that @xmath171 and write @xmath172 it is easy to see that @xmath173 for the non - relativistic system eq([massless ] ) simply says that the dispersion relation @xmath174has zero energy excitation . in the @xmath0-model without the nucleons
, we have 3 massless @xmath76 s and a massive @xmath0 field . for the energies much smaller than @xmath175 the massless goldstone bosons are the important physical degrees of freedoms and it is desirable to write down an effective theory with @xmath76 s only . as we have seen , the theory with ssb has many physical consequences , e.g. low energy theorem for the goldstone bosons .
the removal of @xmath0-field should preserve symmetry so that these results are maintained . also , phenomenologically there are no good evidence for the existence of the @xmath176 meson which is the partner of @xmath57 in the chiral symmetry .
we now discuss the explicit steps for carrying out this process .
write the scalar fields as a vector in 4-dimensional space , @xmath177 we want to parametrize the @xmath106 fields in such a way that the non - goldstone field to be eliminated later is @xmath31 invariant .
one simple parametrization for this purpose is @xmath178 where @xmath179 a @xmath180 orthogonal matrix , @xmath181 which gives @xmath182 and @xmath183 so @xmath184 is the magnitude of the vector @xmath39 and is clearly @xmath31 invariant .
thus it can be eliminated without effecting the symmetry .
one simple choice for @xmath185is , ( @xcite ) @xmath186 note that we can invert these relations to get @xmath187 the lagrangian is of the form @xmath188 + \frac{1}{2}% \mu ^{2}s^{2}-\frac{\lambda } { 4}s^{4}\ ] ] so @xmath189 are the massless goldstone bosons . to study the physics of goldstone bosons at low energies , @xmath190 , we can replace the @xmath191 field by a constant , @xmath192 to get @xmath193 in order to get the correct nomoralization we rescale , @xmath194 so that @xmath195 here the interaction terms will always contain derivatives and amplitudes involving @xmath196 will vanish in the limit of zero momenta ( low energy theorem ) . according to the theorem of field redefinition
, this describes the same physics as the usual @xmath0-model lagrangian in the goldstone sector .
for example , we can check that in the simple case of @xmath197 scattering in the tree level the amplitude from this lagrangian in eq ( [ nonlinl ] ) is @xmath198 = \frac{i}{v^{2}}\left ( s+t+u\right ) .\ ] ] this is the same result as in eq([pipi ] ) , obtained in the @xmath0-model with @xmath199 the lagrangian in eq ( [ nonlinl ] ) which contains on the goldstone boson fields , is one example of non - linear realization of chiral symmetry , which will be discussed in detail in the next section . here
we want to mention a useful geometric interpretation of lagrangian in eq ( [ nonlinl ] ) . when we eliminate the @xmath31 invariant field by setting @xmath200 , @xmath201
satisfy the relation , @xmath202 which is just the sphere with radius @xmath54 in 4-dimensional euclidean space , @xmath203 the variables @xmath204 and @xmath205 are just one particular choice of the coordinates of the space @xmath206 the transformation of lagrangian from @xmath106 fields to @xmath112 fields can be understood in terms of metric tensor in @xmath206 for simplicity , consider just the kinetic terms in @xmath207 , @xmath208 where @xmath209 is the trivial metric in the 4-dimensional euclidean space .
then the transformation @xmath210 gives @xmath211 and @xmath212 where @xmath213 is the induced metric on @xmath206 thus in the lagrangian the coefficient of @xmath214 is the metric of the space @xmath215 which is just the coset space @xmath52 in the general case where the symmetry breaking is of the form , @xmath216 the non - linear lagrangian can be written down with the metric on the manifold @xmath217 as in eq ( [ metricl ] ) .
it is interesting to see how the transformations of @xmath218 are realized on this manifold @xmath206 for the infinitesimal isospin rotation , we have @xmath219 which implies from eq([s3para2 ] ) that @xmath220 this is just a rotation on the vector @xmath221 and the lagrangian in eq([metric l ] ) with metric given in eq([metrics3 ] ) is clearly invariant under such transformation .
the axial transformation on @xmath222 and @xmath0 is of the form @xmath223 which gives @xmath224 this transformation is non - linear and inhomogeneous .. but we can get simple transformation for the combination , @xmath225 this looks very much like an isospin rotation except that the parameters for the rotation now depend on the fields @xmath112 and it is easy now to see that the lagrangian in eq([metricl ] ) is invariant under the axial transformations .
: we can transform the metric in eq([metrics3 ] ) into the more familiar robertson - walker metric used in cosmology as follows .
first we use the spherical coordinates for @xmath221 to write the metric in the line element as @xmath226 .
\nonumber\ ] ] define the new variable @xmath227 by @xmath228 in terms of new variable the line element is of form , @xmath229 which is just the usual robertson - walker metric for the case of positive curvature .
as we have seen in the last section , it is useful to write down a lagrangian with only goldstone bosons as an effective theory to describe physics at low energies . in this section
we will discuss the general description of this procedure , which is usually called the non - linear realization .
the usual discussion of this subject is generally rather formal and abstract .
the discussion here will emphasize the intuitive understanding rather than the mathematical rigor . in order to make the discussion here somewhat self contained
, we will first discuss some simple results from group theory ( @xcite ) which are useful for the understanding of non - linear realization .
then we discuss the general features of non - linear realization . here
we will recall the rearrangement theorem which is central to most of the group theoretical result and then discuss the concept of coset space which forms the basis of the non - linear realization .
@xmath87**rearrangement theorem**let @xmath230 be a finite group . if we multiply the whole group by an arbitrary group element @xmath231i.e .
@xmath232 the resulting set is just the group @xmath233 itself .
@xmath100**coset space**coset space decomposes a group into non - overlapping sets with respect to a subgroup .
let @xmath234 be a non - trivial subgroup of @xmath233 .
for any element @xmath235 in @xmath233 but not in @xmath236 , the left coset @xmath237 , or coset for short , is just @xmath238 the coset @xmath239 will not have any element in common with the subgroup @xmath240 and any two such cosets are either identical or have no elements in common .
this can be seen as follows .
consider cosets @xmath241 and @xmath242 suppose that there is one element in common , @xmath243 then we can write @xmath244 which means @xmath245 is one of the element of subgroup @xmath246 then by the rearrangement theorem applied to the subgroup @xmath240 we get @xmath247 i.e. these two cosets have the same group elements .
the group @xmath233 is now decomposed into these non - overlapping cosets and the collection of all the distinct cosets @xmath248 together with @xmath240 will contain all the group elements of @xmath249 this is denoted by @xmath250 this can be generalized to lie group where rearrangement theorem is valid .
: consider points on 2-dimensional plane which form a group under the addition , @xmath251 clearly , points on the @xmath252-axis , @xmath253 form a subgroup , denoted by @xmath246 then the vertical line of the form , @xmath254with a fixed is a coset with respect to the subgroup @xmath246 it is clear that the whole 2-dimensional plane can be decomposed into collection of such vertical lines .
we can label these cosets , vertical lines , by choosing one element from each coset .
clearly there are many ways to choose such representatives .
one convenient parametrization is to choose those points on the @xmath255-axis , so that the cosets are of the form @xmath256 each group element @xmath257 can be written as the product , @xmath258 where @xmath259 under the action of an arbitrary group element @xmath260 this will give @xmath261 ( the computation here is organized in such a way that it parallel to the more complicate case in the non - linear realization . )
thus the group element @xmath262 will move the points in the coset @xmath263 to points in the coset @xmath264 in terms of coset parameters , we have @xmath265 we will first discuss the general machinery of non - linear realization ( @xcite@xcite ) and then take up the special case of chiral symmetry where the parity symmetry will make the realization much simpler .. suppose the symmetry group @xmath233 is spontaneously broken to a subgroup @xmath246 where both @xmath233 and @xmath236 are lie groups .
choose the generators of @xmath233 to be of the form @xmath266 such that @xmath267 are the generators of the subgroup @xmath246 as usual , the group elements in @xmath236 can be written in the form , @xmath268 where @xmath269 are the group parameters for @xmath236 and are taken to be real . from the coset decomposition , we can write an arbitrary group element in @xmath233 as @xmath270 here @xmath271are the parameters which label the coset @xmath272 note that since the vacuum is invariant under @xmath240 we have @xmath273 i.e. @xmath274 also labels the different vacua , which are degenerate .
recall that the different vacua form the manifold @xmath250 thus the coset parameters are also the parameters for the manifold @xmath250 under the action of an arbitrary group element @xmath275 we have the combination @xmath276 since @xmath277 is also a group element in @xmath233 , we can write a coset decomposition , @xmath278 and then @xmath279 where we have used the group property of @xmath236 to write @xmath280 note that the new coset parameters @xmath281 and parameters @xmath282for the subgroup @xmath240all depend on the original coset parameters @xmath283@xmath284 in this way , the group element @xmath285 transform the coset parameters from @xmath286 in the coset space @xmath250 as we will see later , these coset parameters will be identified with the goldstone bosons .
these transformations on @xmath287 and @xmath288 induced by the group elements will have the same group properties as the group elements and are called the _ non - linear realization _ of the group .
this is in contrast to the usual representation of the group where group elements are represented by matrices . in the transformation in eq([coset ] ) @xmath289 is generally not a linear function of @xmath290 but for the special case where @xmath291 is a group element from the unbroken subgroup @xmath240 we get , from eq([g ] ) , @xmath292 where @xmath293 in general , the broken generators @xmath294 transform as some representation @xmath295 with respect to the subgroup @xmath240@xmath296 for example , in the case of @xmath297 model , the broken generators , @xmath298 transform as triplet under the unbroken subgroup @xmath299 we can then write @xmath300 where @xmath301 this means that @xmath302 transform linearly under the subgroup @xmath303 also it is easy to see that the parameters @xmath304 are independent of the coset parameters @xmath305 .
but if the group element is of the form @xmath306 the transformation law for the coset parameters @xmath307 is non - linear and quite complicate . for the case of chiral symmetry ,
there is significant simplification due to parity operation .
we will illustrate this in the simple case of @xmath308 symmetry .
the parity operation is of the form , @xmath309 consider the case where group element @xmath310 consists of left - handed transformation , @xmath311 and write the transformation of coset parameters as @xmath312 under the parity transformation , the left - handed transformation is changed into right - handed one , @xmath313 then applying the parity transformation to eq([l ] ) , we get @xmath314 using the notation @xmath315 we can combine the eqs([l],[r ] ) into @xmath316 note that as we have mentioned before , the parameters @xmath317 depend on the coset parameters @xmath283 the transformation law for @xmath318 here is non - linear because the factor @xmath319 depends on @xmath282 .
however , the combination @xmath320 will have a simple transformation law , @xmath321 since @xmath322 are independent of the coset parameters @xmath283this transformation law is linear and will be useful for constructing lagrangian .
we are interested in the cases where spontaneous symmetry breaking is generated by the scalar fields .
some of these scalar fields become the massless goldstone bosons , like the pions in the @xmath0-model and others remain massive , like @xmath0-field .
consider the scalar fields in the @xmath176-model , where we will use the notation , @xmath323 the vacuum expectation value which gives the classical ground state is of the form , @xmath324 suppose we make a very general assumption that from a given point in @xmath106-space we can reach any other point by some group transformation.(space is transitive ) .
then the general field configuration can be written as @xmath325 : strictly speaking we should use the representation matrices @xmath326 of the group element @xmath310 rather than @xmath310 itself .
however for simplicity of notation , @xmath310 here is a shorthand for @xmath327 from the coset decomposition in eq([g]),we can write @xmath328 here we have chosen coset parameters to be the pion fields
. then we can write the scalar fields as @xmath329 where we have used the fact that the vector @xmath330 is proportional to the vacuum configuration @xmath48 and is invariant under the subgroup @xmath331 since the goldstone bosons are identified as the coset parameters , they transform the same way as @xmath332 in eq([coset ] ) , i.e. @xmath333 where @xmath334 for the case @xmath335we have from eq([linearrep ] ) , @xmath336 : the scalar fields here have property that after separating out the goldstone bosons , the remainders are proportional to the vacuum expectation value and is then invariant under the subgroup @xmath246 this is true only for scalar fields in the vector representation in @xmath337 or @xmath338 groups and is not true for scalars in more general representation . in more general cases ,
we can separate out the goldstone bosons by writing @xmath339 as @xmath340 where @xmath341 contains all the massive fields . under the action of group element @xmath285
, we have @xmath342 where @xmath343 to see how the goldstone bosons transform under the axial transformation , we set @xmath344 in eq([linearr ] ) to get @xmath345 to understand this equation better , we first take the infinitesimal transformation , @xmath346@xmath347 and then expand this in powers of @xmath348s , @xmath349 comparing both sides we see that @xmath350 thus there is a inhomogeneous term in the transformation law for the goldstone bosons @xmath129 and this is why goldstone bosons have derivative coupling .
since the transformation law for @xmath351 is linear and simple , it is easier to construct the chirally invariant interaction in terms of @xmath352 rather than @xmath353 it is easy to see that the only invariants without derivatives will involve trace of some powers of @xmath354which is just an identity matrix .
( this also implies that the goldstone boson coupling will involve derivatives ) .
thus the interaction with lowest numbers of derivative is of the form @xmath355 [ [ covariant - derivative ] ] * covariant derivative * : + + + + + + + + + + + + + + + + + + + + + + + the parity symmetry in the @xmath0-model is responsible for getting the simple combination @xmath356 which transforms linearly .
for the more general case where there is no such simplification , to construct invariant terms involving derivatives is quite complicate because the non - linear transformation law will involve golstone boson fields which are space - time dependent .
this means that we need to construct the covariant derivatives .
futhermore , to exhibit the low energy explicitly we need to couple golstone field with derivative to other matter fields .
we will now discuss briefly in the simple case of @xmath0-model .
write the scalar fields @xmath339 in the form , @xmath357 with @xmath358 @xmath359as before under the action of group element @xmath360 we have @xmath361 since @xmath106 transforms linearly , we have @xmath362 which implies @xmath363 this means that @xmath364 transforms non - linearly because @xmath365 in @xmath366 depend on @xmath76 fields . making use of the simplification for the case of chiral symmetry we have , from transformation law,([linearr ] )
@xmath367 and @xmath368 @xmath369 if we define @xmath370 \label{covarv}\ ] ] @xmath371 \label{covara}\ ] ] we get @xmath372 @xmath373 this means that @xmath374 transforms like gauge field , while @xmath375 transforms as global adjoint field .
therefore @xmath374 can be used to construct the covariant derivative and @xmath375 is like a global axial vector field . [ [ nucleon - field ] ] nucleon field + + + + + + + + + + + + + the nucleon field in the linear @xmath0-model has the transformation properties , @xmath376 thus to couple nucleon fields to the goldstone bosons , we could write down the following @xmath5 invariant coupling , @xmath377 however , this is not of the form of derivative coupling which exhibits the low energy theorem explicitly . for this purpose and general non - linear realization , we define a new nucleon field by , @xmath378 under the action of the group element @xmath360we get @xmath379 where @xmath380 thus @xmath381 transforms according to the representation of the subgroup @xmath236 but with the group parameters depend on @xmath76 fields , @xmath382 then from the transformation properties given in eqs([covarv , covara ] ) , we can write down the derivative coupling as @xmath383 which will yield the low energy theorem explicitly .
in the frame work of relativistic field theory , e.g. in @xmath218 @xmath0-model , spontaneous symmetry breaking seems to be put in by hand , i.e. setting the quadratic terms to have negative sign in the scalar potential in order to develope vacuum expectation value .
this is rather ad hoc and no physical reason is given for why this is the case .
we will now discuss some simple non - relativistic examples of spontaneous symmetry breaking in order to shed some light on this ( @xcite ) .
consider a system of @xmath8 spins on an one dimensional lattice with hamiltonian , @xmath384 where @xmath385 is the coupling constant for spin - spin interaction and @xmath386 is the external magnetic field . in this hamiltonian ,
for calculational simplicity we allow every spin to interact with every other spin , while more realistic situation will be the short range nearest neighbor interaction .
but the interest here is to see how the spontaneous symmetry breaking come about and we will ignore this .
the partition function is given by @xmath387 using the identity for the gaussian integral , @xmath388 we can write the partition function as @xmath389\ ] ] now we can sum over each @xmath390 independently , @xmath391 \}\ ] ] and @xmath392 \right)\ ] ] from the partition function @xmath393 , we can compute the average spin , @xmath394
if @xmath395 in the limit the external field vanishes , @xmath396 , then we have spontaneous symmetry breaking . since we are interested in the case where @xmath8 is very large , we can use saddle point method to compute @xmath397 write @xmath398 where @xmath399\ ] ] the minimum of @xmath400 is given by @xmath401 let @xmath402 be the solutions of this transcendental equation , then @xmath403 suppose @xmath404 is the smallest of these solutions , it will dominate the partition function for large @xmath405 then the average magnetization is then @xmath406 where we have used the equation satisfied by @xmath407thus the minimum of @xmath408 will correspond to the average magnetization . to study spontaneous symmetry breaking , we set @xmath409 in eq([minm ] ) , and get @xmath410 it turns out that this equation has only the trivial solution , @xmath411 if @xmath412 and non - trivial solution exists only for @xmath413 to understand this feature , we expand @xmath400 in powers of @xmath255 for the case @xmath409@xmath414 \\ & = & \frac{j}{2}\left ( 1-\beta j\right ) x^{2}+\frac{1}{12}\beta ^{3}j^{4}x^{4}+\cdots \nonumber\end{aligned}\ ] ] thus @xmath415 corresponds to negative quadratic term , which is the familiar situation in the scalar potential in the @xmath0-model and the like . in terms of temperature
this condition , we have @xmath416 here @xmath417 is the coupling which wants to align the spins in the same direction while the effect of temperature is to randomize spins .
thus the condition in eq([critical ] ) simply means that the interaction of spins has to overcome the thermalization in order to produce significant spin alignment .
the temperature @xmath418 is usually called the critical temperature and spontaneous symmetry breaking is possible only for @xmath419 in this simple example , the non - zero magnetization @xmath420 breaks the symmetry , @xmath421 and originates from the competition between spin - spin interaction which aligns the spin and the thermalization which tends to destroy the alignment.remarks:@xmath422from the partition function in eq([partition ] ) we see that the probability to find the system to have @xmath423 is given by the boltzmann factor @xmath424 if @xmath404 is the absolute minimum for @xmath400 , then in the thermodynamic limit @xmath425 we have @xmath426 and the probability for all the other @xmath423 will be zero . @xmath68for the case @xmath427 is near @xmath428the minimum of @xmath408 is located at small values of @xmath255 .
thus we can expand @xmath408 in power series , @xmath429 the minimum is then @xmath430 this means that near the critical temperature @xmath431the dependence of average magnetization on @xmath432 is non - analytic .
this is a typical behavior of physical quantities near the critical point .
the superfluid @xmath433 provides a simple example of goldstone excitation where the excitation energy , @xmath434 goes to zero when the wave number @xmath435 the helium atoms are tightly bounded and the long - distance attractive force between atoms are very weak while the short distance is strongly repulsive .
thus a system of helium atoms can be described as a gas of weakly interacting bosons with hamiltonian , @xmath436 here @xmath437 is the potential describes the effective interaction between helium atoms and @xmath438 is the field operator for the helium atom and satisfies the commutation relation , @xmath439 = \delta ^{3}\left ( x - y\right)\ ] ] we will assume that the system is in a large box of volume @xmath440 with periodical boundary condition . clearly , this hamiltonian is invariant under the transformation , @xmath441 this is just a @xmath124 symmetry which says that the number of he atoms is conserved .
the conserved charge is just the number operator , @xmath442 with the commutation relations , @xmath443 = \psi \left ( x\right ) , \qquad \left [ q,\psi ^{\dagger } \left ( x\right ) \right ] = -\psi ^{\dagger } \left ( x\right ) .
\label{commutator}\ ] ] we can expand @xmath438 in plane waves , @xmath444 where @xmath445 and @xmath446 are the usual creation and annihilation operators satisfying the commutation relations , @xmath447 = 0,\left [ a_{\stackrel{\rightarrow } { k}},a_{\stackrel{\rightarrow } { % k^{\prime } } } ^{\dagger } \right ] = \delta _ { _ { \stackrel{\rightarrow } { k}},_{% \stackrel{\rightarrow } { k^{\prime } } } } \ ] ] the hamilton is then of the form , @xmath448 where @xmath449 in these two equations and there after we have , for notational simplicity , neglected the vector symbol for the wave vectors @xmath450 s. since @xmath437 is real we have @xmath451 for the trivial case where there is no interaction , @xmath452 the ground state is just the one in which all particles are in the @xmath453 state , @xmath454 state , i.e. @xmath455 where @xmath456 we are interested in the cases where @xmath457 the total number of particles , is very large .
thus @xmath458 is very large . from the properties of the creation and annihilation operators , @xmath459
we will make the assumption that the matrix elements of @xmath460 are of order @xmath461 and @xmath462 of order @xmath463 thus in the limit @xmath464 commutator of @xmath465and @xmath466 is of order unity while @xmath467 are of order @xmath468@xmath469 = 1\ll a_{0}\,\quad \text{or\quad } % a_{0}^{\dagger } \symbol{126}\sqrt{n_{0}}\ ] ] thus we can neglect the commutator @xmath470 since @xmath471 and @xmath462 commute with all the other operators , @xmath472 = \left [ a_{0}^{\dagger } , a_{k}\right ] = 0\qquad \text{for } k\neq 0.\ ] ] we can then take @xmath460 and @xmath462 to be c - numbers ( schur s lemma ) , @xmath473 thus we will replace @xmath460 and @xmath462 by @xmath474 then the coefficients of terms quadratic in @xmath475 and @xmath476 @xmath477 in the interaction will be of order @xmath478and those of quartic term is of order 1 .
therefore we can make the approximation that neglect the quartic terms and get the hamiltonian in the form , @xmath479+\frac{n_{0}^{2}}{2\omega } \stackrel{\_% } { v}\left ( 0\right)\end{aligned}\ ] ] or @xmath480 where @xmath481 and we have used @xmath482 note that this hamiltonian does not conserve the particle number but it conserves the momentum because the removal of @xmath483 mode effects the particle number but not the momentum . since this hamiltonian contains only quadratic terms , we can solve this by bogoliubove transformation as follows .
define the quasi - particle operators by @xmath484 then we have @xmath485 where @xmath486is an arbitrary parameter at our disposal .
we now write the hamiltonian in terms of the quasi particle operators by inverting the relations in eq([bogo1,bogo2 ] ) , @xmath487 and choose the parameter @xmath488 so that the coefficient of the non - diagonal terms @xmath489 is zero . the computation is straightforward and the result is @xmath490 and the hamiltonian is @xmath491 where @xmath492 this is just the hamiltonian for the uncoupled harmonic oscillators and the eigenvalues are @xmath493 the quai - particle energy excitation has the property that @xmath494 which is just the goldstone excitation .
clearly , the ground state @xmath495 is the one which is annihilated by all quasi particle operators @xmath496@xmath497 and the excited states are of the form , @xmath498 it is straightforward to show that the quasi particle ground state can be written in terms of the original creation operators @xmath499 as @xmath500 where @xmath501 this shows that the quasi particle ground state is a complicate combination of the vacuum , 2 particle states , 4 particle states,@xmath502etc . to elucidate the goldstone theorem
, we note that the vacuum expectation value of @xmath503 in the ground state is non - zero , @xmath504 where we have used the fact that @xmath505 for @xmath506from the momentum conservation . from the commutation relation in eq ( [ commutator ] ) we see that this is the symmetry breaking condition which implies that @xmath507 the quasi particle excitation which has the property that its energy @xmath508 goes to zero in the limit @xmath509 is the goldstone excitation implies by the goldstone s theorem . | in this introductory lecture , some basic features of the spontaneous symmetry breaking are discussed .
more specifically , @xmath0-model , non - linear realization , and some examples of spontaneous symmetry breaking in the non - relativistic system are discussed in details .
the approach here is more pedagogical than rigorous and the purpose is to get some simple explanation of some useful topics in this rather wide area . . |
our main goal in this paper is to obtain a simpler proof that , under the sweep map , the @xmath3 statistic of a rational dyck path @xmath4 becomes the @xmath5 statistic of its image path @xmath6 .
the first proof of this remarkable identity is due to loehr and warrington in @xcite .
there is also a second proof due to gorsky and mazin in @xcite . inspired by a recent work of nathan williams @xcite ,
we have come to depict rational dyck paths in a manner which makes the ranks of the vertices of a path consistent with its visual representation .
this very simple change turns out to be conducive to considerable simplifications in proving many of the properties of rational dyck paths .
witness the proof of the invertibity of the sweep map given in @xcite . in figure
[ fig : goodpair ] we have depicted a path @xmath7 in the @xmath8 lattice rectangle and its sweep map image @xmath6 as they are traditionally depicted .
the ranks of the vertices of an @xmath2-path are constructed by assigning @xmath9 to the origin @xmath10 and adding an @xmath0 after a north step and subtracting an @xmath1 after an east step .
the sweep image of @xmath4 is obtained by letting the main diagonal sweep ( parallel to its original position ) from right to left , the vertices of @xmath11 and drawing a north step of @xmath6 when we sweep the south end of a north step of @xmath7 and draw an east step when we sweep the west end of an east step .
the steps of @xmath6 can also be obtained by rearranging the steps of @xmath7 by increasing ranks of their starting vertices .
the ` area ` of a rational dyck path is equal to the number of lattice cells between the path and the main diagonal . by a result in (
? * lemma 11 ) a cell @xmath12 above an @xmath2-dyck path @xmath6 contributes a unit to its dinv statistic if and only if the starting rank @xmath13 of the east step of @xmath6 below @xmath12 and the starting rank @xmath14 of the north step of d to the right of @xmath12 satisfy the inequality @xmath15 . in our display , the cells contributing to @xmath16 are distinguished by a green square .
a quick count reveals that @xmath17 to proceed we need some notation .
let @xmath2 be a co - prime pair of positive integers .
a tableau @xmath18 consists of an ordered set of @xmath1 red arrows and @xmath0 blue arrows , placed on a @xmath19 lattice rectangle . where @xmath20 is a large positive integer to be specified .
a red arrow is the up vector @xmath21 and a blue arrow is the down vector @xmath22 .
the rows of lattice cells will be referred to as rows and the horizontal lattice lines will be simply referred to as lines . on the left of each line
we have placed its @xmath23 coordinate which we will refer to as its level .
the level of the starting point of an arrow is called its starting rank , and similarly its end rank is the level of its end point
. it will be convenient to call row @xmath24 the row of lattice cells delimited by the lines at levels @xmath24 and @xmath25 . given a word @xmath26 with @xmath1 letters @xmath27 and @xmath0 letters @xmath28 , and a sequence of @xmath29 weakly positive ranks @xmath30 , the tableau @xmath31 ( see figure [ fig : tableau ] ) is obtained by placing the letters of @xmath26 at the bottom of the lattice columns and drawing in the @xmath32 column an arrow with starting rank @xmath33 and @xmath34 if the @xmath32 letter of @xmath26 is an @xmath27 or @xmath35 if the @xmath32 letter of @xmath26 is a @xmath28 .
figure [ fig : tableau ] depicts our manner of drawing the path @xmath4 .
the ranks of @xmath4 are now the circled levels of the starting vertices .
notice that each lattice cell may contain a segment of a red arrow or a segment of a blue arrow or no segment at all .
the red segment count of row @xmath36 will be denoted @xmath37 and the blue segment count is denoted @xmath38 .
we will set @xmath39 and refer to it as the count of row @xmath36 .
notice that , for obvious reasons , the row counts in the above tableau are invariably all equal to zero .
this property has the following immediate consequence .
[ p-1.1]the starting rank of any arrow @xmath40 of @xmath6 may be simply obtained by drawing in green the line at the starting level of its preimage @xmath41 , then counting the segments above the green line of any red arrow of @xmath4 that starts below the green line and adding to that count the number of segments below the green line of any blue arrow of @xmath4 that starts above the green line .
the desired rank is @xmath42 , when @xmath4 has @xmath14 red arrows and @xmath13 blue arrows that start below the green line .
the reason this number can be obtained as stated is due to the zero sum property .
indeed , the count of red segments above the green line is certainly needed .
the remaining portion of @xmath42 is obtained by counting blue segments below the green line since the latter are subtracted from the remaining portion to make equal to zero the line counts below the green line .
see figure [ fig : bigpair ] for an example .
this proposition leads to the basic formula we will use to compute the area of the image @xmath6 by working on the preimage @xmath4 .
[ t-1.1 ] for any @xmath2-dyck path @xmath6 , with @xmath2 co - prime , we have @xmath43 where `` @xmath44 '' denotes the rank of the south end of the @xmath45 north step of @xmath6 .
notice that that if we define the rank of the lattice cell whose south - west coordinates are @xmath46 by setting @xmath47 , then the least positive rank in the @xmath45 row is none other than the remainder of @xmath48 mod @xmath1 . since the residues modulo @xmath1 are @xmath49 and the least positive ranks are distinct , it follows that their sum is @xmath50 .
calling @xmath51 the least positive rank in row @xmath36 , it is evident that @xmath52 is the contribution of row @xmath36 to the area of @xmath6 .
this given , we see that is simply obtained by summing all these contributions .
given a co - prime pair @xmath2 , our argument is to show that as we decrease the area of the preimage @xmath4 by one unit , both dinv and area satisfy the same recursion .
we have depicted in figure [ fig : redblueb ] the cell that we have subtracted from the preimage @xmath4 to obtain a preimage @xmath53 with one less unit of area .
this operation may be viewed as replacing a red arrow @xmath27 of @xmath53 by a dashed red arrow @xmath54 and a blue arrow @xmath28 by a dashed blue arrow @xmath55 .
calling @xmath6 and @xmath56 the sweep map images of @xmath4 and @xmath53 , our task is to determine the difference @xmath57 .
our tool will be formula and the fact that the starting rank of any arrow @xmath40 of @xmath6 is @xmath58 where @xmath4 has @xmath14 red arrows and @xmath13 blue arrows that start below the preimage @xmath59 of @xmath40 in @xmath4 .
this difference will be calculated in @xmath4 and @xmath53 by means of our tool .
now there are @xmath60 distinct cases .
firstly , red arrows that start above level @xmath61 ( see figure ) or below level @xmath62 are not affected by the replacements .
thus their contribution to the difference is @xmath9 .
secondly in the case of any red arrow starting strictly below level @xmath63 and strictly above level @xmath64 , its contribution to the difference is @xmath1 .
the reason for this is that both red arrows in the display will increase the ranks of the arrows of @xmath6 and @xmath56 , whose starting levels are in this range by an equal amount , therefore they cancel each other .
by contrast the dashed blue arrow will affect the ranks of the red arrows of @xmath56 so that the contribution of each to the difference is @xmath65 .
thirdly , notice that each red arrow of @xmath6 or @xmath56 that starts strictly below rank @xmath64 and strictly above rank @xmath62 , is not affected by either of the blue arrows .
but each of the red arrows of @xmath56 is affected by the dashed red arrow and thus contributes a @xmath66 to the difference .
the reason is that the arrows that start below level @xmath62 contribute equally to the areas of @xmath27 and @xmath54 .
thus they cancel in computing the desired difference . on the other hand
all the arrows accounted for in do contribute to the area of @xmath27 but not to the area of @xmath54 .
notice first that the contribution to the area difference in the third case is the negative of the first part of the contribution obtained in .
it results from this that all the remaining contributions are multiples of @xmath1 .
( a convenient fact since , according to , @xmath1 has to be divided out . ) furthermore , in the second case this multiple counts the number of red arrows that have a red segment in @xmath68 or @xmath69 .
finally , we see in that the factor of @xmath1 that survives the cancellation , counts exactly the arrows that have a blue segment in @xmath70 or @xmath71 .
[ p-2.1]let @xmath72 be obtained from @xmath4 by removing an area cell and let @xmath6 and @xmath56 be their sweep map images .
let @xmath70 and @xmath71 be the blue regions and @xmath68 and @xmath69 be the red regions in figure [ fig : redblueb ] .
then @xmath73 where @xmath74 denote red segment counts and @xmath75 denote blue segment counts in the corresponding regions .
[ p-2.2 ] the dinv statistic of a rational dyck path @xmath6 ( given in our stretched form ) may simply be obtained by counting the pairs @xmath76 consisting of a blue arrow to the left of a red arrow , such that @xmath77 must touch @xmath78 as it moves to the right past @xmath78 .
suppose that @xmath77 has starting rank @xmath13 and @xmath78 has starting rank @xmath14 . by the quoted result of loehr - warrington , the pair @xmath76 contribute a unit to @xmath79 if and only if @xmath80 .
this is equivalent to requiring that @xmath81 and @xmath82 .
but these two inequalities are precisely what is needed to guarantee that @xmath77 must touch @xmath78 as it moves horizontally past it .
[ p-2.3]let @xmath53 be obtained from @xmath4 by removing an area cell .
let @xmath70 and @xmath68 be the blue regions and @xmath71 and @xmath69 be the red regions in figure [ fig : redbluea ] . then @xmath83 where @xmath84 denote blue segment counts and @xmath85 denote red segment counts in the corresponding regions .
since @xmath53 is obtained from @xmath4 by replacing the solid arrows @xmath87 by dashed arrows @xmath88 ( see figure [ fig : redbluea ] ) , we can divide the contribution of a pair @xmath89 to the difference @xmath90 into four cases .
\(2 ) both @xmath77 and @xmath78 are in the displayed arrows .
this can only happen in two ways : @xmath91 in @xmath4 ( no dinv ) or @xmath92 in @xmath53 ( @xmath93 dinv ) .
therefore the contribution to the difference in this case is @xmath94 .
\(3 ) only @xmath77 is one of the displayed arrows .
then we need to consider the pairs @xmath95 in @xmath4 and @xmath96 in @xmath53 .
their contribution to the difference is @xmath93 if @xmath78 has a red segment in @xmath69 , @xmath94 if @xmath78 has a red segment in @xmath71 and @xmath9 if @xmath78 does not have a segment in neither @xmath69 or @xmath71 .
\(4 ) only @xmath78 is in the displayed arrows .
then we need to consider the pairs @xmath97 in @xmath4 and @xmath98 in @xmath53 .
their contribution to the difference is @xmath93 if @xmath77 has a blue segment in @xmath68 , @xmath94 if @xmath77 has a blue segment in @xmath70 and @xmath9 if @xmath77 does not have a segment in neither @xmath68 or @xmath70 .
we fist show that the recursions in and are identical . to do this
it is sufficient to observe that @xmath99 and @xmath100 .
the reason for this is the alternating colors of segments in each row that always begin with a red segment and end with a blue segment .
this reduces us to verify the identity in the basic case where @xmath101 .
the only @xmath2-dyck path with area @xmath9 is the path @xmath102 that remains as close as possible to the main diagonal .
thus the ranks of the south ends of such a path are a rearrangement of @xmath103 .
this forces the image @xmath6 of @xmath102 to start with @xmath1 north steps and end with @xmath0 east steps .
this is the path of maximum area .
it remains to prove that @xmath102 has maximum dinv , or equivalently every cell above @xmath102 contributes to its dinv . by contradiction , suppose that for a pair @xmath76 , @xmath77 does not sweep @xmath78 .
we have the following two cases .
see figure [ fig : basecase ] .
\(1 ) @xmath78 starts at a level above @xmath77 .
assume we have @xmath104 for some @xmath105 .
then consider the path obtained from @xmath4 by changing this to @xmath106 .
this is a new dyck path with area one less than that of @xmath4 .
a contradiction !
\(2 ) @xmath78 ends below the level of @xmath77 , then assume we have @xmath107 for some @xmath105 .
then consider the path obtained from @xmath4 by changing this to @xmath108 .
this is a new dyck path with area one less than that of @xmath4 .
a contradiction !
we terminate our presentation by a few comments .
to begin we should note that our argument does not use proposition [ p-1.1 ] .
we have nevertheless included it in this writing for two reasons .
firstly because it is too surprising a result to leave out , but more importantly , because it gives a simple proof of the not trivial result that the sweep image of a dyck path is also a dyck path . the reason for this is that it is implicit in the conclusion of the proposition that the starting ranks of all the arrows of the image are non - negative .
the latter is the only property needed to guarantee that the image of an @xmath2-dyck path is an @xmath2-dyck path .
finally we remark that the dinv sweeps to area " result was known to hold in the non - coprime case , but no proof appeared in the literature .
( gregory warrington told us their proof in @xcite can be extended for this case . )
our proof naturally extends for the non - coprime case as we explain below .
let @xmath2 be coprime and let @xmath64 be a positive integer . we are considering lattice paths in the @xmath109 rectangle .
the rank of lattice points are defined as before , but now different points may have the same rank .
the @xmath5 statistic extends naturally , but we must be careful about the ties for sweep map and the @xmath3 statistic .
the dinv statistic we are using is the same as the @xmath110 statistic in ( * ? ? ? * lemma 11 ) : a cell @xmath12 above an @xmath111-dyck path @xmath4 contributes a unit to its dinv statistic if and only if the starting rank @xmath13 of the east step of @xmath4 below @xmath12 and the starting rank @xmath14 of the north step of @xmath4 to the right of @xmath12 satisfy the inequality @xmath112 , which is equivalent to @xmath113 . in our stretched form
, a pair @xmath76 contributes a dinv if @xmath77 sweeps @xmath78 when moving to the right , including the case they start at the same level but excluding the case they end at the same level .
an alternative description is to move @xmath77 to the right along a line of slope @xmath114 for sufficiently small @xmath115 . with this description and
redo the computation we will see that proposition [ p-2.3 ] holds for @xmath111-dyck paths , except that we need to move region @xmath69 one level up and region @xmath70 one level down . for the corresponding sweep map ,
we only describe using the stretched form of dyck paths .
then the sweep order is from bottom to top and at each level from right to left .
this corresponds to sweep from bottom to top , but with a line of slope @xmath114 for sufficiently small @xmath115 .
if we redo the computation with this description , we will see that theorem [ t-1.1 ] naturally extends , and proposition [ p-2.1 ] holds for @xmath111-dyck paths , except that we need to move region @xmath69 one level up and region @xmath70 one level down . | we give a new combinatorial proof of the well known result that , when @xmath0 and @xmath1 are co - prime , the dinv of an @xmath2-dyck path is equal to the area of its sweep map image .
the first proof of this remarkable identity is due to loehr and warrington in [ 4 ] .
there is also a second proof due to gorsky and mazin in [ 3 ] . |
recent experiments on dynamical fracture @xcite have shown that cracks in brittle materials become unstable above a critical velocity .
this instability is associated with the formation of a roughened fracture surface , with an increased dissipation of elastic energy , and with complicated velocity dynamics - for a review , see @xcite .
although there are hints of such an instability in the traditional continuum formulations of ideally brittle cracks @xcite , a systematic treatment does not appear possible .
indeed , recent attempts @xcite to utilize a cohesive stress modification @xcite of the continuum elastic equations in order to address this problem have been rather unsuccessful . in the absence of a compelling continuum approach
, one must deal in some manner with discrete dynamics on the atomic scale " . in this
regard , slepyan @xcite pioneered the idea of studying lattice models in which atoms interact ( via piece - wise linear springs ) with their neighbors in a predetermined lattice geometry .
this conceptual framework was further developed by marder and collaborators @xcite . here
, one can directly obtain the relationship between the crack tip velocity and the imposed driving and furthermore one can search for conditions which do not allow stable steady crack motion .
clearly , lattice models are not fully realistic , especially at displacements that are not small compared to the underlying lattice spacing ; to be more realistic , one must resort to molecular dynamics simulations including all possible atomic interactions @xcite .
however , the analytic tractability of these models as well as indications @xcite that they do contain the essential mechanism responsible for the observed dynamical instability make them well worthy of serious attention .
slepyan recognized that in an ideally brittle material ( where each spring is linear until a displacement at which it completely breaks ) one can use fourier methods to find the lattice analog of a traveling wave solution .
in this solution , each point on the lattice ( at the same transverse coordinate ) undergoes the same time history of motion as any other , albeit with some time delay .
once one obtains the solution , one must check that all springs assumed to be linear have displacements that are in fact below the breaking threshold .
the violation of this assumption by the steadily propagating solution signals the onset of more complex dynamical behavior , at least in qualitative accord to what is seen experimentally @xcite . using the wiener - hopf technique ,
slepyan solved for the velocity - driving curve in the limit where the width of the lattice transverse to the crack direction goes to infinity . one important consideration in all models of brittle fracture concerns dissipation mechanisms . from our perspective
, it makes sense to put in dissipation at the lattice scale in such a way so as not dominate the large - scale continuum elastic field .
if in addition one demands an equation that is local in time , one is led @xcite to the introduction of a kelvin viscosity which dissipates energy proportional to the rate of change of spring lengths . in the naive continuum limit , this gives rise to a third derivative ( two space , one time ) term .
that means that if the viscosity is chosen to be o(1 ) on the lattice scale , it will scale to zero as far as the macroscopic dynamics is concerned .
we saw this explicitly in a previous set of papers @xcite which solved this problem for finite width lattices and considered the nature of the solution as the width became large .
nevertheless , the viscosity can have a considerable effect on aspects of the solution which depend on the microscopic details , namely the crack speed ( for a fixed stress intensity factor ) and the self - consistency of the traveling wave ansatz .
the rest of this paper is organized as follows . in the next section ,
we formally define the model , employing slepyan s idea @xcite of replacing the driving at some external boundary with some local forcing on the crack surface .
this allows for the solution in section iii of the strictly infinite lattice model ; the connection to a finite problem with some displacement driving is made by appealing to the universality of the microscopic crack solution given a fixed stress intensity factor . in the following section
, we discuss the numerical evaluation of the velocity curve from its formal expression . following that , we turn to the aforementioned self - consistency condition and study the effect of dissipation on the critical velocity .
we conclude with some observations in the final section .
we wish to study mode iii cracks and the effects thereupon of kelvin viscosity .
our model thus consists of a square lattice of mass points undergoing ( scalar ) displacements out of the plane .
the lattice extends infinitely long in both the @xmath0 ( along - crack ) and @xmath1 ( transverse ) directions .
the lattice points are connected to their nearest neighbors by ideally brittle springs which behave linearly with spring constant 1 until some threshold elongation @xmath2 at which point they irreversibly crack .
all the ( uncracked ) springs have a viscous damping @xmath3 .
let us assume that the crack corresponds to a sequential breaking of the vertical bonds between the mass points at rows @xmath4 and @xmath5 . the equation of motion for the masses in rows @xmath6 reads @xmath7 we allow the vertical bonds along the crack surface to have a different spring constant @xmath8 and damping parameter @xmath9 , to model the effect of having weak links along the crack surface .
the equation for @xmath4 then reads , using the assumed symmetry @xmath10 , @xmath11 it is easily verified that the equations for @xmath12 are consistent with the above symmetry .
we can , without loss of generality choose @xmath13 ( so that the threshold elongation is unity ) .
note that in these units , the elastic wave speed is unity , so all velocities are dimensionless , expressed as fractions of the wave speed .
we are interested in steady - state cracks , described by the slepyan traveling wave ansatz , @xmath14 which implies that every mass point in a given row undergoes the same time history , translated in time .
we choose the origin of time such that @xmath15 , so that it represents the moment of cracking of the spring in column 0 .
if we define @xmath16 , the equations of motion for @xmath6 become @xmath17 for @xmath4 , we separate out the terms proportional to @xmath18 , giving @xmath19 the driving term for this equation has the form @xmath20 \ .\ ] ] in this expression , we have inserted an external force @xmath21 which acts on the crack boundaries .
this force is an artifice which serves as an inhomogeneous source term , in place of introducing the driving through the boundary condition on the top and bottom surfaces .
we will see later how to choose this function .
we fourier transform with respect to @xmath22 ( see the appendix for our conventions regarding fourier transforms and fourier integrals ) .
if we assume that @xmath23^{y-1 } $ ] for @xmath24 , we obtain from eqn .
( [ eomy ] ) the dispersion relationship @xmath25 this can be rewritten as @xmath26 which therefore implies that @xmath27 .
if we define @xmath28 then we can write @xmath29 note that we have must choose the sign that corresponds to a decaying solution in the @xmath1 direction , i.e. @xmath30 . if we substitute this ansatz into eqn .
( [ eom1 ] ) , we get @xmath31 where @xmath32 is the fourier transform of @xmath33 . noting that @xmath34
, we obtain @xmath35 with @xmath36 this equation is implicit , since the driving term @xmath32 on the right hand side still depends on the displacement field @xmath37 . in the next section , we solve this equation via the wiener - hopf technique .
let us define @xmath38 , the bond elongation between rows @xmath5 and @xmath4 . in the appendix , we define @xmath39 and @xmath40 as the decomposition of @xmath41 into terms which are analytic in the upper and lower half planes respectively .
this allows us to write eq .
( [ forcek ] ) as @xmath42 where we have employed a similar breakup for the driving term @xmath21 , and where we have utilized eq .
( [ derivative ] ) . substituting the formula for @xmath43 into eqn .
( [ chi ] ) for @xmath44 leads to a closed form equation for @xmath45 ; namely @xmath46 defining @xmath47 by @xmath48 we can rewrite eq .
( [ rawk ] ) as @xmath49 we will solve this equation using the wiener - hopf technique . the key is to factor @xmath47 into a product of two pieces , @xmath50 , each of which is regular in the upper ( @xmath51 ) or lower ( @xmath52 ) half plane .
doing so , we get @xmath53 we then have to deal with the extra singularity at @xmath54 .
we do this by a simple subtraction @xmath55 we have now to choose @xmath21 , our external forcing . in taking the width to infinity , in essence
we are solving for the `` inner '' solution in the sense of boundary - layer theory ; i.e. , only on the scale of the lattice spacing . the external forcing in the finite width problem only acts on the large scale , and does not vary on the lattice scale .
thus , @xmath21 must have support only at @xmath56 . following slepyan , then
, we choose @xmath21 such that @xmath57 where @xmath58 is a constant .
this then gives us the the two separate equations [ vpm ] @xmath59 as shown in eqs .
( [ limp ] ) and ( [ limm ] ) we can solve for @xmath60 from eqs .
( [ vp ] ) or ( [ vm ] ) by multiplying the first by @xmath61 or the second by @xmath62 .
this yields @xmath63 substituting this back into eqns .
( [ vpm ] ) , and defining @xmath64 , we find @xmath65 these equations directly determine the displacements given the strength of the external driving , @xmath58 . to get a physical handle on the meaning of this constant
, we note that eqns .
( [ vpm ] ) can be directly solved for the leading behavior of @xmath66 as @xmath67 goes to @xmath68 , using the fact that in this limit , @xmath69 where @xmath70 , and @xmath71 is a constant , reflecting the arbitrariness in how we perform the decomposition of @xmath47 .
we find @xmath72 this is just the expected ( fourier transform of ) the singular solution of continuum elasticity as one approaches the crack tip .
thus , as in all boundary - layer problems , the large - distance limit of the `` inner '' solution matches on to the short - distance limit of the `` outer '' solution , confirming the justification of the choice of @xmath21 above . we need to relate the strength of the short - distance singularity of the continuum solution to the driving displacement @xmath73 .
this can be done directly by solving the continuum elastic equations .
it is easier , however , to follow slepyan and derive this relation by energy considerations .
one can calculate @xcite the flux of energy into the boundary - layer , or process " zone , obtaining @xmath74 now consider the flux of energy @xmath75 being used to break bonds , which is simply related to @xmath60 @xmath76 then , using the idea that the griffith s displacement is determined by exactly the condition that all the energy flux is needed just to break the bonds , we have @xmath77 notice that the factor of @xmath71 guarantees that the result is invariant with respect to how the decomposition of @xmath47 is performed . in order to make contact with other treatments , we consider the standard case @xmath78 and @xmath79 , in which all bonds are equivalent . in this case , the function @xmath47 has a particularly simple form , @xmath80 with @xmath81 and @xmath82 . we can represent these functions in term of the products over all their roots @xmath83 where @xmath84 are the ( nonzero ) roots of @xmath85(@xmath86 ) in the lower half plane and @xmath87 are corresponding roots of @xmath85(@xmath86 ) in the upper half plane .
this decomposition allows us to write down explicitly the factors @xmath51 , @xmath52 : @xmath88 examining the small @xmath67 behavior , we find the @xmath89 , so that substitution of @xmath51 into eqn .
( [ delta ] ) gives @xmath90 if we explicitly take out of the product in eq .
( [ e1 ] ) the one imaginary @xmath91 , which we will denote as @xmath92 , we get @xmath93 this formula was obtained by @xcite , who started from a finite lattice of transverse size @xmath94 and considered the @xmath95 limit .
note that this can be done for @xmath78 on a square lattice , ( the case being considered here ) , but not for general @xmath8 or for modes i or iii on a triangular lattice ; the method here works for these cases as well @xcite .
next , it is easy to see that in the limit @xmath96 , we have @xmath97 and that therefore @xmath98 furthermore , in this limit @xmath99 .
all other complex roots cancel out , leaving only real roots .
hence , @xmath100 a result originally obtained by slepyan .
note that in our treatment , the presence of one of these real roots in either the numerator or the denominator of the product expression depends entirely on the half - plane from where it came as we took the limit @xmath96 .
to see what this condition implies , let us denote by @xmath85(or @xmath86 ) as @xmath101 ; then the small imaginary part of a specific real root , @xmath102 , satisfies the equation @xmath103 which gives as the determining factor @xmath104 this condition is exactly what appears in slepyan s work .
in this section we derive an integral form for the basic result eq .
( [ e1 ] ) .
we then present a numerical procedure for finding the displacement - velocity curve .
the basic notion is to utilize eq .
( [ log ] ) to find @xmath51 .
it is more convenient for the numerical work to isolate explicitly the singularity at @xmath56 , defining @xmath105 by @xmath106 where @xmath107 is an arbitrary positive constant . because of the factor @xmath108 in the denominator , @xmath109 ; the factor @xmath110 does not alter the desired asymptotic behavior at infinity : @xmath111 as @xmath112 . applying eq .
( [ log ] ) , we get for @xmath113 @xmath114 and for @xmath115 @xmath116 with @xmath117 where the upper sign is for @xmath113 and the lower sign for @xmath115
. there is one point regarding these expressions that must be clarified .
we must make the correct choice of the branch cuts for the square roots in @xmath118 and @xmath119 appearing in @xmath47 .
it is convenient to express things in terms of @xmath120 , the simple @xmath78 and @xmath121 , form of @xmath47 , so that from eqn . ( [ seqn ] ) @xmath122 and from eqn .
( [ xi ] ) @xmath123 we must choose @xmath124 such that @xmath30 as noted below eqn .
( [ xi ] ) .
it is easy to see from the above expression for @xmath125 that it is sufficient to choose @xmath124 to lie in the first and fourth quadrants to guarantee this . for @xmath124 ,
we can then generate @xmath47 directly via eqn .
( [ seqn1 ] ) , and from this @xmath126 . as @xmath109
, there is no pole in the integral ( [ kpm ] ) at @xmath127 when @xmath56 .
this means that @xmath128 and the asymptotic expressions for @xmath129 are @xmath130 comparing this to eqn .
( [ kappa ] ) , we have @xmath131 so that @xmath132 explicitly writing out the integrals , we get @xmath133 the extra degree of convergence of the integrand at infinity is very desirable , as it helps control the fact that @xmath134 is a rapidly oscillating function . also , the fact that the answer must be independent of @xmath107 provides a check on the numerical routines . a change of variables @xmath135 allows the integral to be rewritten as @xmath136 to actually compute this integral , we divided the region of integration into two parts from @xmath68 to @xmath137 and from @xmath137 to @xmath138 .
the second integral was further transformed so as to have in the numerator the factor @xmath139 which goes to zero at infinity ; the subtracted integral with the numerator @xmath140 was performed analytically .
after all these manipulations , the integrals were successfully computed by standard mathematical library subroutines .
results from our numerical calculations for @xmath78 and @xmath79 , are presented in figs .
[ figure1 ] , [ figure2 ] , which reproduce , of course , those of @xcite . at small @xmath3 , there are large oscillations in the @xmath141 vs @xmath73 curve , as found initially by marder and gross for lattices with small transverse sizes .
the curves all hit the @xmath142 line at the point which marks the end of the band of lattice - trapped static cracks @xcite .
for larger damping , the oscillations disappear and the moving crack branch bifurcates smoothly ( in the backward direction ) from this point , eventually turning around and becoming stable . in fig .
[ vsmallk ] , we contrast the standard result to that for smaller @xmath8 , keeping @xmath121 .
we see that the @xmath142 limit has moved closer to unity , in accord with the finding in @xcite that the window of arrested cracks narrows as @xmath8 decreases ( note for purposes of comparison that our @xmath8 is a factor of 2 smaller than that defined in @xcite .
the velocity at the minimum @xmath143 however is hardly affected , so the minimum stable velocity is essentially unchanged .
past this point , the velocity is higher at smaller @xmath8 , even when the change in @xmath144 is factored in .
this is further evidence of how the microscopic details influence the velocity versus driving displacement curve .
on dimensionless driving @xmath145 for @xmath78 and @xmath121 .
the curves are for @xmath146 , @xmath147 , @xmath148 , @xmath149 , @xmath150,width=312 ] to smooth behavior at large @xmath3 . here
@xmath78 and @xmath121.,width=312 ] on dimensionless driving @xmath143 for @xmath151 , @xmath147 , @xmath152 , @xmath153 , @xmath154 .
here @xmath155,width=312 ]
as mentioned in the introduction , cracks become unstable above a critical velocity . in the framework of the model
being considered here , one might assume that this velocity is associated with the solution as found by the wiener - hopf method becoming inconsistent .
as we shall see , there are two kinds of inconsistencies .
the first sets in for small @xmath141 , for small enough @xmath3 . here
the vertical bond between @xmath5 and @xmath4 first achieves critical extension prematurely , at some positive @xmath22 , contradicting the assumption that the bond breaks at @xmath156 . in the second kind of inconsistency ,
a horizontal bond achieves critical extension .
if the model is defined so that only the vertical bonds between @xmath5 and @xmath4 are breakable , then this does not present a problem .
if one the other hand , all the springs are chosen identical , so that the horizontal springs also break at the same critical extension as the vertical ones , then the solution is indeed inconsistent at this point . this is what occurs at large enough velocity , for all @xmath3 . to proceed
, we must calculate the bond lengths , back transforming our fourier space solutions so as to obtain the physical displacements .
we find the time dependence of the function @xmath41 . from eqs .
( [ vp ] ) , ( [ vm ] ) and ( [ v0bk ] ) , we have @xmath157 @xmath158 performing the inverse fourier transform , we get @xmath159 for @xmath160 and @xmath161 for @xmath162 . here
we have used the expressions for @xmath129 derived in the last section .
clearly , these integrals must be done numerically . to proceed
, we transform the pieces containing the factors @xmath163 by dividing both the numerator and denominator by @xmath164 . in the denominator ,
we get @xmath165 ; we have already discussed how this can be computed .
we now have integrands containing factors of the form @xmath166 the @xmath167 refers on the right hand side to the sign of the imaginary part of @xmath67 .
as we will see later on , we need to calculate this function only for positive @xmath67 .
then , the integral in the exponent can be written as @xmath168 we break up the region of integrations into three parts,@xmath169 , @xmath170 and @xmath171 . in the first interval ,
we numerically calculate the integral directly as written . for the third interval
, we proceed as discussed in the last section for a similar semi - infinite integral ; this leaves us with having to integrate numerically a function which behaves as @xmath172 near infinity . finally ,
for the integral near @xmath137 we add an integral with @xmath173 replaced by @xmath174 ; this added integral is done analytically and the resultant subtracted integral with the integrand now containing @xmath175 is done numerically . applying the described procedure
, we can evaluate the functions @xmath176 for any real positive @xmath67 .
we must now perform the final integration over the variable @xmath67 .
first , we note that the requirement of having a real displacement necessitates @xmath177 .
this allows us to change the region of integration to @xmath178 in ( [ vpos ] ) ; we will return shortly to [ vneg ] .
the trickiest part of this calculation concerns the behavior near @xmath56 . for @xmath179 , there is a @xmath180 behavior . here
the leading order term can be integrated analytically over the interval @xmath181 and then the overall function with this term subtracted can be used for a numerical integration .
this works in a straightforward manner .
we need to be more careful with the integral ( [ vneg ] ) , because here the integrand behaves near zero as @xmath182 .
to proceed , we first divide the interval of integration into three parts : @xmath183 , @xmath184 , and @xmath185 .
the integrals over @xmath183 and @xmath185 can again be combined to yield an integral twice the real part of the integrand over the latter range . for the range spanning zero , we first subtract from the integrand the leading term @xmath186 after this subtraction ,
the integrand becomes of order @xmath180 .
now , the one - sided integral converges , and thus we can again transform the range to @xmath187 .
this subtracted integral is evaluated numerically from a very small value of @xmath67 ( typically @xmath188 ) to @xmath137 . in the remaining part of the range ,
i.e. from @xmath68 to @xmath189 , we evaluate the integral analytically after approximating it by its leading small @xmath67 behavior , @xmath190 ; the value of the constant @xmath191 is determined by fitting the the behavior of the function near the point @xmath189 .
finally we need to perform analytically the integral of the subtracted piece @xmath192 . to accomplish this
, we deform the contour of integration from @xmath193 on the real axis to the curve from @xmath194 to @xmath137 going over the lower half of the unit circle .
note that the branch - cut for this function must be taken as before over negative real axis ; this choice guarantees the fact that @xmath195 . note that once all our transformations are complete , we only need values for the integrand for positive ( real ) values of @xmath67 ; this was used earlier in the technique for calculating @xmath196 . a good check of this complex numerical technique of computing the fourier transformation is applying it to the fourier transform @xmath197 for @xmath162 and @xmath198 for
@xmath160 , as in this case result must be equal to zero .
we first examine what happens in the standard case , @xmath78 and @xmath79 .
figure [ figure3 ] shows the behavior of @xmath41 for @xmath199 and various values of the velocity . as discussed above , we see the two different kinds of inconsistencies exhibited by these solutions . for example , the solution for @xmath200 is inconsistent as @xmath201 for @xmath160 .
this type of inconsistency , seen at small velocities , corresponds to the region of backward behavior on @xmath202 graph . as already claimed by marder and gross , this region is therefore unphysical . .
here @xmath78 and @xmath121,width=312 ] more interesting is the inconsistency which sets in at large velocity , and has been argued to be related to the experimentally observed microbranching .
what happens here is that the elongation of the horizontal bond , given by @xmath203 increases as a function of increasing driving . at some critical speed it reaches the value @xmath204 and the bond breaks .
figure [ figure4 ] shows the dependence of the critical speed on @xmath3 .
we note in passing that these graphs agree with what can be gotten by extrapolating a similar graph for the finite @xmath94 lattice , a graph that can be obtained using the methods of @xcite . the actual location in @xmath22 at which the horizontal break occurs at the critical velocity is shown in fig .
[ figure5 ] .
several details of our findings are worthy of note .
first , there is always a critical speed for any dissipation , but this speed gets very close to the maximal crack speed for large @xmath3 .
the value of the critical speed for the dissipation - less limit is @xmath205 ; this is above what has been seen in some experiments but recall that we are doing mode iii on a square lattice , a far cry from mode i in an amorphous system .
the is a curious break in the curve at @xmath206 , which corresponds in fig .
[ figure5 ] to the point where the break location varies most strongly with @xmath3 .
it should be noted that this inconsistency does not appear related in any obvious way to the @xcite criterion .
first , the yoffe criterion , concerning the direction of maximal stress , is calculated for the continuum elastic field , and is thus completely independent of the viscosity @xmath3 @xcite .
the onset of the inconsistency is strongly @xmath3 dependent however .
also , the yoffe criterion is stated for mode i cracks , and its applicability to mode iii is subject to question .
we turn now to the case of @xmath207 , with weakened bonds along the crack path , again keeping @xmath121 .
figures [ figure4 ] , [ figure5 ] show our results . as is reasonable ,
the inconsistencies persist , but are pushed to higher speeds as @xmath8 decreases .
it is interesting to speculate that since for finite width systems , velocities greater than the wave speed are possible , sufficiently small @xmath8 might enable stable steady - state fracture at supersonic velocities .
also , intersonic wave speeds are possible in the case of mode ii fracture , and therefore weakened bonds on the path of the crack would in this case as well allow steady - state propagation at higher speeds @xcite .
on @xmath3 for @xmath78 , @xmath208 , @xmath209 .
here @xmath121,width=312 ] on @xmath3 for @xmath78 , @xmath208 , @xmath209 . here
we have shown how to extend the slepyan approach to cracks in ideally brittle infinite lattice systems to the case of having a kelvin viscosity for each lattice spring .
in addition , we have examined the onset of the additional , off - axis cracking at high velocity .
we showed how dissipation delays but does not eliminate the microbranching instability , an effect impossible to see within the context of any continuum elastic treatment . in the following work
, we will present the results of similar calculations for mode i cracks on a triangular lattice .
that system is much closer to those that have been studied experimentally and one can therefore hope for more direct lessons to emerge . also , we should mention complementary studies on nonlinear lattice models which relax the assumption of ideally brittle springs and thereby allow one to view the onset of inconsistent behavior as the limiting case of more standard bifurcations @xcite .
dak acknowledges the support of the israel science foundation . the work of hl and lp is supported in part by the nsf ,
dmr94 - 15460 .
dak and lp thank prof .
a. chorin and the lawrence berkeley national laboratory for their hospitality during the inital phase of this work .
we discuss here some of our conventions and notations regarding fourier transformations .
we use fourier transformation in the form @xmath210 we define also @xmath211 and @xmath212 as the fourier transform of @xmath221 is given by @xmath222 finally , we consider separating an arbitrary function @xmath223 into the product of two pieces @xmath51 and @xmath52 , with poles and zeroes in lower and upper half - planes respectively . these can be done using the identity , @xmath224 abraham , f.f . ,
brodbeck , d. , rafey , r.a . ,
rudge , w.e . , 1994 .
instability dynamics of fracture - a computer - simulation investigation .
( 2 ) , 272 - 275 .
barenblatt , g.i . , 1959 . the formation of equilibrium cracks during brittle fracture : general ideas and hypothesis , axially symmetric cracks .
appl . math . and
* 23 * , 622 - 636 ( 1959 ) .
fineberg , j. , gross , s.p . ,
marder , m. , swinney , h.l .
instability in dynamic fracture . , ( 4 ) 457 - 460 .
fineberg , j. , gross , s.p . ,
marder , m. , swinney , h.l .
instability in the propagation of fast cracks . ,
( 10 ) 5146 - 5154 .
fineberg , j. , marder , m. , 1999 .
instability in dynamic fracture .
. repts . *
313 * , ( 1 - 2 ) 1 - 108 .
gumbsch , p. , zhou , s.j . ,
holian , b.l . , 1997 .
molecular dynamics investigation of dynamic crack stability .
( 6 ) , 3445 - 3455 .
holland , d. , marder , m. , 1997 .
ideal brittle fracture of silicon studied with molecular dynamics . ,
( 4 ) 746 - 749 .
holland , d. , marder , m. , 1998 .
erratum : ideal brittle fracture of silicon studied with molecular dynamics . ,
( 18 ) 4029 .
kessler , d.a . , 2000 .
steady - state cracks in viscoelastic lattice models ii . , to appear .
kessler , d.a . , levine , h. , 1998 .
steady - state cracks in viscoelastic lattice models .
( 5 ) , 5154 - 5164 .
kessler , d.a . , levine , h. , 1999 .
arrested cracks in nonlinear lattice models of brittle fracture .
( 6 ) 7569 - 7571 .
kessler , d.a . , levine , h. , 2000 .
stability of cracks in nonlinear viscoelastic lattice models . in preparation .
kulamekhtova , sh.a . ,
saraikin , v.a . ,
slepyan , l.i . , 1984 . plane problem of a crack in a lattice .
mekhanika tverdogo tela , * 19 * , ( 3 ) 112 - 118 [ mech .
solids * 19 * , ( 3 ) 102 - 108 ] .
langer , j.s . ,
1992 . models of crack propagation .
( 6 ) , 3123 - 3131 .
langer , j.s . ,
lobkovsky , a.e . , 1998
. critical examination of cohesive - zone models in the theory of dynamic fracture .
solids * 46 * ( 9 ) , 1521 - 1556 .
marder , m. , gross , s.p .
origin of crack tip instabilities .
solids * 43 * ( 1 ) , 1 - 48 .
marder , m. , liu , x. , 1993 .
instability in lattice fracture .
( 15 ) , 2417 - 2420 .
omeltchenko , a. , yu , j. , kalia , r.k . , vashishta , p. , 1997 .
crack front propagation and fracture in a graphite sheet : a molecular - dynamics study on parallel computers .
, * 78 * ( 11 ) , 2148 - 2151 .
pechenik , l. , levine , h. , kessler , d.a . ,
steady - state mode
i cracks in a viscoelastic triangular lattice . submitted for publication pla , o. , guinea , f. , louis ,
e.,ghasias , s.v . , sander , l.m . , 1998
. viscous effects in brittle fracture .
( 22 ) , r13981-r13984 .
rosakis , a.j . ,
samudrala , o. , coker , d. , 1999 .
cracks faster than the shear wave speed .
science * 284 * ( 5418 ) , 1337 - 40 .
sander , l.m . ,
ghasias , s.v . , 1999 .
thermal noise and the branching threshold in brittle fracture . ,
( 10 ) 1994 - 1997 .
sharon , e. , gross , s.p . and fineberg , j. , 1996 .
local crack branching as a mechanism for instability in dynamic fracture .
( 25 ) , 5096 - 5099 .
sharon , e. , gross , s.p . and fineberg , j. , 1996 .
energy dissipation in dynamic fracture .
( 12 ) , 2117 - 2120 .
slepyan , l.i . ,
dynamics of a crack in a lattice .
doklady akademii nauk sssr * 258 * ( 1 - 3 ) 561 - 564 . [ sov .
* 26 * ( 5 ) , 538 - 540 ] .
slepyan , l.i . ,
1982.the relation between the solutions of mixed dynamical problems for a continuous elastic medium and a lattice .
doklady akademii nauk sssr , * 266 * ( 1 - 3 ) 581 - 584 .
* 27 * ( 9 ) , 771 - 772 .
yoffe , e.y . , 1951 .
philos . mag .
* 42 * , 739 - 750 .
zhou , s.j . ,
beazley , d.m .
, lomdahl , p.s . ,
holian , b.l . , 1997 .
large - scale molecular dynamics simulations of three - dimensional ductile failure .
( 3 ) , ( 479 - 482 ) . | we extend the slepyan solution of the problem of a steady - state crack in an infinite ideally brittle lattice model to include dissipation in the form of kelvin viscosity . as a demonstration of this technique , based on the wiener - hopf method ,
we apply the method to mode iii cracks in a square lattice .
we use this solution to find the critical velocity at which the steady - state solution becomes inconsistent due to additional bond - breaking ; this point signaling the onset of complex dynamical behavior.keywords : crack propagation and arrest , dynamic fracture , crack branching and bifurcation |
heavy - to - light decay processes become more and more important nowadays with copious data from the @xmath0 factories .
especially , @xmath1 decays provide information about the poorly known cabibbo - kobayashi - maskawa ( ckm ) element @xmath2 . in a theoretical viewpoint ,
heavy - to - light transition shows interesting kinematical configurations which provoke new effective theories of strong interaction .
soft - collinear effective theory ( scet ) is a useful framework to deal with light and energetic particles @xcite .
but nonperturbative quantities are always involved in the analysis .
the way how to deal with them is crucial and distinctively differs in various approaches . among the nonperturbative calculational methods , light - cone sum rule ( lcsr )
has been very successful @xcite .
compared to the traditional sum rule by shifman , vainshtein , and zakharov ( svz ) @xcite , lcsr is more adequate for the heavy - to - light decays @xcite . in the svz sum rules ,
nonperturbative effects are encoded by the so called vacuum condensates .
but it is argued that the condensates exaggerate the end - point behavior of the final state meson distribution amplitudes ( da ) in the heavy - to - light decays @xcite .
the lcsr is based on an expansion of nonlocal operators in twist whose matrix elements between vacuum and the meson define the meson distribution amplitudes .
one of the most representative nonperturbative quantity for the heavy - to - light transition is the @xmath3 form factor , @xmath4 .
it was already calculated by lcsr in @xcite .
a new development of qcd factorization ( qcdf ) @xcite enlarged our understanding of various @xmath0 decays , and the spectator interactions are systematically examined in @xcite .
the advent of scet enabled us to analyze @xmath4 more profoundly @xcite . what makes the analysis more complicated is the possible existence of the feynman mechanism , or soft overlap .
in this situation , some of the partons carry very small momentum compared to others to make the final state meson .
it is quite controversial how large the soft overlap is , or even how to _ define _ this soft contribution @xcite .
for example , @xmath5 in @xcite includes only collinear quarks for the final pion , implicitly neglecting the soft - overlap configuration .
that is the reason why the @xmath5 is power - counted from @xmath6 , where @xmath7 @xmath8 is a large ( hadronic ) scale . on the other hand ,
the soft form factor @xmath9 in @xcite includes a hard - collinear spectator interactions as its @xmath6 corrections which are not momentum asymmetric . and
the authors of @xcite define the
universal form factor @xmath10 in terms of soft - collinear messenger modes to describe the soft overlap . a clearer and more definite construction will be needed to remove any confusions .
we will use the terminology of spin - symmetric and spin - nonsymmetric contributions to the form factors @xcite .
recently , there was a try to establish the lcsr in scet to calculate the heavy - to - light form factors @xcite .
actually , there have been many efforts to construct the lcsr in the effective theory , like hqet @xcite . in a conventional method ,
one constructs the correlation function typically with the interpolating current for the initial state and the weak current .
after possible contractions , the remaining fields between the vacuum and the final state define the distribution amplitudes .
the authors of @xcite claim that when the momentum configuration of the final partons is highly asymmetric , then the light - cone expansion of the remaining fields is not guaranteed .
they propose an alternative method where the final state pion is described by the interpolating fields .
but as we will see later , new method for lcsr does not show a fundamental difference from the conventional one .
rather , the two methods are _ equivalent _ under proper conditions . in this paper ,
a natural extension of the conventional lcsr in the effective theory is pursued .
in other words , we describe the initial @xmath0 meson in terms of the usual interpolating fields .
we reproduce the previous results for the form factors of lcsr . + as for the soft form factors , they will be still suffering from the criticism from @xcite .
but we argue that the defects are not fully overcome by a new method also .
still the essential point is how to combine the momentum - asymmetric partons into the final state pion . in spite of this difficulty
, we would like to show that lcsr in scet can be well established , especially when the collinear gluon is exchanged with the spectator quark , hoping a deeper understanding about scet , much like the traditional svz sum rule in the hqet .
the paper is organized as follows . in the next section , the basic formalism of scet and lcsr
is summarized .
sum rule calculations are given in sec .
we distinguish spin - symmetric and spin - nonsymmetric contributions to the form factor , and reproduce the well known results .
discussions and conclusions appear in sec .
effective fields in scet include collinear quark fields @xmath11 , ultrasoft(usoft ) heavy quark fields @xmath12 , usoft light quark fields @xmath13 , collinear gluon fields @xmath14 , and usoft gluon fields @xmath15 .
these are the relevant degrees of freedom for the usoft - collinear interactions .
collinear covariant derivatives can be defined as @xmath16 where @xmath17 is the usual label operator @xcite , and usoft covariant derivatives are @xmath18 the @xmath19-components of the derivative appear as @xcite @xmath20 we begin with the scet lagrangians for the usoft - collinear interactions @xcite : @xmath21 where the field strength operators are defined as @xmath22~,\nonumber\\ ign\cdot m&=&\left[i\nbar\cdot d^c,~in\cdot d\right]~,\end{aligned}\ ] ] and @xmath23 is the collinear wilson line , @xmath24~.\ ] ] the superscripts on @xmath25 are the power suppression in @xmath26 .
heavy - to - light currents are constructed in scet by matching from full qcd . up to the next - to - leading order ( nlo ) in @xmath27 at tree level
, we have @xmath28 the essence of lcsr is to calculate the correlation function @xmath29|0\rangle~. \label{pi}\ ] ] the heavy - to - light current @xmath30 and the @xmath0-meson interpolating field @xmath31 are @xmath32 using the effective fields , the correlator @xmath33 can be written in the hadronic language as @xmath34 where @xmath35 , @xmath36 , and @xmath37 .
@xmath0-meson effective decay constant @xmath38 is defined by @xmath39 by the dispersion relation and quark - hadron duality , the lcsr is established as @xmath40 where @xmath41 , and the borel transformed version is @xmath42 where @xmath43 is the borel parameter .
note that we use @xmath44 as a dynamical variable rather than the usual @xmath45 , so the lower limit of the dispersion integral is @xmath46 which corresponds to @xmath47 when @xmath48 .
the matrix elements @xmath49 are proportional to the effective form factors . in the next section ,
we evaluate @xmath50 with the quark - gluon fields for various @xmath51 to complete the sum rule .
spin - symmetric contributions come from the operators which preserve the spin - symmetric form factor relations @xcite .
after the field redefinition of @xmath52 , @xmath53 where @xmath54 is the soft wilson line , and scaling down to the @xmath55 , their matrix elements are parameterized as @xcite @xmath56 the leading order diagram is shown in fig .
[ pis ] .
diagram for the leading order spin - symmetric form factor.,height=188 ] sum rules for @xmath57 can be obtained via the correlation function @xmath58 : @xmath59|0\rangle~.\ ] ] using the heavy quark propagator @xmath60 we have @xmath61 by expanding @xmath62 and @xmath63 into their light - cone components , @xmath64 here the terms containing @xmath65 do not contribute , since in the phenomenological sector the matrix elements of the currents are proportional to @xmath66 or @xmath67 which has no perpendicular components . in the full theory , the matrix elements of nonlocal operators between vacuum and meson are described by the das . up to the twist 3 , @xmath68 where @xmath69 .
one delicate point at this stage is how to describe the final state pion . in the literature ,
the energetic pion is constructed by two collinear quark fields in the effective theory .
but this picture does not fully appreciate the soft contributions where the constituents momenta are asymmetric . for the leading order @xmath57
, we try to set the relevant matrix element as @xmath70~ , \label{softda}\ ] ] and consequently , @xmath71 if the collinear quark @xmath72 and the soft quark @xmath73 scale like @xmath74 then the combined momentum has a large virtuality @xmath75 .
this is not desirable to form a pion in the final state . to reduce the large virtuality
, it is assumed that the soft quark @xmath73 participating in the final pion scales as @xcite @xmath76 the imaginary part of @xmath58 is proportional to a delta function , which restricts the range of @xmath77 after integrating over @xmath78 in the borel improved sum rule eq .
( [ bt ] ) ; @xmath79 where we used the fact that @xmath80 fluctuates around @xmath81 with the amount of @xmath82 and thus its maximum value @xmath83 is roughly @xmath84 .
the constraint of ( [ u0 ] ) ensures the collinearity of a parton from the weak vertex , @xmath85 , and the softness of the spectator quark @xmath86 in ( [ softda ] ) @xcite .
the final result for @xmath57 is @xmath87 ~e^{(2\lmdbar-\ubar\eta)/t}~,\ ] ] where @xmath43 is the borel parameter . in the so called local duality limit where @xmath88 ,
@xmath89 this is exactly the full theory lcsr result @xcite . to see this , note that @xmath90 in @xcite is given by @xmath91
it is quite interesting that eq .
( [ zetas ] ) is also consistent with the result of @xcite . in @xcite , the soft form factor @xmath9
is given by @xmath92 then the wandzura - wilczek ( ww ) approximation , @xmath93 and the sum rule results @xmath94 are used for their analysis .
but if we combine the sum rule result for @xmath95 @xcite , @xmath96 we arrive at @xmath97 this is nothing but the result of ( [ zetas ] ) when @xmath98 . at nlo of @xmath99 , @xmath57 includes spectator interactions with collinear gluon exchanges .
the spectator quark hit by the collinear gluon becomes collinear quark at this order , so the momentum configuration of the partons can be symmetric .
figure [ pisym ] shows some of them . [ cols="^,^ " , ] as an illustration , we construct the correlation functions for these diagrams as @xmath100|0\rangle~ , \nonumber\\ % \pi_{sym}^{(n)}&=&i\int d^4x~e^{-ip_b\cdot x}\langle\pi(p)|t\left [ j^{(0)}(0)j_b^\dagger(x ) , \int d^4z~\call_{\xi\xi}^{(0)}(z ) , \int d^4y~\call_{\xi q}^{(n)}(y)\right]0\rangle~. \label{symnlo}\end{aligned}\ ] ] note the presence of the vertex @xmath25 which converts the usoft spectator into a collinear quark . with the dimensional regularization where the dimension @xmath101 , we have @xmath102&= & \frac{g^2c_f}{16\pi^2}~f_\pi n^\mu\int_0 ^ 1 du ~\frac{\phi_\pi(u)}{\ubar}\nonumber\\ & & \times\left [ -\frac{1}{\epsilon}~\frac{r}{\ubar - r } + \frac{2r}{\ubar - r}~\ln\left(\frac{r\eta}{\mu}\right ) -\frac{\ubar}{\ubar - r}~\ln\left(\frac{\ubar}{\ubar - r}\right)\right]~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{sym}^{(1)}\right]&= & \frac{g^2c_f}{16\pi^2}~f_\pi n^\mu\int_0 ^ 1 du ~\frac{\phi_\pi(u)}{\ubar}\nonumber\\ & & \times u\left [ \frac{1}{\epsilon}~\frac{r}{\ubar - r } -\frac{2r}{\ubar - r}~\ln\left(\frac{r\eta}{\mu}\right ) + \frac{\ubar}{\ubar - r}~\ln\left(\frac{\ubar}{\ubar - r}\right ) \right]~,\nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{sym}^{(2a)}\right]&= & \frac{g^2c_f}{16\pi^2}~f_\pi n^\mu\int_0 ^ 1 du ~\frac{\phi_\pi(u)}{\ubar}\nonumber\\ & & \times\left [ -\frac{1}{\epsilon}~\frac{r(1-r)}{\ubar - r } + \frac{2r(1-r)}{\ubar - r}~\ln\left(\frac{r\eta}{\mu}\right)-r -\frac{u\ubar}{\ubar - r}~\ln\left(\frac{\ubar}{\ubar - r}\right ) \right]~,\nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{sym}^{(2b)}\right]&= & \frac{g^2c_f}{16\pi^2}~f_\pi n^\mu\int_0 ^ 1 du ~\frac{\phi_\pi(u)}{\ubar } \left[-\frac{1}{\epsilon}~\frac{ur^2}{(\ubar - r)^2 } + \frac{2ur^2}{(\ubar - r)^2}~\ln\left(\frac{r\eta}{\mu}\right)\right.\nonumber\\ & & \left .
-\frac{ru\ubar}{(\ubar - r)^2 } + \frac{\ubar}{(\ubar - r)^2}(-2ru+\ubar+\ubar^2)~\ln\left(\frac{\ubar}{\ubar - r}\right ) \right]~,\end{aligned}\ ] ] where @xmath103 . since @xmath80 describes a small fluctuation of order @xmath104 around @xmath105 , @xmath106 when @xmath107 .
summing up these terms gives the finite part as @xmath108 \nonumber\\ & = & \frac{g^2c_f}{16\pi^2}~f_\pi n^\mu\int_0 ^ 1 du ~\frac{\phi_\pi(u)}{\ubar}\left [ -r + \frac{2r}{\ubar - r}\left\{\ubar+(1-r)+\frac{ur}{\ubar - r}\right\ } \ln\left(\frac{r\eta}{\mu}\right)\right.\nonumber\\ & & \left .
-\frac{ru\ubar}{(\ubar - r)^2 } + \frac{\ubar}{(\ubar - r)^2}\left\{\ubar^2+r(1 - 2u)\right\ } \ln\left(\frac{\ubar}{\ubar - r}\right)\right]~.\end{aligned}\ ] ] in the limit of @xmath109 , @xmath110~.\ ] ] the light - cone sum rule , eq .
( [ bt ] ) , is now established as @xmath111~,\ ] ] when @xmath88 .
this result must be compared with eq .
( 14 ) of @xcite .
terms which are not proportional to @xmath112 are successfully reproduced .
the matrix elements of power suppressed currents @xmath113 are proportional to the spin non - symmetric form factors .
nonzero contributions involve a collinear gluon exchange between the weak vertex and the usoft spectator which becomes collinear after the interaction .
thus the correlation function which gives sum rules for the factorizable form factors can be written as ( @xmath114 ) @xmath115|0\rangle~. \label{pimn}\ ] ] figure [ pif ] shows the @xmath116 and kinematics .
diagram for the spin non - symmetric form factors where @xmath114.,height=188 ] using the feynman rules for @xmath51 and @xmath117 from @xcite , we have @xmath118&= & \frac{g^2 c_f}{16\pi^2}f_\pi\nbar^\mu\int_0 ^ 1 du~\frac{\phi_\pi(u)}{\ubar } \ln\frac{1}{1-r}~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1a,2a)}\right]&=&0~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1a,2b)}\right]&= & \frac{g^2 c_f}{16\pi^2}f_\pi\nbar^\mu\int_0 ^ 1 du~\frac{\phi_\pi(u)}{u\ubar } \left(\ln\frac{1}{1-r}-\ubar\ln\frac{\ubar}{\ubar - r}\right)~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1b,1)}\right]&= & \frac{g^2 c_f}{16\pi^2}f_\pi n^\mu\int_0 ^ 1 du~\frac{\phi_\pi(u)}{\ubar } ~\frac{r\eta}{m_b}~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1b,2a)}\right]&= & \frac{g^2 c_f}{16\pi^2}f_\pi n^\mu\int_0 ^ 1 du~\frac{\phi_\pi(u)}{\ubar } ~\frac{\eta}{m_b}\left(r-\ubar\ln\frac{\ubar}{\ubar - r}\right)~ , \nonumber\\ % % % \frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1b,2b)}\right]&= & -\frac{1}{\pi}{\rm im}\left[\pi_{ns}^{(1b,2a)}\right]~. \label{results}\end{aligned}\ ] ] it is easy to check that the above results reproduce the previous calculations for the factorizable form factors by full theory lcsr and qcdf . to see this ,
first consider the sum rule for @xmath119 .
since @xmath120=-{\rm im}\left[\pi_{ns}^{(1b,2a)}\right]$ ] , we have @xmath121 e^{-s / t}~ds\nonumber\\ & = & \frac{g^2c_f}{16\pi^2}f_\pi n^\mu\int_{m_b-\lmdbar}^{s_0}~ds~e^{-s / t } \int_0 ^ 1~du~\frac{\phi_\pi(u)}{\ubar}~\left(\frac{s+\eta-2m_b}{m_b}\right)~.\end{aligned}\ ] ] in case of @xmath88 and @xmath122 , the right - hand - side of sum rule becomes @xmath123 where @xmath124 on the other hand , the left - hand - side is @xmath125 the matrix element is proportional to the factorizable form factor @xmath126 in qcdf @xcite : @xmath127 after contracting @xmath128 on both sides , combining eqs .
( [ rhs ] ) , ( [ lhs ] ) , and ( [ deltaf ] ) provides the sum rule for @xmath126 : @xmath129 now that the well - known qcdf result for @xmath126 is @xmath130 eq . ( [ sr1 ] ) is consistent with qcdf provided that @xmath131 this is indeed the case as already mentioned in eq .
( [ lambdab ] ) .
sum rules for @xmath132 also give the same result for @xmath126 . since @xmath133 is a small quantity in eq .
( [ results ] ) , @xmath134\sim\ln\frac{1}{1-r } = r+\calo(r^2)\approx{\rm im}\left[\pi_{ns}^{(1b,1)}\right]~.\ ] ] as for @xmath135 $ ] , we can estimate its size after integrating over @xmath77 with the asymptotic form of @xmath136 , @xmath137 : @xmath138\sim \int_0 ^ 1~du~\frac{\phi_\pi^{\rm asy}}{u\ubar}\left ( \ln\frac{1}{1-r}-\ubar\ln\frac{\ubar}{\ubar - r}\right ) = \calo(r^2)~.\ ] ]
in the present analysis , @xmath44 describes a small fluctuation of order @xmath104 around @xmath105 ; @xmath139 so @xmath140 .
this parameterization is consistent with @xmath90 in @xcite when @xmath141 . for a numerical analysis one
usually treats @xmath83 ( or @xmath90 ) as a free parameter with some reasonable constraints , but the numerics will not be considered here . as shown in the previous section ,
the sum rule result from @xcite for the soft form factor is coincident with that of this work or the previous conventional lcsr .
the reason is that when the final state pion is described by the interpolating current , only terms like @xmath142 are contributing to the soft form factor ( or , the leading spin - symmetric form factor ) @xmath9 . here
the soft field @xmath86 should exist to define the @xmath0 meson distribution amplitude together with the heavy quark field @xmath12 from the weak current . in this picture , the presence of both @xmath143 and @xmath73 at the point @xmath144 is _ a priori _ , without any dynamical explanations .
the original problem of how to form an energetic pion in the final state with one soft and one collinear quark remains in principle unresolved .
descriptions of eqs .
( [ softda ] ) and ( [ jpi ] ) are _ equivalent _ in the sense that the pion is depicted by highly momentum - asymmetric configuration with the collinear field @xmath145 and the soft field @xmath73 .
the potential problem of large nonlocality in eq .
( [ softda ] ) is alleviated by requiring an accidental smallness of the plus component of @xmath73 , @xmath146 @xcite .
we classify the collinear gluon exchange diagrams for the spin - symmetric currents as the nlo correction to the spin - symmetric form factor .
note that the operators appearing in eq .
( [ symnlo ] ) are similar to the non - factorizable operators of @xcite .
but the non - factorizable operators contain the soft gluons in a nontrivial way , which makes it difficult to factorize their matrix elements .
in general , the non - factorizable soft gluon effects can include the three - particle distribution amplitudes of the form @xcite @xmath147 where @xmath148 is the gluon field strength .
the importance of three - particle das was already pointed out in @xcite .
they introduce additional four distribution functions of twist 4 .
thus it would be quite interesting to investigate the three - particle das in the full theory and in scet and compare the results . the analysis will check if the operator set of @xcite is complete at this accuracy .
but as for the predictive power , there will be few improvements because three - particle da analysis trades single nonperturbative parameter @xmath5 for four more das .
it must be emphasized that the nlo spin - symmetric contributions contain potential end - point singularity terms @xmath149 . in the full theory ,
these terms are combined with terms @xmath150 to regularize the end - point divergence @xcite .
we expect a similar situation in the effective theory .
terms like @xmath150 would appear for the diagrams of spectator interactions with soft gluons .
since the feynman rules of the soft sector are the same as those in the full qcd , momentum - asymmetric features are exactly reproduced in the effective theory . and also
, this should happen to satisfy the evolution equation of @xmath151 in the full theory @xcite .
spin - nonsymmetric contributions are exactly the
factorizable
part of @xcite , or @xmath126 in @xcite . in scet , both @xmath0 and @xmath152 are described by the corresponding das in a convoluted manner : @xmath153 where @xmath154 is the jet function .
but in lcsr , the initial @xmath0 meson is described by an interpolating current .
thus a comparison between lcsr and scet will provide some relation for @xmath155 , or the moment of @xmath155 , more exactly . actually , the relation is no other than eq .
( [ lambdab2 ] ) .
it is remarkable that @xmath156 from this work is coincident with that from lcsr in @xmath157 @xcite . on the other hand ,
a new approach of @xcite takes the pion to be described by an interpolating current .
consequently , a comparison between @xcite and scet gives a relation for the moment of @xmath158 , @xmath159 . in this sense ,
the present work and @xcite is complementary . for the nlo spin - symmetric or spin - nonsymmetric contributions where the collinear gluon is exchanged , the main motivations of new sum rule of @xcite become weak since the final quarks are all collinear .
further , what we are mainly concerned about heavy - to - light decay is _ how to _ form an energetic light meson in the final state . in the new approach , the final state meson is described by a local interpolating current , leaving other information in the @xmath0 meson da .
another merit of the conventional lcsr over the new one is that the variety of @xmath0 decays can be easily encapsulated by the final - state mesons das . as for the numerical size of each contributions
, present analysis provides nothing new .
relative size of each component of the form factor is still very disputable @xcite , and more efforts should be made in this direction . in summary ,
heavy - to - light decay form factor is reexamined in the framework of lcsr in scet . in this work , unlike a recent approach in @xcite , conventional lcsr is naturally extended for an effective theory .
establishments of the sum rules in the effective theory is useful in that different kinematic configurations are separated from the beginning at the operator level . after assembling all the pieces
, we successfully reproduce the full theory results for the form factor .
there still remain complicated and delicate problems of how to describe the feynman mechanism clearly with operators in the effective theory , or how large each contributions is . 9 c.w .
bauer , s. fleming and m.e .
luke , phys .
d * 63 * , 014006 ( 2001 ) ; c.w .
bauer , s. fleming , d. pirjol and i.w .
stewart , phys .
d * 63 * , 114020 ( 2001 ) ; c.w .
bauer and i.w .
stewart , phys .
b * 516 * , 134 ( 2001 ) ; c.w .
bauer , d. pirjol and i.w .
stewart , phys .
d * 65 * , 054022 ( 2002 ) .
i. i. balitsky , v. m. braun and a. v. kolesnichenko , nucl .
b * 312 * , 509 ( 1989 ) ; v. m. braun and i. e. filyanov , z. phys .
c * 44 * , 157 ( 1989 ) [ sov . j. nucl . phys . *
50 * , 511.1989 yafia,50,818 ( 1989 yafia,50,818 - 830.1989 ) ] ; v. l. chernyak and i. r. zhitnitsky , nucl .
b * 345 * , 137 ( 1990 ) ; p. ball , v. m. braun and h. g. dosch , phys .
d * 44 * , 3567 ( 1991 ) . v. m. belyaev , a. khodjamirian and r. ruckl , z. phys .
c * 60 * , 349 ( 1993 ) [ arxiv : hep - ph/9305348 ] ; a. khodjamirian , r. ruckl , s. weinzierl and o. i. yakovlev , phys .
b * 410 * , 275 ( 1997 ) [ arxiv : hep - ph/9706303 ] ; a. khodjamirian and r. ruckl , adv .
ser . direct .
high energy phys . * 15 * , 345 ( 1998 ) [ arxiv : hep - ph/9801443 ] ; a. khodjamirian , r. ruckl , s. weinzierl , c. w. winhart and o. i. yakovlev , phys .
d * 62 * , 114002 ( 2000 ) [ arxiv : hep - ph/0001297 ] .
p. ball and v. m. braun , phys .
d * 55 * , 5561 ( 1997 ) [ arxiv : hep - ph/9701238 ] .
p. ball and v. m. braun , phys .
d * 58 * , 094016 ( 1998 ) [ arxiv : hep - ph/9805422 ] .
e. bagan , p. ball and v. m. braun , phys .
b * 417 * , 154 ( 1998 ) [ arxiv : hep - ph/9709243 ] .
p. ball , jhep * 9809 * , 005 ( 1998 ) [ arxiv : hep - ph/9802394 ] .
p. ball and r. zwicky , jhep * 0110 * , 019 ( 2001 ) [ arxiv : hep - ph/0110115 ] .
p. ball and e. kou , jhep * 0304 * , 029 ( 2003 ) [ arxiv : hep - ph/0301135 ] .
p. ball and r. zwicky , phys .
d * 71 * , 014015 ( 2005 ) [ arxiv : hep - ph/0406232 ] .
m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl .
b * 147 * , 385 ( 1979 ) ; nucl .
b * 147 * , 448 ( 1979 ) .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , phys .
* 83 * , 1914 ( 1999 ) [ arxiv : hep - ph/9905312 ] ; nucl .
b * 591 * , 313 ( 2000 ) [ arxiv : hep - ph/0006124 ] ; nucl .
b * 606 * , 245 ( 2001 ) [ arxiv : hep - ph/0104110 ] .
m. beneke and t. feldmann , nucl .
b * 592 * , 3 ( 2001 ) [ arxiv : hep - ph/0008255 ] .
m. beneke , a. p. chapovsky , m. diehl and t. feldmann , nucl .
b * 643 * , 431 ( 2002 ) [ arxiv : hep - ph/0206152 ] .
r. j. hill and m. neubert , nucl .
b * 657 * , 229 ( 2003 ) [ arxiv : hep - ph/0211018 ] .
m. beneke and t. feldmann , nucl .
b * 685 * , 249 ( 2004 ) [ arxiv : hep - ph/0311335 ] .
b. o. lange and m. neubert , nucl .
b * 690 * , 249 ( 2004 ) [ arxiv : hep - ph/0311345 ] . c. w. bauer , d. pirjol and i. w. stewart , phys .
d * 67 * , 071502 ( 2003 ) [ arxiv : hep - ph/0211069 ] . c. w. bauer , d. pirjol , i. z. rothstein and i. w. stewart , phys .
d * 70 * , 054015 ( 2004 ) [ arxiv : hep - ph/0401188 ] . c. w. bauer , d. pirjol , i. z. rothstein and i. w. stewart , arxiv : hep - ph/0502094 .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , arxiv : hep - ph/0411171 .
p. ball , arxiv : hep - ph/0308249 .
f. de fazio , t. feldmann and t. hurth , arxiv : hep - ph/0504088 . j. g. korner , c. liu and c. t. yan , phys . rev .
d * 66 * , 076007 ( 2002 ) [ arxiv : hep - ph/0207179 ]
. w. y. wang and y. l. wu , phys .
b * 515 * , 57 ( 2001 ) [ arxiv : hep - ph/0105154 ] ; phys .
b * 519 * , 219 ( 2001 ) [ arxiv : hep - ph/0106208 ] . d. pirjol and i. w. stewart , phys .
d * 67 * , 094005 ( 2003 ) [ erratum - ibid .
d * 69 * , 019903 ( 2004 ) ] [ arxiv : hep - ph/0211251 ] . c. w. bauer , d. pirjol and i. w. stewart , phys .
d * 68 * , 034021 ( 2003 ) [ arxiv : hep - ph/0303156 ] .
p. ball , jhep * 9901 * , 010 ( 1999 ) [ arxiv : hep - ph/9812375 ] . j. chay and j .- p .
lee , in preparation . | heavy - to - light weak form factor is calculated using the light - cone sum rule ( lcsr ) in the framework of soft - collinear effective theory ( scet ) .
there are spin - symmetric and spin - nonsymmetric contributions .
leading order spin - symmetric contribution corresponds to the soft overlap where some of the partons carry very small momentum .
the next - to - leading order spin - symmetric and spin - nonsymmetric parts are characterized by a collinear gluon exchange with the spectator quark .
we reproduce the full theory lcsr results and give comments on recent lcsr in scet . |
molecular cloud complexes ( mccs ) in the galaxy show an abundance of internal structure , both in density and velocity ( e.g. @xcite ; @xcite ; @xcite ) .
the density contrasts are non - linear independent of the importance of gravity @xcite , and the mccs exhibit non - thermal line - widths ( @xcite ; @xcite ) , generally interpreted as supersonic turbulence .
filaments seem to dominate the morphologies in ( column ) density and velocity ( e.g. @xcite ; @xcite ; @xcite ; @xcite ) .
the spatial distribution of densities and velocities is consistent with a turbulent spectrum , the details of which are a matter of debate , though @xcite . extremely puzzling
is the source of this wealth of structure .
while stellar feedback certainly is a powerful driver ( see e.g. @xcite ) , it only acts locally , and definitely only after the first stars have formed within the cloud .
moreover , stellar driving might be difficult to reconcile with the observed , nearly self - similar spatial energy distribution .
external drivers suffer from the fact that the cold dense gas essentially acts like a wall to any incoming wave ( e.g. @xcite ; @xcite ; @xcite ) , preventing an efficient energy transfer from the warm diffuse component to the cold dense phase ( see , however , @xcite for 1d models ) .
@xcite argue that openings ( holes , channels ) in a mcc serve as an inroad for alfvn - waves , however , the question of what forms these channels remains . with growing evidence that the lifetimes of molecular clouds at least in the solar neighborhood range around a few ( @xmath1 to @xmath2 ) myrs ( @xcite ; @xcite ; @xcite ) , a picture in which mccs are envisaged as transient objects in large - scale atomic colliding flows rather than as well - defined entities in a quasi - equilibrium state ( giant molecular cloud , gmc ) is emerging .
the colliding flows would accumulate atomic gas which might eventually reach column densities high enough for h@xmath0-formation at which point the `` lifetime '' of the molecular cloud would start , i.e. the accumulation time could be much longer .
some of the molecular regions might become self - gravitating and form stars instantaneously ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the large - scale flows might be driven by supernova explosions , galactic shear , or might occur in galaxy interactions ( for recent evidence of cloud collisions , see @xcite ) .
we propose and extend the idea ( see [ s : prevcurwork ] ) that the structures observed in mccs arise from the very processes that form mccs , i.e. they arise without recourse to non - linear perturbations put in by hand , an approach which defers the problem to an even earlier stage .
we investigate numerically the formation of structure in mcc precursors via colliding flows . to avoid the somewhat unwieldy `` precursor of molecular cloud complexes '' , we will abbreviate this to _ pomcloc_s . by this
we mean cold clouds ( usually identified as cold neutral medium , cnm ) , which may be fully atomic or may already contain traces of h@xmath0 . in any case
, the pomclocs serve as the initial stage of molecular cloud complexes , mccs .
our numerical models emphasize the ease with which colliding flows generate structure via a combination of dynamical and thermal instabilities . due to the thermal instability
, structure grows predominantly on small scales , leading to early non - linear density perturbations as possible seeds of gravitational collapse and star formation .
the resulting line - widths in the cold gas when seen in projection agree with observed values , however , the cold gas does not seem to exhibit internal supersonic turbulence .
only a few percent of the energy input remains available for driving turbulent motions in the cold gas , the bulk of the energy is lost due to radiative cooling . within the scope of the models , we find that the time scale for h@xmath0 formation is limited not by a minimum temperature , but by the highly unsteady environment .
this work extends our previous study @xcite , in dimensions ( here , we present corresponding 3d models ) , in resolution and in scope of physical applications .
of course , we rely heavily on earlier work ( [ s : prevcurwork ] ) .
the physics of the problem is described in [ s : physics ] , and the numerical realization as well as numerical artifacts are discussed in [ s : numerics ] and [ ss : effectbounds ] , in order to provide a background for the interpretation of the results ( [ s : results ] ) .
these are summarized in [
s : summary ] , while
[ s : wheretogo ] suggests possible routes to follow in the future .
the concept behind this study is that molecular cloud formation should be seen as a non - equilibrium process . when the perception of the ism as a dynamical medium gained acceptance , many aspects of molecular cloud and star formation theory were revisited . @xcite and @xcite re - examined the jeans criteria and found that converging velocity fields could provoke the gravitational collapse of sub - jeans mass gas clouds .
@xcite followed this line of thought , further emphasizing that diffuse gas cooling upon compression by even mild velocity and pressure disturbances could easily be transformed to high densities .
they applied this idea to externally perturbed spherical clouds .
motivated by these investigations into the role of compressive velocity fields in promoting fast cooling and gravitational collapse , @xcite used two - dimensional numerical simulations to study the cooling and fragmentation of gas compressed at the interface of two identical , oppositely directed supersonic colliding gas streams . since these early works which explored some consequences of a dynamical ism , molecular cloud formation and evolution has been studied within the framework of compressible turbulence . a strongly fluctuating medium with motions on many scales results in high density albeit transient regions which might be identified as molecular clouds .
large - scale ( kiloparsec ) models of molecular cloud formation were presented in two - dimensional simulations by @xcite and @xcite , throwing light on the role of turbulence in the cycle of cloud and star formation @xcite , and specifying conditions for cloud formation in the presence of magnetic fields .
but the numerical resolution needed to yield a detailed explanation of the structure and dynamics of molecular clouds makes it necessary to tackle the problem with smaller - scale simulations ( tens of parsecs ) .
for instance , the role played by thermal instability continues to be actively explored .
this instability alone has been shown to be an efficient mechanism for generating substructure and even driving turbulence @xcite .
as indicated by earlier work ( e.g. @xcite in an application to the solar atmosphere ) , thermal instability can be triggered by compression .
indeed one - dimensional plane - parallel simulations of transonic converging flows @xcite show how this mechanism transforms inflowing diffuse atomic gas to dense cold gas that is long - lived .
shock waves ( driven e.g. by supernova explosions , @xcite ; @xcite ) are another means to create cold , dense gas from warm , diffuse gas . propagating into the warm ism , shock waves have been shown to fragment in the presence of thermal instability @xcite and linear perturbations @xcite leading to h@xmath0-formation within a few myrs @xcite .
however the passage of a single shock might not leave in its wake enough dense , cold , cloudlets to constitute a mcc .
encounters between streams of transsonically transported gas , on the other hand might be a way to collect , compress , and cool large quantities of gas while at the same time possibly endowing the cold , dense gas with the dynamical and structural characteristics of molecular cloud complexes .
this scenario has recently come into the spotlight . in high - resolution two - dimensional simulations of supersonic , colliding gas flows ,
@xcite confirm that even in a very dynamical environment a bistable medium develops as compressions initiate thermal instability .
however their study concentrates on how turbulence generates and influences thermally unstable gas , not on the reverse problem of how thermal instability could feed the turbulence characteristic of molecular clouds . in three - dimensional simulations of transonic colliding flows ,
@xcite show that a thin cold sheet , reminiscent of those observed by @xcite and @xcite forms at the junction of the two flows .
it develops turbulence which they attribute to the non - linear thin shell instability ( ntsi , @xcite ) .
furthermore they find that even in simulations without gravity , the highest density gas is overpressured with respect to the mean warm neutral medium pressure , suggesting that the ram pressure of the colliding flows is responsible . to be fair , we should mention that favorable conditions for molecular cloud formation not only exist in converging flows , but also in `` focal planes '' of ( sustained ) mhd - waves , as @xcite shows .
however , he notes that while highly structured mccs form in such a system , the cold gas does not acquire significant turbulence .
the notion of generating the turbulent substructure of molecular clouds by their formation process has been discussed in various contexts ( e.g. @xcite ; @xcite ; @xcite ; @xcite ) .
@xcite emphasized the evolution of the thermal states in the colliding flows , and provided a semi - analytical model for the evolution of a gas parcel .
one of the main differences to the present study is that they impose velocity perturbations on their incoming gas flow as do @xcite , although as `` random noise '' at a percent - level , whereas we defer the structure generation to the actual flow collision site .
@xcite compared in detail analytical solutions with one - dimensional and three - dimensional models . @xcite and their subsequent work discuss molecular cloud formation behind a shock - compressed layer , more resembling the situation of an expanding shell . in their case ,
the initial perturbations reside in the density field .
the present study extends the `` proof of concept '' of `` structure formation from scratch '' in two dimensions presented in @xcite by focusing on the dynamical state and the turbulent properties of the pomcloc during its formation , and on the conditions necessary for the onset of h@xmath0 formation .
we restrict the models to hydrodynamics with radiative cooling , leaving out the effects of gravity and magnetic fields .
gravity would lead to further fragmentation , and magnetic fields could have a stabilizing effect ( see also discussion in [ s : wheretogo ] ) .
for this regime , then , we identify three relevant instabilities , namely the ntsi , the kelvin - helmholtz - instability ( khi ) and the thermal instability ( ti ) .
this enumeration of course does not mean that the instabilities work independently of each other .
rather , the resulting dynamical system is a consequence of a combination of all three instabilities , however , in degrees depending on the local flow properties . the ntsi ( @xcite , fig . [
f : ntsisketch ] ) arises in a shock - bounded slab , when ripples in a two - dimensional slab focus incoming shocked material and produce density fluctuations .
gas streaming along the @xmath3-direction will be deflected at perturbation peaks and collect in the troughs .
thus , @xmath3-momentum is transported laterally ( along the @xmath4-direction ) . for a weak equation of state , the slab can act as a wall to deflect the incoming gas streams , while for an adiabatic one , the bounding shocks will travel faster at the @xmath4-locations of the troughs because of the higher compression , thus leveling out the initial perturbations ( see @xcite for a discussion of the speed of the bounding shock ) .
the growth rate is @xmath5 , where @xmath6 is the sound speed , @xmath7 is the wave number along the slab , and @xmath8 is the amplitude of the spatial perturbation of the slab . various aspects of the ntsi have been investigated numerically . @xcite and @xcite focused on the combination of the ntsi and self - gravity .
the former authors found a criterion for forming self - gravitating fragments either under isothermal conditions or with line - cooling .
@xcite discussed the effect of the cooling strength on the evolution of the ntsi , stating that strong cooling leads to small - scale fragmentation which wipes out the ntsi - effects .
@xcite investigated vishniac s analysis numerically , noting that the shear flows along the slab lead to kelvin - helmholtz modes and identifying them as the main cause for the internal structure of the slab .
although in their radiative shock study , @xcite focused mainly on the interaction of stellar winds , they extended the argument and the implications to the more general interstellar medium @xcite .
@xcite studied the ntsi in the context of cloud collisions , applying the process to the formation of clumpy filaments in e.g. the orion molecular cloud .
they partially included magnetic fields , following the magnetic pressure term and the induction equation , but leaving out the tension term of the lorentz - force .
the flows deflected at the slab ( as indicated in fig .
[ f : ntsisketch ] ) can cause shear instability modes , which have been studied at great length ( e.g. @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite , to name a few ) . in the incompressible case , for a step function profile in the velocity and constant densities across the shear layer , the growth rate is given by the velocity difference @xmath9 , i.e. all wave modes get unstable .
if aligned with the flow , magnetic fields can stabilize against the khi via the tension component of the lorentz force , and the growth rate is given by @xmath10 , i.e the modes that will get unstable are those for which the flow velocity difference is larger than the alfvn speed @xmath11 . for compressible gas
, the situation changes substantially : the system will be stable for all those wave numbers whose effective mach number is larger than a critical value @xcite .
thus , while kelvin - helmholtz modes definitely _ will _ play a role in the suggested scenario , figure [ f : ntsisketch ] certainly can give only a very simplified picture .
the ti @xcite rests on the balancing of heating and cooling processes in the ism .
the ti develops an isobaric condensation mode and an acoustic mode , which under ism - conditions is mostly damped .
the condensation mode s growth rate is independent of the wave length , however , since it is an isobaric mode , smaller perturbations will grow first @xcite .
heat conduction sets the smallest growth scale .
if this scale is not resolved numerically , perturbations will grow at the resolution scale @xcite .
the signatures of the ti are fragmentation and clumping .
they persist as long as the sound crossing time across a density perturbation is smaller than the cooling time scale @xmath12 the ti can drive turbulence in an otherwise quiescent medium , even continuously , if an episodic heating source is available @xcite .
all three instabilities grow fastest or at least first on the smallest scales .
this poses a dire challenge for the numerical method .
we chose a method based on the 2nd order bhatnagar - gross - krook formalism ( @xcite ; @xcite ; @xcite ; @xcite ) , allowing control of viscosity and heat conduction .
the code evolves the navier - stokes equations in their conservative form to second order in time and space .
the hydrodynamical quantities are updated in time unsplit form .
as shown below , statistical properties of the models are resolved with respect to grid resolution , viscosity and heat conduction , although the flow patterns change in detail as is expected in a turbulent environment .
the heating and cooling rates are restricted to optically thin atomic lines following @xcite , so that we are able to study the precursors of mccs up to the point when they could form h@xmath0 .
dust extinction becomes important above column densities of @xmath13@xmath14 , which are reached only in the densest regions modeled .
thus , we use the unattenuated uv radiation field for grain heating @xcite , expecting substantial uncertainties in cooling rates only for the densest regions . the ionization degree is derived from a balance between ionization by cosmic rays and recombination , assuming that ly @xmath15 photons are directly reabsorbed .
numerically , heating and cooling is implemented iteratively as a source term for the internal energy @xmath16 of the form @xmath17 in units of energy per volume per second . here
, @xmath18 is the heating contribution ( mainly photo - electric heating from grains ) , @xmath19 the cooling contribution ( mainly due to the cii hfs line at @xmath20 m ) .
since the cooling and heating prescription has to be added outside the flux computations , it lowers the time order of the scheme . to speed up the calculations , equation ( [ e : cooling ] )
is tabulated on a @xmath21 grid in density and temperature .
for each cell and iteration , the actual energy change is then bilinearly interpolated from this grid .
two opposing , uniform , identical flows in the @xmath3-@xmath4 computational plane initially collide head - on at a sinusoidal interface with wave number @xmath22 ( and @xmath23 for 3d models ) and amplitude @xmath8 ( fig .
[ f : ntsisketch ] ) .
the incoming flows are in thermal equilibrium .
the system is thermally unstable for densities @xmath24@xmath25 .
the cooling curve covers a density range of @xmath26 @xmath25 and a temperature range of @xmath27 k. the box side length is @xmath28pc .
thus , coriolis forces from galactic rotation are negligible , however , the simulation domain is large enough to accommodate a moderately - sized molecular cloud complex . for an interface with @xmath29 , a cold high - density slab devoid of inner structure forms .
the onset of the dynamical instabilities thus can be controlled by varying the amplitude of the interface perturbation .
this allows us to test turbulence generation under minimally favorable conditions .
this setup might seem artificial : ( 1 ) the incoming flows are not expected to be perfectly uniform , however , we chose to defer the moment of structure generation to the last possible moment in the simulation , instead of imposing perturbations on the incoming flows ( see [ ss : asymmetries ] ) .
( 2 ) as we will see below , the model run times extend considerably beyond @xmath30myrs . at an inflow speed of @xmath30
km s@xmath31 this would correspond to a total extent of the system of @xmath32pc .
this suggests that the initial densities of the flows are more likely to be a few @xmath25 to form molecular clouds with flows of order @xmath33-@xmath34 pc length .
spiral density waves can also produce coherent flows of the length required , at least in principle .
the boundary conditions in the transverse ( i.e. @xmath4 and @xmath35 ) directions are periodic , while in the @xmath3-direction , the boundary values are defined as uniform inflow at constant density @xmath36 and inflow speed @xmath37 .
consequently , the boundaries can not treat outgoing waves or flows .
thus , the models have to be halted once the bounding shocks reach the boundaries .
the corresponding time scales have been discussed by @xcite .
sketch of the ntsi mechanism and geometry of the initial conditions .
gas is streaming along the @xmath3-direction and colliding at the perturbed interface .
the resulting shear flows excite kh - modes . ] to facilitate the analysis of the dynamics and time history of the cold gas , the code is equipped with lagrangian tracer particles , that are advected after each flux update .
tracer particles are deployed at grid cell centers where @xmath38k and in four generations , starting at approximately @xmath39 myrs in intervals of @xmath40 myrs .
this allows us to study the effect of turbulent motions on the fluid parcels as well as the time history of ( representative ) regions of cold gas . extending the model sequence of @xcite
, we varied the parameters , dimensions and resolution of the models ( see table [ t : modparams ] ) .
c|lccccc 2b10 & @xmath41,@xmath42 & @xmath43&@xmath44 & @xmath45&@xmath43 + 2b20 & @xmath41,@xmath42 & @xmath46&@xmath47&@xmath45&@xmath43 + 2b30 & @xmath41,@xmath42 & @xmath48&@xmath49&@xmath45&@xmath43 + 2c10 & @xmath41,@xmath42 & @xmath43&@xmath50 & @xmath51&@xmath52 + 2c15 & @xmath41 & @xmath53&@xmath54&@xmath51&@xmath52 + 2c20 & @xmath41,@xmath42,@xmath55&@xmath46&@xmath56&@xmath51&@xmath52 + s / l2c20 & @xmath41 & @xmath46&@xmath56&@xmath51&@xmath52 + 2c25 & @xmath41 & @xmath40&@xmath57&@xmath51&@xmath52 + 2c30 & @xmath41,@xmath42,@xmath55&@xmath48&@xmath58&@xmath51&@xmath52 + s / l2c30 & @xmath41 & @xmath48&@xmath58&@xmath51&@xmath52 + 2c35 & @xmath41 & @xmath59&@xmath60&@xmath51&@xmath52 + 3c10 & @xmath61 & @xmath43&@xmath50 & @xmath51&@xmath52 + 3c20 & @xmath61 & @xmath46&@xmath56&@xmath51&@xmath52 + 3c30 & @xmath61 & @xmath48&@xmath58&@xmath51&@xmath52 before we discuss the simulation results and their physical implications , we assess the numerical reliability of the code for the problem at hand .
we will restrict the numerical discussion to the 2d models , since similar effects pertain to the 3d models .
figure [ f : prettypic ] shows stills of models 2c10c , 2c20c and 2c30c ( i.e. a sequence in mach numbers 1 , 2 and 3 ) taken @xmath62 myrs after flow contact time .
while model 2c10c preserves the symmetry of the initial conditions , at larger mach numbers , the structures develop strong asymmetrical features with respect to the mid - plane . since this could be cause for concern , we measured the degree of asymmetry in the models with time ( fig .
[ f : asymmetry ] ) .
asymmetry is defined in terms of the density differences as @xmath63 where the average is over all cells with @xmath64 ( and not over the whole domain in order to exclude the inflow initial conditions which are symmetric ) , and @xmath65 is the absolute value of the density difference between cells which should be symmetric across the upper and lower half of the simulation midplane .
the initial conditions are perfectly symmetric , but slight differences at the machine accuracy level in the reconstruction of hydrodynamical variables eventually lead to a difference in the cooling strength and thus cause perceptible asymmetries .
the code preserves perfect symmetry for a purely adiabatic equation of state ( i.e. without the additional heating and cooling terms ) .
after onset , the asymmetries grow linearly in time until they reach a saturation level between @xmath66 and @xmath34 .
this corresponds approximately to the temperature ( and density ) contrast between the warm and cold gas and thus to the maximum asymmetry reachable for the system .
although the asymmetries increase with mach number and to a lesser extent with resolution , they only appear well after the system has evolved .
@xcite , @xcite and @xcite chose an alternative route : they added perturbations ( in density or velocity ) to the incoming flow , thus breaking the symmetry of the initial conditions .
while the physical reason for adding perturbations to the inflow is perfectly obvious , in the present study we want to emphasize the point that even with the least possible perturbation to the flow , ( turbulent ) substructures are generated with ease . in some sense , this is an attempt to carry the argument to extremes .
conversely , one could argue that imposing perturbations on the incoming flow helps to hide the symmetry breaking due to truncation errors . to get a stronger handle on how our somewhat extreme initial conditions are affecting the results , we repeated model 2c20 twice , once with an interface perturbation mode of @xmath67 instead of @xmath22 , and in the second repetition with the @xmath67-mode overlaid on the @xmath22 mode at @xmath68% of the amplitude at @xmath22 , motivated by a turbulent cascade . in both cases ,
the system develops small - scale structures at @xmath67 , meaning that the instability grows at the smallest imposed ( @xmath67 ) scales , as long as these are larger than the field length ( see next section ) .
@xcite demonstrated with 1d models that the choice of the heat conductivity @xmath69 can strongly influence the dynamics of the system . especially they argue that a too low heat conductivity can damp turbulent motions in the cold phase and can lead to artificial fragmentation at the grid scale ( see however @xcite ) .
the critical parameter here is the field length @xcite the length scale below which heat conduction can stabilize the ti
@xmath70 where @xmath69 is the heat conductivity ( see below ) , and @xmath71 is the cooling function from equation ( [ e : cooling ] ) . with a heat conduction of @xmath72)^{1/2}$ ] @xcite ,
@xcite conclude that for ism - conditions , a linear resolution of several thousand cells is needed ( they get convergence at @xmath73 cells ) .
the heat conduction in the bgk scheme can be controlled explicitly ( see @xcite for an analysis ) by varying the kinematic viscosity , since the prandtl number in the code is @xmath74 by construction .
the choice of the viscosity is controlled by two considerations , namely ( a ) it should be large enough to prevent numerical artifacts on small scales , and ( b ) it should be small enough to leave enough dynamical range . obviously , both requirements are difficult to meet simultaneously . to establish to what degree our models are resolved , we begin by measuring the field lengths ( eq . [
[ e : field ] ] ) cell - wise , excluding all cells which belong to the inflow , since their thermal time scale @xmath75 .
figure [ f : field ] shows the volume and mass fractions of the field - unresolved cells against time for models 2c20c and 2c30c . note
that when the volume and mass fractions agree , the bulk of the non - equilibrium gas is in the cold phase ( see also fig .
[ f : massfrac ] ) . in other words ,
figure [ f : field ] and the following ones refer ( mostly ) to the cold gas phase .
thus , between @xmath30% and @xmath76% of the volume , and between @xmath30% and up to @xmath77% of the mass of the cold gas are not resolved .
this is acceptable as long as we devise a selection criterion for the subsequent analysis . ] ) in the cold gas ( other temperature regimes are not affected ) against time for runs with resolution n = 512 ( @xmath78pc).[f : field ] ] figure [ f : fieldhisto ] lets us estimate the size of the smallest structures we can resolve .
again , we distinguish between volume and mass fractions , which however in this case are close to identical .
the figure indicates that the field length for the bulk of the mass and volume lies between @xmath79 and @xmath80 pc , scales which are well resolved by all our models . for our lowest resolution 2d runs ( @xmath81 ) ,
the resolution limit is @xmath78pc .
thus , for the subsequent discussions , we will only consider structures with sizes larger than @xmath82pc for 2d runs , and @xmath83pc for 3d runs . if the field - length is not resolved , structures tend to grow on grid scales , i.e with increasing resolution
, there should be more and more small - scale structures .
this effect is clearly visible in the left panel of figure [ f : fieldselect ] , which shows a histogram of the size of cold regions ( @xmath38k ) .
the size of a cold region is defined as the geometric mean of its minimum and maximum diameter .
the inset numbers give the average number of cold regions per line of sight , i.e. normalized to resolution .
if the field length were resolved , heat conduction would lead to a cutoff at small @xmath84 . selecting for regions with @xmath85pc , the histograms ( and the average number of cold regions )
agree sufficiently to proceed with the above selection criterion .
molecular clouds consistently show non - thermal line - widths of a few km s@xmath31 ( e.g. @xcite ) that together with temperatures of @xmath86k are generally interpreted as supersonic turbulence .
the line - widths in our models are consistent with the observed values ( fig .
[ f : veldis_mach ] ) .
the broad line wings are non - gaussian .
this may be a sign of intermittency ( e.g. @xcite ) .
histogram of the `` observed '' line - of - sight velocity dispersion in the cold gas ( @xmath38k ) for a model sequence in mach numbers , @xmath87 myrs after initial flow contact . shown
are models at @xmath81 .
line - of - sight velocity dispersions are determined along the inflow direction ( @xmath3-axis , see fig .
[ f : ntsisketch ] ) , and then laterally averaged . observed velocity dispersion values are reached easily .
see text for the choice of @xmath88k as `` defining '' temperature for the pomclocs . ]
the `` observed '' linewidth is derived from the histogram of the density - weighted line - of - sight velocity dispersion in the cold gas at @xmath38k ( fig .
[ f : veldis_mach_obsint ] , filled symbols ) .
this linewidth would correspond to linewidths in the cold neutral medium as e.g. traced by hi .
since the _ internal _ line - widths of coherent cold regions ( open symbols ) range around the sound speed of the cold gas ( @xmath89 km s@xmath31 ) , the _ internal _ velocity dispersions do not reach mach numbers @xmath90 ( see also @xcite ; @xcite ) .
hence , the `` supersonic '' line - widths are a consequence of cold regions moving with respect to each other within a warmer and more diffuse medium , but not a result of internal supersonic turbulence in the cold gas which eventually would be hosting star formation @xcite .
note from figure [ f : veldis_dens_obsint ] , that this result is independent of resolution and geometry .
because of the thermal instability , to make figure [ f : veldis_mach_obsint ] and [ f : veldis_dens_obsint ] cold coherent regions are identified via a temperature threshold at @xmath38k .
figure [ f : tempthresh ] shows the temperature distribution according to volume and mass for models 2c20 and 2c30 ( at @xmath91 ) .
note that the above temperature criterion identifies the bulk of the cold gas .
the choice of @xmath88k as a `` defining '' temperature for the pomclocs is motivated by figure [ f : tempthresh ] .
although the histogram does not include gas at the inflow temperature , the volume is dominated by warm gas ( @xmath92 k ) , while the mass is dominated by cold gas ( @xmath93 k ) .
the variations of the line - of - sight velocity dispersion with time is within the error bars shown .
`` observed '' ( filled symbols ) and internal ( open symbols ) velocity dispersion in the cold gas ( @xmath38k ) against inflow speed for the same sequence of models as in figure [ f : veldis_mach ] , @xmath87 myrs after initial flow contact .
resolution is @xmath81 .
the dashed line denotes the sound speed in the cold gas .
the internal velocity dispersion reaches mach @xmath79 at most . ]
`` observed '' ( filled symbols ) and internal ( open symbols ) velocity dispersion in the cold gas ( @xmath38k ) for a selection of models against inflow speed , @xmath87 myrs after initial flow contact .
question marks in the model names are wild cards for the mach number . the dashed line denotes the sound speed in the cold gas . ]
are the line - widths actually indicating turbulent motions , or are we seeing the inflow motions of ( already ) cold gas ( e.g. @xcite ) ?
if the gas motions are truly turbulent , the average lyapunov exponents @xmath94 should be positive , indicating a growing separation between neighboring particles . in equation ( [ e : lyapunov ] ) ,
@xmath95 is the start time of particle advection , and @xmath55 are the corresponding separations of neighboring particles .
the average refers to the simulation domain .
after an initial compression phase with @xmath96 , the cold gas becomes turbulent ( fig .
[ f : lyapunov ] ) . with higher mach number
, the onset of internal turbulence in the slab is delayed , which is mirrored by the lower lyapunov exponents .
higher resolution leads to larger @xmath97 at the time of the onset of turbulence , which is defined as @xmath98 . for @xmath91 ,
the initial separation of tracer particles is smaller , so that a larger spatial range can be covered .
all exponents are positive , independent of resolution , and converge for late times . the power spectrum of the specific kinetic energy @xmath99 ( fig .
[ f : specekin ] for model 3c20 ) can be used as another diagnostic of turbulence in the models . shown
are the three linear spectra , taken along the inflow direction ( @xmath100 ) and the transversal directions ( @xmath101 ) .
the spectral index @xmath102 is consistent with the fourier transform of a step function , as to be expected since the strong decelerations along the @xmath3-direction effectively lead to a discontinuity in @xmath99 ( note that for the initial conditions , not @xmath99 but @xmath103 is discontinuous . ) .
the spectral indices in the transversal directions , @xmath104 and @xmath105 , are consistent with a kolmogorov spectrum , indicating fully developed turbulence .
the lower spectrum ( denoted by triangles and a corresponding slope @xmath106 ) shows the spectrum averaged over spheres of constant wavenumber @xmath7 .
error bars denote errors on the mean . within the errors ,
the slope is still consistent with a kolmogorov spectrum of @xmath107 in three dimensions .
the last half decade is dominated by numerical diffusion , leaving us approximately @xmath53 decades for physical analysis .
specific kinetic energy spectrum for model 3c20 at @xmath108 myrs .
the linear spectral indices @xmath109 refer to the coordinate directions , and @xmath110 gives the index of the spherically averaged spectrum . along the inflow
, the strong decelerations effectively lead to a step function in @xmath111 , mirrored in the steeper index . due to numerical dissipation ,
the last half decade is not available for physical analysis . ]
the spectral indices differ depending on whether the spectrum is taken along the inflow or transversally .
another way to see this is to split the specific kinetic energy into a compressible ( and translational ) part with @xmath112 and a solenoidal part with @xmath113 . for the full domain ( models 2c20 and 3c20 ) ,
the specific kinetic energy is clearly dominated by the compressible part , because of the inflows ( fig .
[ f : ecmpsolfull ] , _ left _ ) .
note however , that the solenoidal part ( thin lines ) is steadily increasing .
even though the energy input rate is constant , the total specific kinetic energy ( thick lines ) decays with time as a consequence of the radiative cooling ( see below ) . the specific energy restricted to the cold gas ( at @xmath38k , fig .
[ f : ecmpsolfull ] , _ right _ ) is dominated by the solenoidal component , and all components increase with time , indicating that once the gas has cooled down to its minimum temperature , it starts to store the kinetic energy from the inflows .
the solenoidal components are slightly larger for the 3d models , which is not surprising since the extra degree of freedom allows the gas to evade compression more efficiently
. one might speculate as to whether the total specific kinetic energy will decrease until it has reached a minimum at the moment when ( most of ) the gas has cooled down to the minimum temperature given by the cooling curve , or whether it does find some kind of equilibrium between kinetic energy input and thermal energy loss . since with increasing density
, the cooling will get more and more efficient , the latter seems unlikely . however , it is probably even more unlikely that the inflows are maintained for times long enough to establish equilibrium between kinetic energy input and thermal energy loss
. moreover , once gravity dominates the cold , dense regions , rather than reach a state of equilibrium , the clouds might pass through the phases of initial compression , turbulence generation , cooling and finally gravitational collapse .
_ left : _ specific kinetic energy for the whole simulation domain of models 2c20 and 3c20 against time , split into compressible , solenoidal and total part .
compressible motions dominate because of the inflows . _
right : _ specific kinetic energy for the cold gas ( @xmath38k ) of models 2c20 and 3c20 against time , split into compressible solenoidal and total part .
solenoidal motions dominate in the cold gas . ]
structure forms in colliding flows as a result of an interplay between dynamical and thermal instabilities .
the dynamical instabilities generate compressions and shear flows , while the thermal instabilities amplify density perturbations to the non - linear regime where they become potential seeds for self - gravitating structures .
the effect of the ti on the thermal state of the gas is shown in figure [ f : pn ] .
the solid line denotes the @xmath114 relation in thermal equilibrium .
it is a direct result of the heating and cooling processes included and displays the usual three regimes , namely an `` atomic '' regime for @xmath115 with an effective @xmath116 , an `` isothermal '' regime for @xmath117 with @xmath118 , and an unstable regime with @xmath119 .
each cell in the scatterplots figures [ f : pn ] and [ f : vn ] is color - coded with its thermal time scale ( eq . [
[ e : tauc ] ] ) .
positive time scales denote heating ( towards red colors ) and negative time scales correspond to cooling ( towards blue colors ) . at high densities ,
the thermal time scales are short ( greenish colors ) , which is mirrored in the very small scatter in @xmath120 around the equilibrium curve . moving to lower densities , the cooling time scale increases dramatically , until we reach the instability regime , which consists of a black triangular region in @xmath114 .
this consists mostly of gas which has passed the bounding shocks ( slightly increased densities due to the ( nearly ) adiabatic shock ) , but has not yet reached the cold dense phase .
since there is no gravity in the simulation , the ram pressure of the inflow ( see dashed lines in fig .
[ f : pn ] ) limits the maximum pressure gas can reach in the system , apart from slight overshoots due to waves . a substantial amount of gas resides at higher pressure than that of the ambient inflow ( higher by a factor of approximately @xmath121 for model 2c20 , and by @xmath122 for model 2c30 ) .
this gas has been `` overpressured '' by the highly compressible inflow .
even the gas in the isothermal branch ( i.e. gas which has cooled down to the minimum temperature ) resides at pressures between the initial thermal equilibrium pressure and the ram pressure of the inflow ( see also @xcite an effect which traditionally has been assigned to self - gravity , e.g. @xcite ) .
the more or less linear black regions at @xmath123 are a result of the adiabatic reaction of the shock - compressed gas , i.e. this gas has had not yet time to react to thermal effects ( see also @xcite ) .
the dashed red line denotes the hugoniot curve @xmath124 that is , the pairs of values @xmath125 for the state on one side ( index 1 ) of the shock front that are compatible with the rankine - hugoniot jump conditions across ( index 2 ) the shock front ( @xcite ) .
this has also been noted by @xcite . from figure
[ f : vn ] we see that the thermally unstable gas is generally still fast - moving , i.e. it has passed the bounding shocks but has not yet met the slab of dense , cold material ( of course the initial inflow just shows up as a single dot in the @xmath114 and @xmath126 plots ) .
high - density gas is mostly moving at a few km s@xmath31 , consistent with the line - of - sight velocity dispersion discussed in
[ ss : turbulence ] . a small fraction of cells ( @xmath127 for model 2c20 , and @xmath128 for model 2c30 ) seems to exhibit velocities larger than the inflow velocity .
this is hard to understand because there is no physical mechanism to accelerate gas to velocities higher than inflow velocities .
all of those cells coincide with the largest density / temperature jumps occurring in the simulation and have slightly increased temperatures as well .
thus , they are a consequence of a slight overshoot from the reconstruction step in the scheme .
since the fraction of faulty cells does not vary over time ( once the cold gas has formed ) , we simply neglect these cells for the analysis . for our later discussion of critical masses for molecular cloud formation
, figure [ f : massfrac ] summarizes the mass content in the three thermal regimes , defined by the warm stable range ( @xmath129k ) , the cold stable range ( @xmath38k ) and the instability range in between . obviously , once the cold regions form , most of the gas is locked there .
the amount grows linearly with time , indicating that although the pomcloc is turbulent , it essentially acts like a slab for collecting cold material .
the warm phase ( center panel ) contains a close to constant amount of mass with time
it is just a transitory phase in our models . at any given time only a small fraction ( but still a few percent ) of the total mass is passing through the unstable regime
this mass fraction depends slightly on the mach number of the inflow , because the shorter dynamical time scales compete with the thermal time scales , so that gas which has been thrown out of thermal equilibrium will have less chance of attaining equilibrium conditions again .
a similar effect can be seen in the temperature maps ( fig .
[ f : prettypic ] ) : for lower mach numbers , the transition from warm ( yellow ) to cold ( blue ) gas is much more pronounced ( see also @xcite and @xcite ) .
conversely , a high fraction of gas in the unstable regime @xcite might indicate a highly dynamical environment .
a similar effect exists for 3d versus 2d models : since the compressible modes are stronger in the 2d case , gas tends to be forced to cool in compressions , while in the 3d case , it can evade the compression and cooling by shearing motions , which even would lead to additional heating .
as figure [ f : massfrac ] shows , neither resolution nor geometry affects the mass fractions in the stable temperature regimes significantly .
especially , the fraction of mass in the cold phase ( fig .
[ f : massfrac ] , left panel ) is indistinguishable for models at resolution @xmath81 and @xmath91 . comparing the kinetic energy of the inflow to that in the unstable and cold ( see [ ss : thermal ] ) gas phases allows us to estimate the efficiency of turbulent `` driving '' in our models , i.e. the efficiency with which the highly ordered kinetic energy of the inflow is converted to the turbulent motions within the pomcloc .
figure [ f : efficiency ] shows the ratio of the kinetic energy change to the energy input rate of the inflow , the efficiency , @xmath130 in the cold ( filled symbols ) and unstable ( open symbols ) gas .
the integral extends over all cells within the chosen temperature regime , and @xmath131 is the area of the inflow .
overall , the efficiency @xmath132 in the cold gas ranges between @xmath1% and @xmath39% .
this is consistent with the ratio of the inflow speed to the line - of - sight velocity dispersion ( see e.g. fig . [
f : veldis_dens_obsint ] ) . in other words ,
most of the energy is lost due to atomic line cooling .
gas in the unstable phase reaches an efficiency which is smaller by a factor of @xmath133 : an effect of the lower densities ( the velocities are generally higher , see fig [ f : vn ] ) .
there seems to be a weak trend to smaller efficiency @xmath132 with larger energy inflow @xmath134 : larger inflow velocities result in higher compressions and thus in faster cooling , by which a growing fraction of the input energy is lost .
ratio of kinetic energy change to energy input rate in the unstable ( open symbols ) and cold ( filled symbols ) gas phases ( see expression [ e : efficiency ] ) for the models indicated .
question marks in the model names are wild cards for the mach numbers . only a few percent of the energy input is transferred to the cold gas . ] a crucial point in our analysis is whether the cold gas reaches conditions favorable for h@xmath0 formation , and thus eventually for molecular clouds ( so far we have only been talking about the precursors of molecular clouds ) .
three criteria need to be met .
\(1 ) for h@xmath0 formation , the gas temperature has to drop below a threshold temperature which we take to be @xmath135k @xcite .
this might appear to be somewhat high , since the sticking probability decreases rapidly with higher temperatures , and is near unity only for @xmath136 ( @xcite ; @xcite ) .
nevertheless , we chose the above temperature limit , because it turns out that once the temperature of a gas parcel has dropped below @xmath88k , it will quickly cool down to the minimum temperature allowed by the cooling prescription . in that sense ,
the time scales give a lower limit , i.e. denote the point at which the first h@xmath0 could potentially be formed .
since the cooling curve only extends down to @xmath77k anyway , we can not make more detailed statements about h@xmath0 formation . from figure
[ f : massfrac ] we already saw that for reasonable inflow speeds , a few @xmath137 pc@xmath31 at @xmath38k accumulate within approximately @xmath138 myrs . this material would be in principle available for h@xmath0 formation at least as far as the temperature is concerned .
\(2 ) the ( now cold ) gas has to stay cold `` long '' enough to allow for h@xmath0 formation .
time estimates for h@xmath0 formation vary . in their analysis of h@xmath0 formation behind shock fronts ,
@xcite quote time scales between @xmath39 and @xmath30 myrs , depending on the inflow momentum .
once h@xmath0 exists , its further formation could well be a run - away process , since self - shielding is much more efficient than dust shielding .
this means that the critical time scale is given by the onset of h@xmath0 formation .
a minimum requirement therefore is that the cold gas is not re - heated during this time scale . to measure this
, we followed the temperature history of the tracer particles and determined how long each particle stays cold .
the top panel of figure [ f : temphistory ] is a cumulative histogram of the time intervals over which tracer particles have temperatures @xmath38k .
apart from a small fraction at short time intervals @xmath139 ( these are particles at the rims of the cold regions ) , most of the particles stay cold for at least @xmath121 myrs .
in fact , for most models , the cold gas parcels stay cold for over @xmath140 myrs , i.e. once it has cooled down , by far most of the gas stays cold .
model 2c35 ended @xmath121 myrs after the initial tracer deployment , so in order to compare to the other models , all had to be evaluated to a maximum time interval of @xmath141 myrs , since otherwise it would look as if particles are re - heated in model 2c35 after @xmath121 myrs . to summarize the top panel of figure [ f : temphistory ] , the gas stays cold long enough to allow for h@xmath0 formation .
cumulative histogram of time intervals over which tracer particles stay at temperatures @xmath38k ( _ top _ ) , and over which tracer particles `` see '' column densities @xmath142 @xmath14 with respect to the ambient uv radiation field ( _ bottom _ ) . model sequence in mach number .
model 2c35 ended @xmath121 myrs after initial tracer deployment .
all models were run at @xmath81 . ]
\(3 ) finally , the cold , dense gas could be re - exposed to the ambient uv radiation field .
formation requires a critical column density of @xmath143 @xmath14 .
this we can determine again with the tracer particles , resulting in the bottom panel of figure [ f : temphistory ] , which combines the temperature criterion of the top panel and the column density threshold .
dropping the temperature criterion does not affect the result .
thus the critical quantity is the shielding column density , not the temperature . in other words ,
once the gas enters the `` cold '' phase , its thermal time scale is short compared to the dynamical time scale , so that the gas stays isothermal ( see fig . [
f : pn ] ) .
however , due to the continuous re - structuring of the cloud , gas is repeatedly re - exposed to the uv radiation field and the column density of the cold gas only stays above @xmath144 @xmath14 for about @xmath79 myr .
this is a direct consequence of the pomcloc s highly dynamical nature .
note that figure [ f : temphistory ] gives a pessimistic view : once a small fraction of the particles has reached conditions beneficial for h@xmath0 formation , self - shielding will set in .
the analysis up to now only includes shielding by dust .
the inflow velocity of model 2c10 is too low to reach sufficiently large column densities for h@xmath0 formation in the elapsed time of the simulation .
as @xcite found in one dimension , average column densities must achieve values of @xmath144 @xmath14 for h@xmath0 formation to occur because of the shielding requirement .
molecular clouds in our galaxy are complex and highly - structured , with broad , non - thermal line - widths suggesting substantial turbulent motions .
thus , molecular clouds very likely are not static entities and might not necessarily be in an equilibrium state , but their properties could well be determined by their formation process .
we presented numerical models of the formation of precursors of molecular cloud complexes ( pomclocs ) , in large - scale colliding hi - flows .
we will now summarize our findings and discuss their astrophysical relevance .
in [ s : numerics ] we discussed to what extent numerical artifacts might influence our conclusions .
one last effect needs to be mentioned , namely the choice of the boundary conditions . while the boundary conditions in the horizontal direction are prescribed by the inflow , we are free to define the boundaries in the transverse direction . as mentioned in
[ ss : initbound ] , our standard choice is periodic boundary conditions , i.e material leaving the simulation domain at the bottom reenters at the top and vice versa ( same holds for the tracer particles ) .
this might raise the question of whether the level of turbulence and the amount of cold gas in the simulations is a `` closed box '' effect in the sense that if the boundaries were open the compressed gas could leave the simulation domain before cooling down and contributing to the cold gas mass . in order to assess the effects of this specific boundary choice ,
we ran a set of models with open boundary conditions in the direction transverse to the inflow , i.e. gas is free to leave the simulation domain .
models s2c20 and s2c30 ( fig .
[ f : open ] , top ) have the same box size as models 2c20 and 2c30 ( fig .
[ f : prettypic ] ) , while in models l2c20 and l2c30 , the simulation domain was enlarged in the transversal direction ( fig .
[ f : open ] , bottom ) in such a way that the inflow region has the same size as in the square models ( e.g.2c20 , 2c30 ) , but above and below this inflow region we place `` inactive '' regions into which the compressed gas can stream . in other words ,
not only are the boundaries in the transverse direction open , but there is no inflow in the x - direction in the `` inactive '' regions ( in fact , they are open boundaries , i.e. material is free to leave the box ) . . _
bottom : _ stills of models l2c20 and l2c30 with open boundary conditions in the transversal direction and an `` inactive '' region above and below the inflow .
the resolution is @xmath145.[f : open ] ] the morphologies of models with open and periodic boundaries are slightly different , but not disquietingly so . for the non - periodic models , note that while there is a clear outflow observable in l2c20 and l2c30 , its signature is weak in s2c20 and s2c30 .
the open boundaries are implemented via a constant extrapolation of the last active cells , so that the pressure gradient at the boundary in the transverse direction is constant .
however because for models l2c20 and l2c30 we implement an `` inactive '' region above and below the inflow region , the gas in the dense compressed region , i.e. the inflow region , sees a pressure gradient along the transverse direction at the boundaries between the `` active '' and `` inactive '' regions . for models l2c20 and l2c30
, dense gas is definitely squeezed between the colliding flows out into the `` inactive '' regions .
how does this affect the mass budget ?
figure [ f : massfracbound ] shows the mass per length contained in the three temperature regimes as in figure [ f : massfrac ] , but for models 2c20 , s2c20 and l2c20 .
k ) , unstable ( @xmath146k ) and cold ( @xmath38k ) phase , for models 2c20 , s2c20 and l2c20 . for convenience ,
the mass fraction in the unstable regime has been multiplied by @xmath30 .
the total mass in the simulation domain increases ( more or less ) linearly for all models , despite the open boundary conditions of models s2c20 and l2c20.[f : massfracbound ] ] clearly , the total mass ( i.e. the sum over the three temperature regimes ) increases linearly , dominated by the growth of the cold mass fraction .
only model l2c20 shows some deviation from a linear evolution at later times .
the cooling time scale is much shorter than the flow time scale and the sound crossing time , so that the gas is compressed and cooled down before it can feel the boundaries . in other words , gas in
the incoming flow cools efficiently , and structures within the slabs form independent of the boundary conditions .
we conclude that our discussions of the mass budget and the level of turbulence are nearly independent of the specific boundary condition choice . observed line - of - sight velocity dispersions of a few km s@xmath31 in the cold gas
are reproduced by the models ( fig .
[ f : veldis_mach ] ) . yet ,
the _ internal _ velocity dispersions are generally subsonic , i.e. it seems as if the term `` supersonic '' is not necessarily an accurate description of the hydrodynamical state of the cold gas ( figs .
[ f : veldis_mach_obsint ] and [ f : veldis_dens_obsint ] ) .
@xcite argues , that because of the ages and small spatial dispersions of young stars in taurus , their velocity dispersions relative to their natal gas are very likely subsonic . the turbulent line - widths amount only to a fraction of the inflow velocity .
the subsonic internal velocity dispersions come as a surprise , and the question arises whether self - gravity could lead to higher internal dispersions .
while this is definitely something to test , to drive internal motions , gravity would have to act on local non - linear perturbations within the cold gas , for whose growth there simply might not be enough time available @xcite .
the above result would affect some aspects of the model of turbulence - controlled star formation ( see @xcite ) .
there , supersonic turbulence leads to non - linear ( isothermal ) shock compressions which in turn trigger ( at least partly ) fragmentation and subsequent gravitational collapse . by this fragmentation mechanism star formation
is `` localized '' in the sense that before the whole cloud can collapse , the non - linear density perturbations caused by supersonic turbulence have formed stars ( see also @xcite ) . in the absence of supersonic turbulence in the cold regions ,
the non - linear density perturbations must arise during the formation process of the cloud .
these perturbations could be envisaged as the cold filaments which are a consequence of the combination of dynamical and thermal instabilities .
the subsonic internal motions are mirrored in the specific kinetic energy modes ( fig [ f : ecmpsolfull ] ) and spectra ( fig .
[ f : specekin ] ) .
they clearly point to weakly compressible turbulence in the cold gas .
turbulence is generated with ease in the colliding flow scenario , and the resulting pomcloc reproduces observational bulk quantities .
one possible consequence of turbulence as an initial - condition effect is the efficient mixing between warm and cold phases in the complex .
there is an ongoing debate about the degree of mixing between atomic and molecular hydrogen in mccs .
while there is no doubt about the intermittent density ( and velocity ) structure in molecular clouds , it is less clear , what the diffuse , possibly warmer medium is made out of .
turbulent mixing would point to cnm , i.e. hi , while h@xmath0 self - shielding and h@xmath0 formation time scale arguments would favor a fully molecular cloud .
the `` openings '' in the clouds ( no matter whether they are e.g. channels , gaps or holes ) might play a crucial role for the internal dynamics of the clouds : @xcite argue that the channels of warm material in mccs allow energy ( in their case in the form of alfvn waves ) to enter the cloud and thus to drive the internal motions .
however , this would only refer to motions of the cold gas regions with respect to each other , but not necessarily to internal motions in the cold gas .
because of its very short thermal time scales , the dense gas closely follows the thermal equilibrium relation between pressure and density ( fig .
[ f : pn ] ) .
this gas phase is approximately isothermal and corresponds to the dense , isolated regions which we identify as precursors of molecular gas .
a perceptible amount of gas resides in the thermally unstable phase , consistent with observations by @xcite , also consistent with the modeling results of other groups ( e.g. @xcite ; @xcite ) .
the thermally unstable gas generally travels at high velocities ( fig . [
f : vn ] ) , meaning that it corresponds to gas which has passed the bounding shock but has not yet `` fallen onto '' the cold , dense objects ( where it would cool down to the isothermal cloud temperature ) .
the amount of unstable gas varies with the inflow mach number , which is a consequence of competing dynamical and thermal time scales . given a sufficiently high inflow speed ,
colliding flows are an efficient way to assemble enough cold gas to form molecular clouds . despite the highly structured objects , the amount of cold mass increases approximately linearly with time as if collected in a cold slab ( fig .
[ f : massfrac ] ) .
once the gas has cooled below @xmath147k in our models , it stays cold ( fig .
[ f : temphistory ] ) : the cold material has very short thermal time scales , i.e. it is close to isothermal ( in fig . [
f : pn ] , @xmath148 for @xmath149 @xmath25 ) .
the main limiting factor for h@xmath0 formation in our models is not the temperature of the gas , but the highly time - dependent column density . because of the dynamical nature of the clouds ,
gas is continuously re - exposed to the ambient radiation field .
most of the gas manages to stay above shielding column densities of @xmath142 @xmath14 only for @xmath1myrs at most ( fig . [ f : temphistory ] ) .
however , these estimates are pessimistic : ( a ) we only considered extinction by dust , not self - shielding by already formed h@xmath0 .
( b ) molecular clouds might not form `` from scratch '' , i.e. only from hi .
instead , the inflows might already contain a substantial amount of h@xmath0 ( see e.g. @xcite ) .
this not only would shift the mass budget in favor of h@xmath0 , but also would provide efficient self - shielding early on in the molecular cloud formation process . even more than @xmath30myrs after flow contact time , in all our runs the total mass collected in the cold gas falls short of typical molecular cloud masses by a factor of approximately @xmath30 .
the average column density essentially evolves one - dimensionally , i.e. @xmath150 for the fastest inflow speed in our models .
thus , at least @xmath30 myrs are needed to reach an _
average _ column density high enough for self - shielding ( since the dynamics lead to local focusing of the gas flows , isolated regions might reach this stage earlier , but not markedly so , see @xcite ) .
two remedies come to mind : ( 1 ) the cold mass growth rate depends nearly linearly on the inflow momentum ( fig . [
f : massfrac ] ) . however , the observed velocities and densities in the galactic hi limits the inflow momentum to approximately @xmath151 km s@xmath31 @xmath25 , or
@xmath152 @xmath14 . thus
, molecular clouds might preferentially form in flows where the ambient atomic density is higher than the average density , i.e. @xmath153 @xmath25 , @xmath154 km s@xmath31 . then , the mechanism described above might well dominate the formation of low - mass molecular clouds .
more massive objects need either more time or a different parameter regime as in e.g. galaxy mergers .
( 2 ) the models presented here do not account for gravity .
since a slab - like geometry is favored by construction in the colliding flow scenario , gravity would lead to collapse in the plane , causing the densities ( and column densities ) needed for shielding to be rapidly reached . in that sense ,
any time scales given for conditions favorable for h@xmath0 formation are upper estimates .
note that the _ lateral _ collapse can set in much later , and , if the cloud is large enough , and sufficiently dense substructures exist , might only occur after stars have formed locally .
we would like to stress the point that since the presented models are concerned with the formation of precursors of molecular clouds , the somewhat largish assembly time scales do not constitute a problem for short molecular cloud lifetimes .
as @xcite discussed , the ratio of the precursor s assembly time scale over the actual molecular cloud lifetime can be as large as the relative amounts of atomic over molecular gas ( approximately a factor of @xmath155 in the solar neighborhood , @xcite ) . for a molecular cloud life time of , say , @xmath2myrs in the solar neighborhood ( @xcite and references therein ) , the accumulation time for the pomcloc
would be expected to range around @xmath156myrs .
obviously , the _ molecular cloud _ lifetime starts only once molecules are being formed .
our models generally run longer than @xmath30 myrs . at an inflow velocity of @xmath30 km
s@xmath31 , this corresponds to a coherent flow of approximately @xmath34pc , or strictly speaking , twice that length for a colliding flow . although higher flow velocities and/or densities ( note that at least with the densities we seem to be on the lower side of the parameter range , see e.g @xcite ) would allow shorter time scales and thus shorter coherence lengths , the question nevertheless arises whether such coherent flows are realistic . on smaller scales
( and higher densities ) , the interaction of winds of nearby stars might provide a possible source ( see e.g. @xcite ) , while on larger scales , the most probable source would be interacting supernova shells .
the somewhat longish assembly timescales suggest that the initial densities of the flows are more likely to be a few @xmath25 to form molecular clouds with flows of order @xmath33-@xmath34 pc length .
spiral density waves can also produce coherent flows of the required length , at least in principle .
the main purpose of this work was to demonstrate that precursors of molecular cloud complexes forming in colliding flows reproduce observational measures such as line - widths and column densities , without any recourse to initially imposed perturbations .
the combination of dynamical and thermal instabilities efficiently generates the non - linear substructures necessary for a fast gravitational collapse and a ( close to ) instantaneous onset of star formation , thus meeting the short time scales required by observations ( e.g. @xcite ) .
once the material is cold and dense , it seems difficult to drive highly supersonic motions ( unless there is an internal driver , meaning stars ) ( @xcite ; @xcite , see however @xcite ) or even to mix it with warm gas @xcite .
however , the cold and warm gas must be efficiently mixed in order to allow the cold material to move `` freely '' at velocities which are supersonic with respect to the cold gas temperature .
thus , the cold substructures must arise during the formation process of the cloud , from compression and cooling in the warm phase . we demonstrated that colliding flows provide a generic physical mechanism for creating non - linear structure in pomclocs , without any recourse to perturbed initial conditions . in their further evolution ( i.e beyond the models presented here )
, these non - linear perturbations will be the seeds for gravitational collapse and star formation .
thus , our models show that from a dynamical point of view
the concept of `` short '' cloud life times is feasible .
ultimately , one can begin to see how realistic the concept of `` turbulent support '' of molecular clouds actually is ( see e.g. @xcite ) .
an mcc in the solar neighborhood has an average column density of @xmath157 @xmath14 @xcite , coinciding with the dust - shielding column density necessary for h@xmath0 formation .
once this value is reached , h@xmath0-formation is expected to begin . within @xmath1 or @xmath2 more myrs
, star formation will have occurred , and the cloud will have dispersed again . within this scenario , the average column density of an ensemble of mccs should be roughly constant , implying a mass - size relation of @xmath158 .
the latter would be expected for a @xcite relation of the type @xmath159 , consistent with observed scalings ( e.g. @xcite ) .
clearly , this study leaves us with many unanswered questions .
gravity might lead to global edge effects @xcite as well as to local `` supersonic '' velocities and to additional instabilities ( see e.g. @xcite ; @xcite ) .
actual h@xmath0 formation would further lower the temperature in the cold gas , in which case eventually supersonic velocities might be reached .
the inflows might already be partially molecular , in which case the filling factor of h@xmath0 should be low @xcite . and finally , magnetic fields could influence and maybe control the structure formation ( e.g. @xcite ; @xcite ; @xcite ; @xcite ) .
we thoroughly enjoyed the discussions with j. gallagher , k. menten , l. sparke , j. stutzki and e. zweibel , and we thank the anonymous referee for a critical and constructive report .
computations were performed on ariadne built and maintained by s. jansen at uw - madison , at the ncsa ( ast040026 ) , and on the sgi - altix at the usm , built and maintained by m. wetzstein and r. gabler .
this work was supported by the nsf ( ast-0328821 ) and has made use of nasa s astrophysics data system .
audit , e. , hennebell , p. 2005 , a&a , 433 , 1 de avillez , m. a. , & breitschwerdt , d. 2005 , , 436 , 585 ballesteros - paredes , j. , hartmann , l. , vzquez - semadeni , e. 1999 , apj , 527 , 285 ballesteros - paredes , j. , vzquez - semadeni , e. , & scalo , j. 1999 , , 515 , 286 bally , j. , stark , a. a. , wilson , r. w. , & langer , w. d. 1987 , , 312 , l45 balsara , d. s. 1996 , , 465 , 775 bergin , e. a. , hartmann , l. w. , raymond , j. c. , ballesteros - paredes , j. 2004 , apj , 612 , 921 blondin , j. m. , marks , b. s. 1996 , new ast .
, 1 , 235 burkert , a. , hartmann , l. 2004 , apj , 616 , 288 burkert , a. , lin , d. n. c. 2000 , apj , 537 , 270 burkert , a. , & heitsch , f. 2006 , _ in preparation _ cazaux , s. , & tielens , a. g. g. m. 2004 , , 604 , 222 cazaux , s. , caselli , p. , tielens , a. g. g. m. , le bourlot , j. , walmsley , m. 2005 , j. phys . , 6 , 155 chandrasekhar , s. 1961 , international series of monographs on physics , oxford : clarendon , 1961 , churchwell , e. , et al . 2004 , , 154 , 322 courant , r. , & friedrichs , k. o. 1948 , supersonic flow and shock waves ; interscience publishers , new york dame , t. m. 1993 , aip conf .
proc . 278 : back to the galaxy , 278 , 267 dib , s. , & burkert , a. 2005 , , 630 , 238 elmegreen , b. g. 1993 , , 419 , l29 elmegreen , b. g. 1999 , , 527 , 266 elmegreen , b. g. 2000 , , 530 , 277 elmegreen , b. g. , scalo , j. 2004 , araa , 42 , 211 falgarone , e. 1990 , liege international astrophysical colloquia , 29 , 173 falgarone , e. , philips , t. g. 1990 , apj , 359 , 344 falgarone , e. , & phillips , t. g. 1996 , , 472 , 191 ferrari , a. , & trussoni , e. 1983 , , 205 , 515 field , g. b. 1965 , apj , 142 , 531 gerwin , r. a. 1968 , rev .
, 40 , 652 hartmann , l. , ballesteros - paredes , j. , bergin , e. a. 2001 , apj , 562 , 852 hartmann , l. 2002 , , 578 , 914 hartmann , l. 2003 , , 585 , 398 heiles , c. 2004 , asp conf .
ser . 323 :
star formation in the interstellar medium : in honor of david hollenbach , 323 , 79 heiles , c. , & troland , t. h. 2003 , , 586 , 1067 heitsch , f. , zweibel , e. g. , slyz , a. d. , & devriendt , j. e. g. 2004 , , 603 , 165 heitsch , f. , burkert , a. , hartmann , l. , slyz , a. d. , devriendt , j. e. g. 2005 , , 633 , l113 heitsch , f. slyz , a. d. , devriendt , j. e. g. , burkert , a. 2006 , _ in preparation _ hennebelle , p. , prault , m. 1999 , a&a , 351 , 309 hennebelle , p. , prault , m. 2000 , a&a , 359 , 1124 hennebelle , p. , whitworth , a. p. , gladwin , p.
p. , & andr ' e , p. 2003
, , 340 , 870 hennebelle , p. , whitworth , a. p. , cha , s .- h . , & goodwin , s. p. 2004
, , 348 , 687 hennebelle , p. , & inutsuka , s. 2005 , astro - ph/0510389 hueckstaedt , r. m. 2003 , new ast .
, 8 , 295 hueckstaedt , r. m. , peterson , a. h. , & hunter , j. h. 2005 , , 361 , l35 hunter , j. h. 1979 , , 233 , 946 hunter , j. h. , & fleck , r. c. 1982 , , 256 , 505 hunter jr .
, j. h. , sandford ii , m. t. , whitaker , r. , klein , r.i .
1986 , , 305 , 309 hunter , j. h. , whitaker , r. w. , & lovelace , r. v. e. 1997 , , 482 , 852 keppens , r. , tth , g. , westermann , r. h. j. , goedbloed , j. p. 1999 , j. plas .
phys . , 61 , 1 klein , r. i. , woods , d. t. 1998 , apj , 497 , 777 koyama , h. , & inutsuka , s. 2002 , , 532 , 980 koyama , h. , & inutsuka , s. 2002 , , 564 , l97 koyama , h. , & inutsuka , s. 2004 , , 602 , l25 koyama , h. , & inutsuka , s. 2004 , rmxac , 22 , 26 kritsuk , a. g. , & norman , m. l. 2002 , , 569 , l127 kritsuk , a. g. , & norman , m. l. 2002 , , 580 , l51 kritsuk , a. g. , & norman , m. l. 2004 , , 601 , l55 kwan , j. , & sanders , d. b. 1986 , , 309 , 783 larson , r. b. 1981 , mnras , 194 , 809 lis , d. c. , pety , j. , phillips , t. g. , & falgarone , e. 1996 , , 463 , 623 looney , l. w. , wang , s. , hamidouche , m. , safier , p. n. , & klein , r. 2006 , astro - ph/0601437 mac low , m .-
2004 , , 289 , 323 mac low , m .- m . , & klessen , r. s. 2004 , reviews of modern physics , 76 , 125 mac low , m .- m . , balsara , d. s. , kim , j. , & de avillez , m. a. 2005 , , 626 , 864 malagoli , a. , bodo , g. , & rosner , r. 1996 , , 456 , 708 mckee , c. f. 1999 , nato asic proc .
540 : the origin of stars and planetary systems , 29 miesch , m. s. , & zweibel , e. g. 1994 , , 432 , 622 mizuno , a. , onishi , t. , yonekura , y. , nagahama , t. , ogawa , h. , & fukui , y. 1995 , , 445 , l161 parker , e. n. 1953 , , 117 , 431 passot , t. , vzquez - semadeni , e. , & pouquet , a. 1995 , , 455 , 536 priest , e. r. , & heyvaerts , j. 1974 , , 36 , 433 pringle , j. e. , allen , r. j. , & lubow , s. h. 2001 , , 327 , 663 prendergast , k. h. , xu , k. 1993 , j. comp . phys . , 109 , 53 roberts , b. 1974 , phys . fluids , 17 , 1771 sandham , n. d. , & reynolds , w. c. 1991 , journal of fluid mechanics , 224 , 133 snchez - salcedo , f. j. , vzquez - semadeni , e. , & gazol , a. 2002 , , 577 , 768 savage , b. d. , bohlin , r. c. , drake , j. f. , & budich , w. 1977 , , 216 , 291 sen , a. k. 1964 , phys . fluids , 7 , 1293 slyz , a. d. , prendergast , k. h. 1999 , a&as , 139 , 199 slyz , a. d. , devriendt , j. e. g. , bryan , g. , & silk , j. 2005 , , 356 , 737 slyz , a. d. , devriendt , j. e. g. , bryan , g. , heitsch , f. & silk , j. 2005 , , _ submitted _ stutzki , j. , & gsten , r. 1990 , , 356 , 513 tohline , j. e. , bodenheimer , p. h. , & christodoulou , d. m. 1987 , , 322 , 787 vasquez , b. j. 1990 , , 356 , 693 vzquez - semadeni , e. , passot , t. , & pouquet , a. 1995 , , 441 , 702 vzquez - semadeni , e. , ryu , d. , passot , t. , gonzlez , r. f. , gazol , a. 2006 , , _ accepted _ vishniac , e. t. 1994 , apj , 428 , 186 walder , r. , folini , d. 1998 , apss , 260 , 215 walder , r. , folini , d. 2000 , apss , 274 , 343 williams , j. p. , blitz , l. , & stark , a. a. 1995 , , 451 , 252 williams , j. p. , blitz , l. , & mckee , c. f. 2000 , protostars and planets iv , 97 wolfire , m. g. , hollenbach , d. , mckee , c. f. , tielens , a. g. g. m. , bakes , e. l. o. 1995 , apj , 443 , 152 young , m. k. r. , & mac low , m .-
2006 , astro - ph/0601005 | molecular cloud complexes ( mccs ) are highly structured and `` turbulent '' .
observational evidence suggests that mccs are dynamically dominated systems , rather than quasi - equilibrium entities .
the observed structure is more likely a consequence of the formation process rather than something that is imprinted after the formation of the mcc .
converging flows provide a natural mechanism to generate mcc structure .
we present a detailed numerical analysis of this scenario .
our study addresses the evolution of a mcc from its birth in colliding atomic hydrogen flows up until the point when h@xmath0 may begin to form .
a combination of dynamical and thermal instabilities breaks up coherent flows efficiently , seeding the small - scale non - linear density perturbations necessary for local gravitational collapse and thus allowing ( close to ) instantaneous star formation .
many observed properties of mccs come as a natural consequence of this formation scenario . since converging flows are omnipresent in the ism , we discuss the general applicability of this mechanism , from local star formation regions to galaxy mergers . |
a lot of attention has been paid to the problem of describing the transition form factor of neutral pion . to get new information about the wave function of the neutral pion @xcite , namely , the distribution of the neutral pion light - cone momentum fractions between light @xmath0 and @xmath1 quarks , is the motivation for numerous theoretical approaches to describe the transition form factor .
experimental information about the form factor is obtained in the process @xmath2 ( fig .
[ fg : ee scattering ] ) .
the kinematics , when one photon is almost real and the other is highly virtual with space - like momentum transfer squared @xmath3@xmath4was measured by several experimental collaborations : cello behrend:1990sr , cleo @xcite at low and intermediate @xmath5 and more recently by babar @xcite and belle @xcite at higher @xmath5 .
the babar collaboration fitted their experimental results for the form factor multiplied by @xmath5 as an increasing function of momentum transfer squared ( fig .
[ fig : fitac])@xmath6with @xmath7=1.04 , \label{babarparam } \\ & & a_{\mathrm{babar+}}=0.182~\mathrm{gev,\quad \beta } _ { \mathrm{babar+}}=0.252,\quad \chi _ { \mathrm{babar+}}^{2}/[35]=0.87 , \label{babarchi}\end{aligned}\]]where in square brackets we point out the number of degrees of freedom . in the first line only babar data are taken into account , while in the second line the full set from babar , cleo and cello data are taken into account .
the belle collaboration considered two parametrizations providing very similar goodness of the fit .
one is the same as ( [ parama ] ) ( fig .
fig : fitbc ) and the another corresponds to constant asymptotic behavior@xmath8the corresponding parameters are@xmath9=0.429 , \label{belleparam1 } \\ & & a_{\mathrm{belle+}}=0.172~\mathrm{gev,\quad \beta } _ { \mathrm{belle+}}=0.221,\quad \chi _ { \mathrm{belle+}}^{2}/[33]=0.637 , \label{bellechia } \\ & & b_{\mathrm{belle}}=0.209~\mathrm{gev,\quad c}_{\mathrm{belle}}=2.2~\mathrm{gev}^{2},\quad \chi _ { \mathrm{belle}}^{2}/[13]=0.435 , \label{belleparam2 } \\ & & b_{\mathrm{belle+}}=0.186~\mathrm{gev,\quad c}_{\mathrm{belle+}}=0.948~\mathrm{gev}^{2},\quad \chi _ { \mathrm{belle+}}^{2}/[33]=0.733 . \label{bellechib}\end{aligned}\ ] ] the data on the pion - photon transition form factor obtained by cello , cleo , babar and belle collaborations attract a lot attention mikhailov:2009kf , brodsky:2011yv , arriola:2010aq , noguera:2010fe , agaev:2010aq , kroll:2010bf , klopot:2012hd with aim to extract the pion distribution amplitude , a nonperturbative quantity important in description of hard exclusive hadronic processes
. the growing behavior of the form factor ( [ parama ] ) is in clear contradiction with the prediction of the approach based on the factorization theorem applied to this process ( see @xcite and references therein ) . on the other hand ,
this behavior might indicate a logarithmically enhanced asymptotic behavior of the form factor , as has been argued in dorokhov:2009dg , radyushkin:2009zg , polyakov:2009je , bystritskiy:2009bk , dorokhov:2010zzb . in @xcite ,
the neutral pion transition form factor was considered in the model with momentum independent light quark mass and quark - pion vertex . in this model , the leading order contribution to the form factor is given by the triangle diagram of fig .
[ fig : born ] .
its asymptotic behavior for the kinematics ( [ kinemat ] ) is double logarithmic one @xcite , @xmath10 , where @xmath11 is the quark mass serving as an infrared cutoff parameter . in order to fit the babar data one needs to tune the value of the mass parameter around @xmath12 mev ( see below ) . in @xcite ,
this model has been generalized considering momentum dependent quark mass and quark - pion vertex . in @xcite ,
the neutral pion transition form factor has been considered in the leading - double - logarithmic approximation in the scattering and annihilation channels . in the present work
, we are going to modify this model considering the lowest order gluon radiative corrections .
we shall use the well known expression for the virtual photon - quark vertex ( the so - called sudakov form factor @xcite ) which enters the triangle feynman diagram , describing the conversion of two photons to the neutral pseudoscalar meson .
the motivation of this and other similar studies is , first , to describe existing data in full interval of @xmath5 , and , second , to understand if there are any inconsistency between babar and belle data at large @xmath13 gev@xmath14 in section [ scattering ] both channels of pseudoscalar mesons production in electron - proton @xmath15 and electron - positron @xmath16 collisions are considered
. the second process could be the subject of future experimental investigation . in sections [ born approximation ] , [ nlo ] and [ sudakovexp ]
we discuss the leading order , next - to - leading order , and sudakov exponentiation calculations to the amplitude for the pion - photon transition . in section
fits we compare our model calculations with existing experimental data .
the matrix element for the neutral pion production in the high energy electron - proton scattering ( replacing the @xmath17 line in fig .
[ fg : ee scattering ] by a proton @xmath18 line ) @xmath19can be written as @xmath20for the kinematics of almost forward electron scattering @xmath21 in ( [ matel ] ) , @xmath22 mev is the pion decay constant , @xmath23 and @xmath24 is the pion transition form factor . in lowest order the form factor is given by the triangle quark loop integral ( fig .
[ fig : born ] ) @xmath25where @xmath26 is the pion mass , @xmath11 is the light quark mass parameter and @xmath27 is the current corresponding to the proton vertex @xmath28 u(p_{2 } ) , \label{protoncurr}\]]where @xmath29 is the proton mass , @xmath30 and @xmath31 are the sachs electric and magnetic form factors of the proton
. differential cross section for the process @xmath32 has the form @xmath33with @xmath34 ( \nu , p_{1},q , q_{1})(\nu _ { 1},p_{1},q , q_{1 } ) , \label{protblock}\]]where @xmath35 in the limit of large @xmath36 one has @xmath37 in ( [ difcross ] ) we use the normalization of the matrix element for the subprocess @xmath38 in accordance with the current algebra for the case of two real photons ( @xmath39 ) @xcite @xmath40corresponding to the width @xmath41 for the @xmath42 collisions , @xmath43one should put @xmath44 in the above expressions for @xmath45 .
let us now consider the amplitude @xmath46 in the context of the constituent quark model with momentum independent quark mass @xmath11 @xcite . within this model ,
the pion form factor is given by the quark - loop ( triangle ) diagram ( fig .
fig : born ) .
the result for the form factor in the considered asymmetric kinematics @xmath47 is given by ametller:1983ec@xmath48
\label{fborna}\]]with @xmath49 and the normalization is @xmath50 in @xcite this leading order ( lo ) expression was used to explain growing - type form factor as it was measured by the babar collaboration @xcite .
the quark mass was used as the only fitting parameter with the result ( see below table [ table:1 ] ) @xmath51 the expansion of the log term in the form factor ( [ fborna ] ) at large @xmath52 leads to@xmath53where the large logarithm is@xmath54numerically , @xmath55 and @xmath56 become indistinguishable at @xmath57 gev@xmath58 . in the following
, we would like to understand the role of radiative qcd corrections to the @xmath59 term in the form factor which are more important at large @xmath5 than neglected in ( [ fbornas ] ) power corrections .
to this end , we first reproduce the leading @xmath59 asymptotic term in ( fbornas ) by using two different techniques which become useful when considering the radiative corrections in the leading logarithmic approximation . one of these methods consists in joining the denominators of the integrand in ( [ borntriangle ] ) by using feynman parameters and performing the loop momentum integration . in case of @xmath60
one obtains @xmath61 an alternative calculation is based on the use of the sudakov parametrization of the loop momentum @xmath62with @xmath63 being the light - like four - vectors constructed from the external momenta , @xmath64 . by using the known relations
@xmath65and performing the integration in @xmath66 by the relation @xmath67we obtain @xmath68here , we also take into account both possibilities of positive and negative values of the sudakov parameters @xmath69 .
below , we shall use both the feynman and sudakov approaches .
the lowest order qcd corrections include three vertex subgraphs and three quark self - energy subgraphs ( fig .
[ fig : nlo ] ) .
kinematics of the main contributions of the feynman triangle amplitude correspond to the `` almost on - mass - shell '' quark connecting the `` almost on - mass - shell '' photon and the emission of a real pion , while the two other quark lines are essentially off - mass - shell .
thus , one of the vertex function associated with the off - mass - shell external photon underlies the sudakov conditions : both quarks are off - mass - shell .
the two other vertices describe the situation when the photon and one of quarks are almost on - mass - shell , while the other quark is off - mass - shell .
this configuration corresponds to the so - called landau case @xcite .
the contribution from the triangle amplitude with a mass operator insertion to the `` almost real '' quark line does not contain logarithmically enhanced terms .
consider first the vertex subgraph with external highly virtual photon with momentum @xmath70 .
the corresponding vertex function has the form ( see fig .
[ fig : vertexnlo]a ) @xmath71here , @xmath72 is a casimir invariant of the su(3 ) color group .
the logarithmically enhanced contributions arise from two kinematically different regions of virtual gluon momentum squared @xmath73 corresponding to small and large values of @xmath74 . using the sudakov parametrization @xmath75with @xmath76 light - like 4-vectors built from @xmath77 , we write @xmath78 in the limit ( [ lima ] ) , as @xmath79j_{\mu } \]]with@xmath80 for the first term in the square brackets of @xmath81 we obtain ( small @xmath74 region ) : @xmath82 j_{\mu } , \label{verta1}\end{aligned}\]]where the factor 2 is due to the two regions of negative and positive values of the sudakov parameters . in the derivation of ( [ verta1 ] ) , we used the relations @xmath83 , with @xmath69 being the sudakov variables associated with the loop momentum of the quark loop . the contribution from the region with large
@xmath74 comes from the second term in the square brackets of @xmath78 @xmath84here , we have to introduce the ultraviolet cut - off parameter @xmath85 , such as @xmath86 .
the usual procedure of joining the denominators and performing the integration leads to @xmath87^{2}}=-\frac{\alpha _ { s}c_{f}}{4\pi } j_{\mu } \ln \frac{\lambda ^{2}}{q^{2}}. \label{verta2}\]]where @xmath88 note , that we systematically omit the logarithmically suppressed terms .
after regularization , the cut - off parameter must be replaced by the quark mass .
the total answer for @xmath89 is @xmath90 j_{\mu } .
\label{vlog}\ ] ] the contribution of other feynman amplitudes ( fig .
[ fig : vertexnlo]b , c ) does not contain logarithmic enhancement .
we will illustrate this statement in the frame of qed @xcite .
let us consider the contribution of two remaining diagrams with one vertex function ( fig .
fig : vertexnlob ) and mass operator insertion ( fig .
[ fig : vertexnlo]c ) @xmath91u(k),\quad \label{vertb } \\ & & t=(p_{2}-q_{2})^{2}-m_{q}^{2},\quad p_{2}-q_{2}=q_{1}-k . \notag\end{aligned}\]]using the explicit expressions for the vertex @xmath92 and mass operator @xmath93 given in the appendix , we obtain @xmath94 , \]]where @xmath95 , \ \
\\ \widetilde{t } & = & \frac{t}{m_{q}^{2}},\ \
l_{t}=\ln \frac{-t}{m_{q}^{2}}.\end{aligned}\]]we see , that the relevant contribution is suppressed by small factor @xmath96 .
other terms do not contribute to the amplitude .
similar statement is valid for two remaining diagrams ( see fig .
[ fig : nlo ] b , e ) . collecting all logarithmically enhanced contributions , we finally obtain @xmath97 \label{fnlo } \\ & = & \frac{1}{2}l^{2}-\frac{\alpha _
{ s}c_{f}}{2\pi } \frac{1}{12}l^{3}\left ( \frac{1}{2}l-1\right ) .
\notag\end{aligned}\]]thus , the form factor at large @xmath5 ( [ fbornas ] ) , modified by next - to - leading ( nlo ) radiative qcd corrections , becomes@xmath98
in order to estimate the effect of the higher orders of qcd perturbation theory we apply to the next - to - leading logarithmic approximation ( [ fnlo ] ) the sudakov exponentiation hypothesis with the result @xmath99 \right\ } \notag \\ & = & \frac{1}{\rho } e^{-\frac{1}{2}l\rho } \int_{0}^{1}\frac{dx}{x-\frac{1}{l}}e^{xl\rho } \left ( 1-e^{-l^{2}\rho ( 1-x)\left ( x-\frac{1}{l}\right ) } \right ) , \label{fallorder}\end{aligned}\]]where @xmath100 for the form factor the resummation effects leads to the generalization of ( [ fnloa ] ) as@xmath101
there are two parameters within the parametrization of the pion transition form factor suggested by the quark model with radiative corrections .
they are the quark mass @xmath11 and the strong coupling constant @xmath102 .
we try to fit the experimental data for the form factor by varying these parameters .
the results of the fit are presented in tables [ table:1]-table:3 and figs .
[ fig : fitac]-[fig : fitbc ] . in fitting procedure
we use the parametrizations ( [ parama ] ) and ( paramb ) as reference parametrizations . from corresponding goodness of
fit ( [ babarparam ] ) , ( [ babarchi ] ) and ( [ belleparam1 ] ) , ( [ bellechia ] ) one concludes , first , that @xmath103 and , second , that the data set extended by experimental points of cello and cleo @xmath104 and @xmath105 . the first fact is due to systematically lower error bars for the babar point set than for the belle .
the second property we interpret as indication on better consistency of the babar data with previous data at lower @xmath5 than for the belle data .
first of all , let us use our model to fit the babar and belle data , considering the quark mass as a free parameter and the strong coupling to be fixed at @xmath106 , which corresponds to a renormalization scale of about 1 gev .
the results are given in table i. for the leading order ( lo ) fit we use expression ( [ fbornas ] ) , for the next - to - leading order ( nlo ) fit we use expression ( [ fnloa ] ) , and the sudakov resummed expression ( exp ) is ( [ fsudorder ] ) . in order to compare goodness of fit based on data used from different collaborations
we introduce the relative parameter@xmath107with @xmath108 is for babar or belle based data set . in ( [ relative chi2 ] ) , @xmath109 is for our model and for the power - like fits @xmath110 is from ( [ babarparam ] ) , ( [ babarchi ] ) and ( [ belleparam1 ] ) , ( bellechia ) , correspondingly . [ cols= " < , < , < , < ,
< , < , < " , ] in table [ table:3 ] we made a two - parametric fit to the babar and belle data .
we see that such a two - parametric fit has equal or even lower @xmath111 with respect to corresponding numbers in table [ table:1 ] . however , the price for that is a growing of the parameter @xmath112 especially for the case of belle data .
we consider such parametrization as not very physical .
in the present work we calculated the transition form factor of the neutral pion where one photon is virtual and another photon is real in the framework of the model where the light constituent quark mass and the quark - pion coupling are momentum independent .
we generalize the previous leading order results dorokhov:2009dg , bystritskiy:2009bk obtained by considering the triangle diagram by including the radiative gluonic corrections to the first order in perturbation theory contain the qcd scale parameter @xmath113 . ] .
the effect of higher order radiative corrections to the virtual photon vertex is estimated by applying the sudakov exponentiation hypothesis .
the results obtained are compared with existing experimental data on the pion transition form factor published by cello , cleo , babar and belle collaborations . in general
the model considered contains two parameters : the quark mass @xmath11 and the strong coupling constant @xmath102 .
first , we fit the data at fixed @xmath102 varying only @xmath11 . taking into account the radiative corrections
increases a little the fitting parameter up to @xmath114 mev and improves a goodness of the fit . considering separate fit of babar and belle data ( table [ table:1 ] )
one gets very close @xmath115 and the difference in @xmath11 is less than 10 mev . including also lower momentum data from cello and cleo ( table [ table:2 ] )
leads to inequality @xmath116 and almost coinciding @xmath11 . to our opinion
it means that fitting procedure is most sensitive to the intermediate momentum interval , where both set of data are in agreement . as for parametrizations discussed in introduction ,
babar data turns out to be more in accordance with lower momentum data than the belle data .
our model is more close to the tendency of the babar data .
resummation effects do not lead to significant changes in goodness of the fit comparing with the nlo results .
second , we try to fit data varying both model parameters ( table [ table:3 ] ) .
the fit of data becomes better , especially in the region of large @xmath5 .
however , @xmath102 has tendency to be close to unit .
this fact is considered as not physically justified and thus not taken into account in our final results .
let us emphasize that at the moment there are two sets of experimental data ( babar and belle ) on the pion - photon transition form factor at high @xmath5 .
they are fully consistent in the range of momentum transfer squared @xmath117 $ ] gev@xmath58 , but have different tendency at higher @xmath5 .
conditionally , the babar data show `` growing '' behavior at large @xmath5 ( see ( [ parama ] ) ) , while the belle data can be interpreted in twofold way as `` quasi - growing quasi - constant '' ( see ( [ parama ] ) and ( [ paramb ] ) ) . from experimental point of view , the data in the range @xmath118 $ ] gev@xmath58 are consistent at the level of @xmath119 standard deviation
. however , from theoretical point of view there are big debates on this difference mikhailov:2009kf , brodsky:2011yv , arriola:2010aq , noguera:2010fe , agaev:2010aq , kroll:2010bf , klopot:2012hd , dorokhov:2009dg , radyushkin:2009zg , polyakov:2009je , bystritskiy:2009bk , dorokhov:2010zzb .
it is clear that only new high statistic experiments can resolve this problem .
the emergence of `` growing '' data was so unexpected that , at first glance , it seemed that it is impossible to explain such behavior from the field theoretical point of view .
however , in dorokhov:2009dg , radyushkin:2009zg , polyakov:2009je , bystritskiy:2009bk , dorokhov:2010zzb there was noted that , if such `` growing '' behavior exists , then it may be related to unusual properties of the pion distribution amplitude in the vicinity of its edge points . this behavior was conditionally called as `` flat '' . in this case
the inverse moment of the pion distribution amplitude is not well defined . in particular ,
such behavior may be modeled if to assume that the pion is almost structureless . in @xcite ( and later in some other works ) it was shown that babar data can be described in the model with momentum independent quark mass and quark - pion coupling by using only one parameter : the constituent quark mass @xmath11 if its value is taken as @xmath12 mev .
this number was considered as a rather small from phenomenological point of view .
one of the main motivation of the present work is to see how sensitive this parameter to gluonic radiative corrections , when fit the babar and belle data .
the result is that the mass parameter becomes a bit heavier @xmath114 mev and the quality of the fit becomes better .
as it was shown in @xcite , a more advanced model , with momentum dependent quark mass and quark - pion vertex , has the same qualitative features as the model considered in this work . in the model @xcite ,
the soften quark propagator and quark - pion vertex lead to the single logarithmic asymptotic dependence on @xmath5 instead of the double logarithmic behavior , if the quark - pion vertex so that it corresponds to the `` flat '' pion distribution amplitude . the pion distribution amplitude
even can vanish at the end - points , but still simulate the logarithmic growth in rather wide range of momentum transfer including quite large values of it . when fit babar data , the mass parameter is still close to @xmath120 mev .
the calculations given in the present paper can be extended by consideration not only momentum dependent nonperturbative quark propagator and quark - pion vertex , but also momentum dependent nonperturbative gluon propagator . finally note that there are available data obtained by the babar collaboration for the @xmath121 , @xmath122 and @xmath123 transition form factors .
the comparison of our model calculations with data for these mesons is given in @xcite . from our point of view
the data show tendency that with increase of the meson mass the form factor changes its behavior from the `` growing '' regime to the `` constant '' regime . in the framework of our model
this change of behavior is related to strong dependence of the shape of the meson distribution amplitude on the meson mass . with growing meson mass
the meson distribution amplitude changes its shape from the `` flat '' one to the `` @xmath124-function '' shape for heavy mesons @xcite .
the study of present work shows that inclusion of the qcd corrections are essential in interpretation of experimental data on the pion transition form factor .
we thank yu.m .
bystritsky , v.v .
bytev , n.i .
kochelev and yu.s .
surovtcev for their interest to the subject of present work and discussions .
this work is supported in part by the russian foundation for basic research ( project no .
11 - 02 - 00112 ) .
the explicit expressions for the vertex function in the landau kinematics and the mass operator are @xcite @xmath125 , \label{vl } \\
\widehat{m}(p ) & = & \frac{\alpha } { 2m_{q}\pi } [ -a+f\frac{1}{m_{q}}(\hat{p}+m_{q})](\hat{p}-m_{q})^{2 } , \label{ml}\end{aligned}\]]with @xmath126 and the coefficients are @xmath127note , that the fictive photon mass @xmath128 introduced here , disappears from the final answer .
99 v. l. chernyak and a. r. zhitnitsky , jetp lett . * 25 * , 510 ( 1977 ) .
g. p. lepage and s. j. brodsky , phys .
b87 * , 359 ( 1979 ) .
a. v. efremov and a. v. radyushkin , phys .
b94 * , 245 ( 1980 ) .
h. j. behrend _ et al . _
[ cello collaboration ] , z. phys .
c * 49 * ( 1991 ) 401 .
j. gronberg _ et al . _
[ cleo collaboration ] , phys . rev . * d57 * , 33 ( 1998 ) , arxiv : hep - ex/9707031 . b. aubert
_ et al . _
[ babar collaboration ] , phys . rev . *
d80 * , 052002 ( 2009 ) , arxiv:0905.4778 .
s. uehara _ et al . _ [ belle collaboration ] , phys .
d * 86 * ( 2012 ) 092007 [ arxiv:1205.3249 [ hep - ex ] ] .
s. v. mikhailov and n. g. stefanis , nucl
b821 * , 291 ( 2009 ) , arxiv:0905.4004 .
brodsky , f .-
cao and g.f .
de teramond , phys . rev .
* d84 * , ( 2011 ) 033001 [ arxiv:1104.3364 [ hep - ph ] ] .
e. ruiz arriola and w. broniowski , phys . rev .
* d81 * , ( 2010 ) 094021 [ arxiv:1004.0837 [ hep - ph ] ] .
s. noguera and v. vento , eur .
j. * a46 * ( 2010 ) 197 [ arxiv:1001.3075 [ hep - ph ] ] .
agaev , v.m .
braun , n. offen and f.a .
porkert , phys .
* d83 * , ( 2011 ) 054020 [ arxiv:1012.4671 [ hep - ph ] ] .
p. kroll , eur .
j. * c71 * ( 2011 ) 1623 [ arxiv:1012.3542 [ hep - ph ] ] .
y. klopot , a. oganesian and o. teryaev , phys . rev . *
d87 * , ( 2013 ) 036013 [ arxiv:1211.0874 [ hep - ph ] ] .
a. e. dorokhov , phys .
* 7 * , 229 ( 2010 ) , arxiv:0905.4577 .
a. v. radyushkin , phys . rev . * d80 * , 094009 ( 2009 ) , arxiv:0906.0323 .
m. v. polyakov , jetp lett . * 90 * , 228 ( 2009 ) , arxiv:0906.0538 .
m. bystritskiy , v. v. bytev , e. a. kuraev and a. n. ilyichev , phys .
* 8 * ( 2011 ) 73 [ arxiv:0912.3668 [ hep - ph ] ] .
a. e. dorokhov , jetp lett .
* 92 * , 707 ( 2010 ) .
a. e. dorokhov , arxiv:1003.4693 [ hep - ph ] .
a. e. dorokhov , nucl .
* 225 - 227 * ( 2012 ) 141 [ arxiv:1109.3754 [ hep - ph ] ] .
v. v. sudakov , sov .
jetp * 3 * , 65 ( 1956 ) .
gerasimov , yad .
* 29 * ( 1979 ) 513 [ sov
* 29 * ( 1979 ) 259 ] , jinr - e2 - 11693 ( 1979 ) .
l. ametller , l. bergstrom , a. bramon , and e. masso , nucl . phys . *
b228 * , 301 ( 1983 ) .
l. b. okun , leptons and quarks , moscow , science , 1981 .
e. a. kuraev , n. p. merenkov and v. s. fadin , yad . fiz . * 42 * ( 1987 ) 782 [ sov
* 45 * ( 1987 ) 486 ] .
n. g. stefanis , a.p .
bakulev , s. v. mikhailov , a.v .
pimikov , arxiv:1202.1781 [ hep - ph ] .
a. a. pivovarov , phys .
* 66 * ( 2003 ) 902 [ yad
. fiz .
* 66 * ( 2003 ) 934 ] [ hep - ph/0110248 ] .
r. boughezal and k. melnikov , phys .
b * 704 * ( 2011 ) 193 [ arxiv:1104.4510 [ hep - ph ] ] .
d. greynat , e. de rafael , jhep * 1207 * ( 2012 ) 020 [ arxiv:1204.3029 [ hep - ph ] ] .
f. del aguila and m.k .
chase , nucl .
b193 * , ( 1981 ) 517 .
e. braaten , phys .
* d28 * , ( 1983 ) 524 .
kadantseva , s.v . mikhailov and a.v .
radyushkin , yad . fiz .
* 44 * ( 1986 ) 507 [ sov
* 44 * ( 1986 ) 326 ] . | we calculate the transition form factor of the neutral pion where one photon is virtual and another photon is real in the model where the light constituent quark mass and the quark - pion vertex are taken to be momentum independent .
radiative corrections to the lowest order triangle quark feynman amplitude are calculated .
the resummation of the lowest radiative corrections to the virtual photon vertex is done by applying the sudakov exponential hypothesis . using fitting parameters , the quark mass and the strong coupling constant , the results on the pion transition form factor are compared with existing data published by cello , cleo , babar and belle collaborations . |
the homogeneous painlev ii equation ( pii ) is the second order nonlinear ode @xmath2 despite its unassuming form , its solutions , known as the _ painlev transcendents _ , appear in exact solutions of many models in mathematical physics .
for example , one particular solution to is the one satisfying the boundary condition @xmath3 where @xmath4 is the airy function .
this solution is known as the _ hastings mcleod solution _ @xcite .
it is particularly important in random matrix theory , for it defines the celebrated tracy
widom distributions which describe the generic soft edge behavior of random matrices from orthogonal- , unitary- , or symplectic - invariant ensembles @xcite , @xcite .
the pii equation is an _ integrable equation _ , and its integrability is characterized by the existence of _ lax pairs_. a lax pair , or more generally a lax system , is a system of overdetermined linear differential equations whose compatibility implies a nonlinear equation .
let @xmath5 be an @xmath6 matrix - valued function with variables @xmath7 .
let @xmath8 be an ( overdetermined ) system of differential equations satisfied by @xmath9 with @xmath10 coefficient matrices @xmath11 .
for the overdetermined differential equations to have nontrivial solutions , we need the compatibility among @xmath11 , the _ frobenius compatibility conditions _ , sometimes called _ zero - curvature relations _ : @xmath12 = 0 , \quad \text{for all $ i , j = 1 , \dotsc , r$}.\ ] ] the frobenius compatibility conditions are in general nonlinear differential equations for the entries of @xmath13 , and we call the system the lax system for the nonlinear equation(s ) . in the most common cases @xmath14 and we call the system a lax pair , but we may also consider the general case @xmath15 . the term lax pair originates with the work of peter lax in the late 1960 s @xcite , in which he used the compatibility of a pair of linear differential equations to study a nonlinear _ partial _ differential equation . in the problem considered by lax
the evolution of the time variable gives an _ isospectral _ deformation of the linear operator . on the other hand ,
painlev equations represent _ isomonodromic _ deformations of the analogous linear equations with respect to the singularities , i.e. , the monodromy data is invariant as the argument of the ( fixed ) painlev function changes , and the isomonodromic relations are expressed in the same form of lax pairs ( * ? ? ?
* chapter 4 ) .
the idea of representing the painlev equations as isomonodromy deformations of a system of linear equations is nearly as old as the painlev equations themselves , dating back to the work of fuchs @xcite and later garnier @xcite
. therefore it may be more appropriate to call the overdetermined systems and rather than lax pairs / systems .
such terminology can be found in the literature , see @xcite and @xcite . however , the phrase lax pair is much more abundant in the literature and this is the nomenclature we use , following the terminology of @xcite , @xcite , @xcite , and @xcite .
nonlinear differential equations which possess a lax system representation are in some sense integrable , although they can be rather complicated .
all of the painlev equations , including , can be represented by lax pairs / systems @xcite .
however , the construction of lax pairs / systems for a given painlev equation is far from trivial , and the relations between different lax pair / systems for a painlev equation deserve investigation for their own sake . in this paper
we demonstrate the relation between one classical lax pair and a recently discovered lax system for the pii equation .
however , the main motivation of our paper is not purely theoretical , but is driven by the appearance of these lax systems in random matrix theory and related problems .
the classical lax pair and the new lax system are both related to random matrix theory , but in quite different aspects .
first we present a classical lax pair for , found by flaschka and newell @xcite .
the flaschka newell lax pair was originally presented for the general pii equation which has a free parameter ( see section [ subsec : outlook ] ) , and we only present it for the homogeneous case . a different lax pair for pii was found by jimbo and miwa around the same time @xcite ( with a precursor in @xcite ) , but in the homogeneous case the jimbo miwa lax pair can be reduced to the flaschka
newell one ( * ? ? ?
* section 4.2 ) .
other lax pairs associated to pii have been found by harnad , tracy , and widom in @xcite ( of size @xmath0 ) and by joshi , kitaev , and treharne in @xcite ( of size @xmath16 ) .
the equivalence among these lax pairs is discussed in @xcite .
let @xmath17 be a @xmath18 matrix - valued function with variables @xmath19 and @xmath20 which satisfies the overdetermined equations [ int:2 ] @xmath21 where @xmath22 and @xmath23 and @xmath24 are parameters which may depend on @xmath20 .
it is an amiable exercise to show that the compatibility of the two equations in is reduced to the fact that @xmath25 solves the painlev equation , and the parameter @xmath24 in is @xmath26 it is known that all solutions to the are meromorphic , so if we choose @xmath27 to be any particular solution to and take @xmath28 , then the system is solvable provided @xmath29 is not a pole of the chosen pii transcendent .
notice then that , given a particular solution @xmath30 and fixing @xmath29 that is not a pole of this solution , we can find a solution to the overdetermined equation using only , given proper initial conditions .
thus below we concentrate on when we talk about the solutions to , where @xmath30 is a fixed solution to , @xmath31 , and @xmath29 is a constant that is not a pole of @xmath23 . in some formulas in this paper
, we suppress the dependence on @xmath20 if it is treated as a constant . since @xmath32 is the only singular point of @xmath33 , and @xmath34 it is natural to construct the fundamental solution @xmath35 such that @xmath36 but @xmath32 is an irregular singularity of @xmath33 , so the stokes phenomenon allows us only to consider the solution @xmath35 that satisfies sectorally . for a rigorous version of the heuristic argument
above see ( * ? ? ?
* section 5.0 ) .
for @xmath37 , define the sectors ( see figure [ fig:2x2jump ] ) , @xmath38 their boundaries are the rays with outward orientation @xmath39 then there are fundamental solutions @xmath40 to such that @xmath41 satisfies the boundary condition in sector @xmath42 .
of course the solution space to is two dimensional and so there are linear relations between the solutions @xmath40 .
these relations depend on the particular painlev transcendent appearing in the coefficient matrices @xmath33 and @xmath43 in , and can be described in the following way ( * ? ? ?
* section 5.0 ) . for each pii solution @xmath30
to , there is a triple of complex numbers @xmath44 satisfying the relation @xmath45 such that the fundamental solutions @xmath46 associated with @xmath47 satisfy @xmath48 the jump matrices @xmath49 are called the _ stokes matrices _ , and the numbers @xmath50 are called the _ stokes multipliers _ corresponding to the given pii solution @xmath30 .
remarkably , each triple
@xmath44 of stokes multipliers satisfying corresponds uniquely to a pii solution , and so the solutions to pii are parametrized by the surface . thus in order to specify a solution to pii , it is enough to specify the stokes multipliers @xmath44 , see ( * ? ? ?
* proposition 5.1 ) . in figure [ fig:2x2jump ]
we show the rays , sectors , and the jump matrices @xmath49
. expressed in @xmath51 and @xmath52 . ] expressed in @xmath51 and @xmath52 . ] for a given set of stokes multipliers , the jump properties determine any of the fundamental solutions in terms of the solution @xmath53 . indeed
if we denote @xmath54 where @xmath51 and @xmath52 are two @xmath55-dimensional vector - valued functions defined on the whole complex plane , then the other @xmath46 are expressed in @xmath51 and @xmath52 as in figure [ fig:2x2phi ] .
the asymptotics of the columns of @xmath46 are summarized below ( with @xmath56 being any small positive constant ) : @xmath57 as mentioned earlier , the hastings mcleod solution to , the one satisfying , is of special importance in random matrix theory .
it is the solution to pii that corresponds to the stokes multipliers @xmath58 , and it is well established that it has no poles on the real line .
thus the solution @xmath59 exists for any real @xmath29 ( * ? ? ?
* section 11.7 ) . consider the one - matrix model given by the probability measure on the space of @xmath6 hermitian matrices @xmath60 , @xmath61 where @xmath62 is the potential and @xmath63 is a scaling factor .
the eigenvalues of @xmath60 are a determinantal process that is characterized by a correlation kernel . in the case that @xmath64 and @xmath65 , the model is in a critical phase if @xmath66 . as @xmath65
, under the double scaling limit @xmath67 , the correlation kernel at @xmath68 and @xmath69 converges to @xmath70 where @xmath71 and @xmath72 are the two components of the 2-vector @xmath51 defined in , see @xcite .
we use the notation @xmath73 to emphasize that this kernel arises in a 1-matrix model and to differentiate it from the kernel which arises in a 2-matrix model , which we denote @xmath74 .
note that although we only state the limiting correlation kernel for a very special potential function , the convergence to @xmath75 holds for a large class of potentials that have a quadratic interior critical point .
see @xcite for the universality of the limiting kernel @xmath75 .
finally we remark that if we give the potential @xmath62 a logarithmic perturbation at @xmath76 , i.e. , let @xmath77 , then the limiting kernel at @xmath76 is changed , and it is expressed by the flaschka newell lax pair for the hastings mcleod solution of the _ inhomogeneous _ pii equation .
see @xcite for detail , and also see section [ subsec : outlook ] .
now we introduce the other lax system for the pii equation , which was discovered recently by delvaux , kuijlaars , and zhang in their study of non - intersecting brownian motions @xcite , by delvaux in the study of non - intersecting squared bessel processes @xcite , and by duits and geudens in their study of the 2-matrix model @xcite , see also @xcite , @xcite . in its most general form this lax system is a @xmath78-dimensional overdetermined differential system consisting of 16 equations . here
we consider a @xmath79 matrix valued function @xmath80 , and the lax system is [ int:7 ] @xmath81 the coefficient matrix @xmath82 is given by @xmath83 where each @xmath84 is a @xmath18 block , such that @xmath85 here the numbers @xmath86 and @xmath87 are positive constants , and @xmath88 , @xmath89 , @xmath23 , @xmath90 , and @xmath91 depend on @xmath92 .
we relegate the formulas for @xmath93 to appendix [ sec : formulas_vvw ] , since we do not use them in the rest of this paper . in the symmetric case @xmath94 and @xmath95 , see also ( * ? ? ?
* section 5.3 ) , @xcite , and in the @xmath96 case , see also ( * ? ? ?
* section 5.2 ) . by the compatibility of the overdetermined system , which is routine but laborious , see ( * ? ? ?
* section 6.5 ) , we derive @xmath97 and @xmath23 and @xmath91 are functions of @xmath98 furthermore , @xmath99 satisfies the pii equation , @xmath100 is the derivative with respect to @xmath29 , and @xmath91 is the pii hamiltonian @xmath101 which satisfies @xmath102 now as with the lax pair , we fix a particular solution @xmath30 to pii and assume @xmath29 is not a pole of this solution .
we can then solve the lax system by alone , with proper initial conditions .
the authors of @xcite , @xcite , @xcite , and @xcite introduced the lax system as a technical tool to study the tacnode riemann
hilbert problem ( rhp ) , a @xmath79 riemann hilbert problem associated with the pii equation .
the tacnode rhp is only defined for the hastings mcleod solution to pii , but the lax system is algebraic and the frobenius compatibility conditions are independent of boundary condition , so the lax system exists for all solutions to the pii equation . from the lax system we can construct an rhp that is associated with all solutions to the pii equation and thus generalize the tacnode rhp .
see riemann
hilbert problem [ rhp:4x4 ] in section [ subsec : tacnode_rhp ] below . since @xmath32 is the unique singular point of @xmath82 , it is natural to put the boundary condition to the solution @xmath60 at @xmath32 .
the situation is a bit more complicated than for the @xmath0 lax system , since infinity is , in the language of @xcite , a _ general irregular singular point _ of the coefficient matrix @xmath82 .
nonetheless , it is possible to transform the equation into one with a regular singular point by means of an explicit change of variable , and then to derive the asymptotic structure of its solutions using the methods of @xcite .
this asymptotic structure was derived by duits and geudens in @xcite . in order to describe it , we define the functions @xmath103 , \end{aligned}\ ] ] and then the @xmath78-dimensional vector - valued functions @xmath104 and the matrix - valued function @xmath105 for the fractional powers in we take the principal branches , so @xmath106 has cuts on @xmath107 and @xmath108 .
more precisely , the functions @xmath109 and @xmath110 each have cuts on the positive real axis , and the functions @xmath111 and @xmath112 each have cuts on the negative real axis .
we also define the function @xmath113 to be the continuation of @xmath106 from the upper half plane with a cut on the negative imaginary axis , and and @xmath114 to be the continuation of @xmath106 from the lower half plane with a cut on the positive imaginary axis .
to be concrete , we denote @xmath115 such that for all @xmath116 , @xmath117 in @xmath118 , and the branch cut for @xmath119 is @xmath120 . if we denote by @xmath121 ( resp .
@xmath122 ) the limiting value of @xmath123 from the upper ( resp .
lower ) half - plane for @xmath124 , then we have the following relations : @xmath125 again due to the stokes phenomenon , we can not find solutions that satisfy the boundary conditions at @xmath32 from all directions , but only sectorally . here
we follow the notation in @xcite and define six overlapping sectors in the complex plane @xmath126 as shown in figure [ fig : omega_k ] .
the following result was proved in ( * ? ? ?
* lemma 5.2 ) .
[ prop : duits - geudens ] for fixed @xmath127 and @xmath128 one of the sectors defined in , the equation has a unique fundamental solution @xmath129 such that as @xmath130 within @xmath128 , @xmath131 in ( * ? ? ?
* lemma 5.2 ) , the above result is stated for @xmath132 , and @xmath133 , but it is trivial to extend to the more general parameters @xmath134 and @xmath135 . the general theory outlined in ( * ? ? ?
* theorem 19.1 ) would indicate a weaker result , namely an asymptotic expansion in powers of @xmath136 rather than in powers of @xmath137 .
the stronger asymptotics above are the result of some symmetry in the equation , see the proof in @xcite .
below we construct six 4-vector - valued functions solutions to @xmath138 which we denote by @xmath139 , explicitly from the solutions to the flaschka newell lax pair .
it is then shown that the solution @xmath140 is recessive in the sector @xmath42 which was defined in .
thus these solutions comprise the essential components of the fundamental solutions @xmath129 satisfying . in order to state the construction and properties of @xmath139
, we first introduce some notations .
suppose @xmath141 is a trivalent contour , where @xmath142 and @xmath143 are three rays in the complex plane which meet at the origin such that @xmath144 and @xmath145 are oriented away from the origin , and @xmath143 is oriented towards the origin .
denote @xmath146 , @xmath147 , @xmath148 , and @xmath149 as @xmath150 , \\ { \gamma}_1=\exp\left(-\frac{8r_1 ^ 4\tau^3}{3(r_1 ^ 2+r_2 ^ 2)^3}+\frac{4r_1s_1\tau}{r_1 ^ 2+r_2 ^ 2}\right ) , \quad { \gamma}_2=\exp\left(-\frac{8r_2 ^ 4\tau^3}{3(r_1 ^ 2+r_2 ^ 2)^3}+\frac{4r_2s_2\tau}{r_1 ^ 2+r_2 ^ 2}\right ) , \end{gathered}\ ] ] and then the function @xmath151 and the related functions @xmath152 define now an integral transform @xmath153 that transforms two @xmath55-dimensional vector - valued functions @xmath154 and @xmath155 into a @xmath78-dimensional vector - valued function given by , @xmath156 where @xmath86 , @xmath87 , @xmath157 , @xmath158 , @xmath159 , and @xmath88 are the parameters in and , and @xmath160 where @xmath161 is any fixed pii solution evaluated at @xmath29 defined in , and @xmath162 is defined in .
we have the following proposition . [ diffeq ]
fix some solution @xmath23 to and let @xmath29 be as in such that it is not a pole of @xmath23 .
let @xmath163 and @xmath164 be any two 2-vector solutions to .
assume that for a particular choice of @xmath165 the integral transform @xmath166 exists and is finite for every @xmath167 .
then @xmath168 solves the differential equation .
the proof of this proposition is given in section [ diff_eq_proof ] .
now we make a special choice of @xmath169 and @xmath170 in proposition [ diffeq ] and define the particular solutions @xmath139 of . recall the rays @xmath171 defined in ( see also figure [ fig:2x2jump ] ) .
we define the trivalent contours @xmath172 as the @xmath165 in proposition [ diffeq ] as follows : @xmath173 where the contours @xmath174 are oriented towards infinity , @xmath175 means the contour @xmath176 oriented in the opposite direction , and @xmath177 . for an illustration of the contours ,
see figure [ fig : six_contours ] .
then we define @xmath178 where @xmath92 are parameters in the formula of @xmath153 , and @xmath179 and @xmath180 are the columns of fundamental solutions to , given as @xmath181 where the parameter @xmath20 is determined by the relation , and @xmath44 are the stokes multipliers corresponding to the chosen pii solution .
we use the notational conventions @xmath182 and @xmath183 , and the subscripts refer to the matrix entries . here
we note that all the @xmath179 and @xmath180 are linear combinations of @xmath51 and @xmath52 , as shown in figure [ fig : six_contours ] , and by the jump condition we see that , @xmath184 from the definitions of the functions @xmath185 and the relation , the linear relations between them are easy to see , especially in figure [ fig : six_contours ] .
we have , for example , the pair of independent relations [ nj_relations ] @xmath186 @xmath187 for the integral representation of @xmath139 . on each ray a two dimensional vector in the form of @xmath188 is given , and they are @xmath189 , @xmath190 , or @xmath191 in the integral formulas @xmath192 . ]
the next result of the paper is that the solutions @xmath139 of satisfy the asymptotics of the columns of the fundamental solutions @xmath193 in some of the sectors @xmath194 , and thus these fundamental solutions can be built from the columns @xmath139 . to state the proposition , we recall the functions @xmath195 and @xmath196 defined in . [ thm : main_asy ]
suppose @xmath197 is a small constant .
for @xmath198 with @xmath199 fixed and @xmath200 , we have the following asymptotic results , where all power functions take the principal branch @xmath201 : 1 .
[ enu : thm : main_asy_1 ] uniformly for @xmath202 @xmath203 uniformly for @xmath204 @xmath205 and uniformly for @xmath206 $ ] @xmath207 2 .
[ enu : thm : main_asy_2 ] uniformly for @xmath208 @xmath205 uniformly for @xmath209 @xmath210 and uniformly for @xmath211 $ ] @xmath212 3 .
[ enu : thm : main_asy_3 ] uniformly for @xmath213 @xmath214 uniformly for @xmath215 @xmath203 and uniformly for @xmath216 $ ] @xmath217 4 .
[ enu : thm : main_asy_4 ] uniformly for @xmath218 @xmath219 uniformly for @xmath220 @xmath221 and uniformly for @xmath222 $ ] @xmath223 5 .
[ enu : thm : main_asy_5 ] uniformly for @xmath224 @xmath221 uniformly for @xmath225 @xmath226 and uniformly for @xmath227 $ ] @xmath228 6 .
[ enu : thm : main_asy_6 ] uniformly for @xmath229 @xmath230 uniformly for @xmath231 @xmath232 and uniformly for @xmath233 $ ] @xmath234 the proof is given in section [ asymptotics ] . to visualize the result of proposition [ thm : main_asy ] , we summarize it in figure [ fig : prop1_3_summary ] . .
the dividing lines are the real and imaginary axes , as well as @xmath235 and @xmath236 .
they separate the complex plane into 8 sectors , and within each sector the leading order behavior of some of the solutions @xmath237 can be identified with the columns of @xmath238 and @xmath239 . if the function @xmath240 is written without any superscript it means that @xmath241 throughout the given sector .
the zigzag lines are the branch cuts for @xmath238 ( the negative imaginary axis ) , and @xmath239 ( the positive imaginary axis ) . ] as a consequence of proposition [ thm : main_asy ] , we see that the vector @xmath140 is recessive in the sector @xmath42 shown in figure [ fig:2x2jump ] .
we now describe the entries of the fundamental solutions defined in proposition [ prop : duits - geudens ] in terms of the solutions @xmath242 .
[ fund_solutions ] fix a pii solution @xmath30 with stokes multipliers @xmath44 . for @xmath243 ,
let @xmath129 be the unique @xmath1 matrix - valued solution to which satisfies .
we have the following explicit formulas : @xmath244 since we know the linear relations for the six fundamental solutions described above , we may describe them as the solution to a riemann hilbert problem ( rhp ) . in order to state the rhp , define the sectors @xmath245 as @xmath246 see figure [ fig:4x4jump ] .
we then define the function @xmath247 piecewise in the complex plane as @xmath248 then @xmath247 satisfies the following rhp . [ rhp:4x4 ] 1 . the @xmath79 matrix - valued function @xmath60 is analytic in each of the sectors @xmath245 defined in , continuous up to the boundaries , and @xmath249 as @xmath250 .
2 . on the boundaries of the sectors
@xmath245 , @xmath251 satisfies the jump conditions @xmath252 for the jump matrices @xmath253 with constant entries specified in figure [ fig:4x4jump ] .
3 . as @xmath130
, @xmath247 satisfies the asymptotics @xmath254 where @xmath255 and @xmath256 are defined in . for rhp [ rhp:4x4].,title="fig : " ] for rhp [ rhp:4x4].,title="fig : " ] for rhp [ rhp:4x4 ] . ]
it is not hard to see that there is at most one @xmath60 that satisfies rhp [ rhp:4x4 ] .
so we have @xmath129 ( @xmath257 ) are uniquely determined by riemann
hilbert problem [ rhp:4x4 ] and .
notice that the jump matrices satisfy the symmetry @xmath258 where @xmath259 .
this implies the symmetry of the solutions @xmath260 which appears in ( * ? ? ?
* lemma 5.1(b ) ) for the hastings mcleod case .
( their result is for the tacnode rhp that is equivalent to rhp [ rhp:4x4 ] in the hastings mcleod case , see section [ subsec : tacnode_rhp ] . ) in this special case , there are additional symmetries with respect to complex conjugation which are not present in the general case .
as informed by an anonymous referee , our integral representation of the fundamental solution of the lax system by the fundamental solution of the flaschka newell lax pair is essentially similar to the representation of fundamental solution of the harnad
widom lax pair by the fundamental solution of the jimbo
miwa lax pair in ( * ? ? ?
* theorem 3.1 ) .
the idea of using a generalized laplace transform to produce a lax pair which is linear in the spectral variable from another which is quadratic in the spectral variable is in fact quite general and has been applied to other painlev equations as well , see @xcite . in the special case @xmath261
, @xmath247 is the solution to the lax system for the hastings mcleod solution to pii , and we refer to this special case as @xmath262 below . in this section
we discuss two occurrences of the entries of @xmath263 , one in the two - matrix model critical kernel and the other in the tacnode kernel in the non - intersecting brownian motion model .
originally these two kernels are expressed in terms of the tacnode rhp , which differs from the hastings mcleod case of our riemann
hilbert problem [ rhp:4x4 ] only by a constant matrix factor , see below .
the integral formulas for the entries of @xmath263 yield contour integral formulas for these two kernels .
consider the two - matrix model in which two @xmath6 random hermitian matrices @xmath264 and @xmath265 have the joint probability measure @xmath266 where @xmath62 and @xmath267 are potentials and @xmath159 is the coupling constant .
we concentrate on the distribution of eigenvalues of @xmath264 , which is a determinantal process and is thus characterized by a correlation kernel . in the case
that @xmath268 , @xmath269 and @xmath65 , the model is in the critical phase if @xmath270 and @xmath271 .
as @xmath65 , under the double scaling limit @xmath272 and @xmath273 , where @xmath274 and @xmath275 are constants , the correlation kernel for the eigenvalues of @xmath264 , at @xmath276 and @xmath277 converges to @xmath278 , whose formula is expressed by the tacnode rhp .
see @xcite for the derivation , and also @xcite .
similar to the critical kernel @xmath75 in the one - matrix model , the limiting kernel @xmath279 is believed to be universal , and it should occur in very general settings of the two - matrix model . if @xmath62 is a quadratic polynomial , the forthcoming paper @xcite will show that @xmath279 occurs for a large class of potentials @xmath267 .
consider @xmath280 non - colliding particles in brownian bridges , with diffusion parameter @xmath281 .
suppose the particles are in two groups , such that the left @xmath282 of them are in the first group and the right @xmath283 of them are in the second .
let particles in the first group start at @xmath284 at time @xmath76 , and end at @xmath284 at time @xmath285 , and let particles in the second group start at @xmath286 at time @xmath76 , and end at @xmath286 at time @xmath285 .
the particles in this model are a determinantal process , and their multi - time correlation functions are given by the multi - time correlation kernel . if @xmath287 and @xmath288 , the model is in the critical phase as @xmath65 , such that the right - most particle in the first group meets narrowly the left - most particle in the second group at time @xmath289 , and their trajectories touch each other like a tacnode . as @xmath65 , under the double scaling @xmath290 and @xmath291 , the multi - time correlation kernel at positions @xmath292 and @xmath293 and
times @xmath294 and @xmath295 converges to @xmath296 , which is expressed by the tacnode riemann
hilbert problem .
the derivation of @xmath297 was achieved by several groups of people : adler , ferrari and van moerbeke got a multi - time tacnode kernel formula with @xmath298 from a discrete random walk model @xcite ; delvaux , kuijlaars and zhang got a single time tacnode kernel formula from the nonintersecting brownian motion model @xcite ; johansson got a multi - time tacnode kernel formula with @xmath298 from the nonintersecting brownian motion model @xcite ; ferrari and vet generalized johansson s result for general @xmath299 @xcite .
follows that in @xcite , but with their @xmath20 replaced by @xmath300 .
the reason is that @xmath20 occurs in our paper everywhere as the argument of @xmath23 , the solution of .
we note that in @xcite , the author took the same change of notation , as explained in ( * ? ? ?
* formula 2.15 ) . ] the formulas of adler ferrari van moerbeke and johansson were both expressed in terms of airy resolvents , but quite different in structure .
they were later was proved to be equivalent @xcite .
delvaux kuijlaars
zhang s formula was expressed by the tacnode rhp ( their paper first defined the tacnode rhp , and the rhp is named thereby ) .
later delvaux showed in @xcite the equivalence of the results in @xcite and @xcite , and furthermore wrote the general multi - time tacnode kernel in the tacnode riemann - hilbert problem .
see also @xcite .
a variation of the model where the nonintersecting brownian bridges are on a circle was studied by the current authors in @xcite .
the tacnode rhp which is mentioned in sections [ subsec:2 mm ] and [ subsec : nibm ] was defined in @xcite , @xcite , and @xcite , with minor variations in generality and formality . the definition in ( * ? ? ?
* section 2.1 ) resembles the hastings mcleod case of our riemann
hilbert problem [ rhp:4x4 ] .
let us denote the solution to the rhp in ( * ? ? ?
* section 2.1 ) by @xmath301 .
then @xmath301 is defined on regions @xmath302 , with the same boundary conditions and asymptotics at @xmath32 , but the jump matrices are slightly different from those of @xmath263 .
it seems perplexing that two similar but different riemann
hilbert problems are associated to lax sytem ( in the hastings mcleod case ) , but the reason is simple : although the solutions @xmath303 are the unique solutions to satisfying the asymptotics throughout the sectors @xmath128 , as stated in proposition [ prop : duits - geudens ] , if the sectors are shrunk to @xmath304 , the asymptotics do not uniquely determine the solution in each sector .
the rhp [ rhp:4x4 ] and the tacnode rhp in @xcite require only asymptotics in sectors @xmath245 , so there is some freedom to choose the jump matrices corresponding to different solutions to .
the relation between the rhp [ rhp:4x4 ] and the tacnode rhp is as follows .
@xmath305 in the paper @xcite , kuijlaars found explicit formulas for the entries of the solution to the tacnode rhp in terms of airy functions and related operators .
he found six solutions to the differential equation with @xmath30 being the hastings mcleod solution to pii , which were labeled @xmath306 .
let us remark here that in the case @xmath261 , the solutions @xmath242 constructed in this paper agree with the ones constructed by kuijlaars up to sign
. specifically we have @xmath307 which follows from comparing ( * ? ? ?
* figure 2 ) with in light of theorem [ fund_solutions ] . we can write the critical kernel @xmath279 for two - matrix model in a contour integral formula where the integrand is expressed by entries of @xmath53 in , the solution to the flaschka newell lax pair .
let @xmath308 be a contour consisting of two infinite pieces : the first passing from @xmath309 to @xmath310 ; and the second passing from @xmath311 to @xmath312 , as pictured in figure [ fig : sigma_t ] .
also let @xmath313 be the contour @xmath314,\ ] ] oriented as shown in figure [ fig : sigma_t_deformed ] .
define the functions @xmath189 and @xmath190 on @xmath315 in a piecewise way : @xmath316 where @xmath317 we then have the following theorem . .
the rays make the angles @xmath318 with the real axis , and the horizontal segment is the interval @xmath319 $ ] . ] .
the rays make the angles @xmath318 with the real axis , and the horizontal segment is the interval @xmath319 $ ] . ]
[ critical ] the critical kernel in the two - matrix model of duits geudens ( * ? ? ?
* equation ( 2.15 ) ) can be written as @xmath320 where @xmath29 is given as @xmath321 this theorem is proved in section [ two_matrix_kernel_proof ] .
similarly we can write the tacnode kernel @xmath297 in a double contour integral formula with integrand given by entries of @xmath53 in .
we have the following theorem .
[ tacnode ] the tacnode kernel of ferrari
vet @xcite can be written as @xmath322 where @xmath323 this theorem is proved in section [ tacnode_kernel_proof ] . in the symmetric case @xmath324 , the formula , up to a rescaling ,
was derived by the current authors from a model of nonintersecting paths on the circle @xcite .
that model was not robust enough to produce the asymmetric tacnode kernel , and the above theorem is new for @xmath325 .
formulas and are analogous , but is more general in the sense that ( i ) it has a @xmath326 parameter and ( ii ) the @xmath327 parameters corresponding to @xmath159 in can be different . in @xcite , a more general two - matrix model as well as its dynamical version is considered , and is generalized to a formula containing @xmath328 parameters like . in this paper
we concentrate on the homogeneous painlev ii equation .
the general pii equation has a constant term : @xmath329 .
both the the flaschka newell lax pair and the @xmath1 lax system can be generalized to the inhomogeneous pii equation , and each appears in the kernel for a determinantal process .
the flaschka newell lax pair for the hastings mcleod solution of inhomogeneous pii equation occurs in the one - matrix model with logarithmic perturbation , see @xcite and a brief discussion in section [ subsubsec:1 mm ] . on the other hand , a @xmath79 riemann hilbert problem associated with the hastings mcleod solution of inhomogeneous pii equation occurs in the limiting critical correlation kernel for the nonintersecting squared bessel processes .
the ( nonintersecting ) squared bessel processes are in some sense a generalization of the ( nonintersecting ) brownian motions , and the limiting critical process for the nonintersecting squared bessel process is a `` hard - edge '' generalization of the tacnode process .
hence the aforementioned @xmath79 rhp , which is then called the hard - edge tacnode rhp , is a natural generalization of the tacnode rhp .
this hard - edge tacnode rhp is also associated with a lax system that is analogous to and more general than .
this hard - edge tacnode rhp is also related to a chiral two - matrix model .
see @xcite and @xcite . it is tempting to conjecture that our construction of the @xmath79 lax system from the @xmath18 flaschka newell lax pair to can be applied to the inhomogeneous case as well , thereby giving formulas for the hard - edge tacnode rhp in terms of solutions to the flaschka newell lax pair
. however , we have so far not been able to derive the relation in a straightforward way .
the algebraic result proposition [ diffeq ] is proved in section [ diff_eq_proof ] , and next the analytic result proposition [ thm : main_asy ] is proved in section [ asymptotics ] . then the main result theorem [ fund_solutions ]
is proved in section [ sec : proof_main ] , based on proposition [ thm : main_asy ] . as the applications of the main theorem , theorems [ critical ] and [ tacnode ] are proved in section [ sec : proof_of_contour_int_form ] .
we thank the anonymous referees for valuable comments , especially for pointing out related literature on isomonodromy deformations .
we also thank peter forrester for comments related to random matrix theory .
first we note that if @xmath330 satisfies equation , then the components @xmath331 are expressed in terms of @xmath332 by @xmath333 and then equation is reduced to the equations in @xmath332 : @xmath334m_2 , \label{int:13 } \\
m_2 '' = { } & -2\tau m_2'-r_2^{3/2}\sqrt{r_1}q({\sigma } ) c^2{\gamma}m_1 ' \notag \\ & + \big(cq({\sigma})^2 r_2 ^ 2+r_2 ^ 2 z+2r_2s_2-\tau^2\big)m_2+\frac{r_2^{3/2}\sqrt{r_1}{\gamma}}{c}\left[q({\sigma } ) \tau\left(\frac{1}{r_1 ^ 2}-\frac{1}{r_2 ^ 2}\right)-c^2q'({\sigma})\right]m_1 .
\label{int:13_2 } \end{aligned}\ ] ] conversely , if the four components of @xmath335 satisfy , , , and , then @xmath335 is a solution to .
suppose @xmath336 where @xmath189 and @xmath190 are any two @xmath55-dimensional vector - valued functions that make @xmath153 well defined on them .
we denote functions @xmath337 , @xmath338 , where @xmath339 such that for @xmath340 , with @xmath341 if @xmath342 and @xmath343
if @xmath344 , @xmath345 then by definition , @xmath346 on the other hand , also by definition , @xmath347 and we can evaluate @xmath348 as follows . since @xmath349 by , we have for @xmath340 , @xmath350 thus @xmath351 and we have @xmath352 using expressions , and , we check that holds .
similarly , we can check that holds .
next we show that if the @xmath189 and @xmath190 in are chosen to be the solutions @xmath163 and @xmath164 to , as in proposition [ diffeq ] , then identities and also hold .
consider first @xmath353 .
we have @xmath354 the first derivative of @xmath355 is already evaluated in , and the second derivative can be computed similarly .
we consider @xmath356 individually , and have @xmath357 now we use the property that @xmath358 is a solution to , and have @xmath359.\ ] ] we therefore have , using , @xmath360.\end{gathered}\ ] ] furthermore , using integration by parts , we have @xmath361 d\zeta + \begin{cases } -h^{(j)}_1(0 ) g_1(0 ) & j = 1 , 2 , \\
h^{(j)}_1(0 ) g_1(0 ) & j = 3 .
\end{cases}\ ] ] noting that @xmath362 we obtain that @xmath363 d\zeta,\ ] ] and then summing up the @xmath340 cases of , @xmath364 d\zeta \right].\end{gathered}\ ] ] since @xmath365 we have , using , that @xmath366 d\zeta = -\frac{3ic^2 a}{4 } \frac{d^2}{dz^2 } i_1 - ibc i_3(\zeta ) + i \left ( c + \frac{2z}{c } \right ) i_1.\ ] ] combining and , we solve that @xmath367 plugging and into , and using the formulas and for the coefficients , we have @xmath368 on the other hand , among the terms on the right - hand side of , @xmath369 are already evaluated in and , while @xmath370 can be evaluated similar to @xmath348 as @xmath371 it is not hard to see that the right - hand side of can also be expressed as the right - hand side of .
thus we prove . in the same way we can prove .
since parts [ enu : thm : main_asy_1 ] [ enu : thm : main_asy_6 ] are similar , we prove part [ enu : thm : main_asy_1 ] in detail in section [ subsec : part_1_of_1.3 ] , and explain how the proof is adapted to other cases in section [ subsec : other_parts_1.3 ] .
parts [ enu : thm : main_asy_2 ] and [ enu : thm : main_asy_3 ] can be proved by the computation as in part [ enu : thm : main_asy_1 ] . for parts [ enu : thm : main_asy_4 ] , [ enu :
thm : main_asy_5 ] , and [ enu : thm : main_asy_6 ] , although the same method works , the computation should be adjusted because @xmath372 in and @xmath373 have different asymptotic behavior at @xmath32 for even and odd @xmath374 . in the proof of parts [ enu : thm
: main_asy_1 ] , [ enu : thm : main_asy_2 ] , and [ enu : thm : main_asy_3 ] , for a computational reason that will be clear later , we take a change of variable @xmath375 where @xmath146 are defined in , @xmath159 and @xmath86 are defined in , and @xmath88 is defined in , and define the cubic polynomial @xmath376 as @xmath377 and @xmath147 is defined in , @xmath20 is defined in , and @xmath378 are defined in .
we note that the leading coefficient of @xmath376 satisfies @xmath379 .
we are interested in the asymptotics of the functions @xmath380 as @xmath130 in various sectors of the complex plane .
note that as @xmath381 at a certain angle , the parameter @xmath382 also approaches @xmath32 at the same angle .
thus we will consider the asymptotic behavior of the integrals which define @xmath380 as @xmath383 .
for brevity we will often use the notation @xmath384 rather than @xmath385 , and we trust the reader can keep in mind that @xmath384 is related to @xmath386 by a scaling and shift of fixed size . [
rem : meaning_of_f ] the function @xmath387 will be useful in the proof of parts [ enu : thm : main_asy_1 ] , [ enu : thm : main_asy_2 ] , and [ enu : thm : main_asy_3 ] , because the essential part of asymptotic analysis is the integrals on @xmath388 ( @xmath389 ) , where the contours @xmath390 are deformed , as explained later in this section .
the integrands on @xmath388 , although various in explicit formulas , all have the asymptotic behavior @xmath391 and under the change of variable , @xmath392 where @xmath149 is defined in .
recall the sectors @xmath302 defined in . in what follows we consider them on the @xmath19-plane and @xmath393-plane by replacing @xmath386 by @xmath19 and @xmath393 respectively in their definitions .
note that if @xmath394 where @xmath202 , then for large enough @xmath24 , @xmath384 defined in has its argument in a compact subset of @xmath395 .
below in the proof we assume that @xmath396 , \quad \delta ' > 0,\ ] ] even if @xmath397 is not large . to be concrete
, we may take @xmath398 . before giving the rigorous argument of the proof
, we describe the strategy . 1 .
[ enu : step_1 ] find the critical points of @xmath387 .
there are two of them , which are denoted as @xmath399 ( on the upper - half plane ) and @xmath400 ( on the lower - half plane ) . then denote @xmath401 2 .
[ enu : step_2 ] deform the contour @xmath402 defined in for @xmath389 , such that @xmath403 is a contour from @xmath312 to @xmath311 and passes through @xmath404 , @xmath405 is a contour from @xmath311 to @xmath406 and passes through @xmath400 , and @xmath407 is a contour from @xmath406 to @xmath312 and passes through @xmath400 .
then @xmath408 goes from a point on @xmath388 to @xmath409 , for @xmath389 .
furthermore , we require that for @xmath410 large enough , @xmath411 @xmath412 and @xmath413 where @xmath414 is a constant depending on @xmath56 .
note that we only define @xmath415 for @xmath416 and only define @xmath417 for @xmath418 . for the saddle point analysis , we require @xmath419 where @xmath420 is @xmath421 for @xmath422 and @xmath423 for @xmath344
. see figure [ fig : contours_part_1 ] for a schematic graph of the contours . the existence of the contours will be carefully justified later . + , @xmath424 and @xmath425 , in the proof of part [ enu : thm : main_asy_3 ] of proposition [ thm : main_asy ] .
@xmath426 and @xmath427 are not labelled , because their major parts overlap . ] + , @xmath424 and @xmath425 , in the proof of part [ enu : thm : main_asy_3 ] of proposition [ thm : main_asy ] .
@xmath426 and @xmath427 are not labelled , because their major parts overlap . ] + , @xmath424 and @xmath425 , in the proof of part [ enu : thm : main_asy_3 ] of proposition [ thm : main_asy ] .
@xmath426 and @xmath427 are not labelled , because their major parts overlap . ]
[ enu : step_3 ] use the standard saddle point analysis to prove the result . in particular , for all integrals on @xmath428 and @xmath429 ( @xmath430 in the entries of the @xmath79 matrix shown in , the integrands are expressed as @xmath431 where @xmath393 is related to @xmath19 by and @xmath387 is given in and .
thus @xmath432 are the saddle points giving the major contributions to the integrals .
we also show that the integrals over @xmath408 are negligible .
below we give the details of the steps .
for each @xmath433 , the equation @xmath434 has two solutions , @xmath435 by , we have @xmath436 and @xmath437 . for the construction of the contours
, we are going to use some planar dynamical system techniques .
we interpret the complex @xmath393-plane as a two - dimensional real coordinate plane by the standard relation @xmath438 .
the function @xmath439 is then harmonic in @xmath393 , or equivalently in @xmath440 , and it has only two critical points @xmath441 . by condition , we have @xmath442 consider the curve @xmath443 with differentiable parametrization @xmath444 such that @xmath445 this curve is the _ level curve through @xmath76_. since by @xmath441 are not on this level curve , the level curve can be extended to @xmath32 in both directions , and we assume it below .
then we have the following result on the directions that @xmath443 approaches @xmath32 .
a numerical plotting of @xmath443 is shown in figure [ fig : l+- ] .
this plot demonstrates the following result .
[ lem : level_curve_0 ] @xmath443 lies in @xmath446 , and it goes to @xmath32 in directions @xmath447 and @xmath448 . by the behavior of @xmath439 at @xmath32 , we know that a level curve , on which @xmath439 is finite , can only go to @xmath32 in six possible directions : @xmath449 , @xmath450 . for @xmath443
, we also know that the tangent direction at @xmath76 is @xmath451 and @xmath452 . in the remaining part of the proof , we consider three cases separately : [ enu : case_3_zero ] @xmath453 , [ enu : case_1_neg ] @xmath454 , and [ enu : case_2_pos ] @xmath455 . in case [ enu : case_3_zero ] , @xmath443 is exactly the real axis and the result of the lemma is obvious . in case [ enu :
case_1_neg ] , we have that one part of @xmath443 goes from @xmath76 to sector @xmath456 and the other part goes from @xmath76 to sector @xmath457 . we denote them @xmath458 and @xmath459 respectively .
we can see that @xmath458 does not go out of @xmath456 , because on one boundary of @xmath456 , @xmath460 , @xmath461 , and on the other boundary of @xmath456 , @xmath462 , @xmath463 .
similarly , @xmath459 does not go out of @xmath457 .
now the possible directions for @xmath458 to approach @xmath32 is limited to @xmath76 and @xmath464 .
next we exclude @xmath464 . for any @xmath414
, we have by direct calculation that @xmath465 for all @xmath466 and large enough @xmath467 , so for @xmath468 , @xmath469 if @xmath470 is large enough . thus @xmath458 goes to @xmath447 . by a similar reason
, @xmath459 goes to @xmath448 .
case [ enu : case_2_pos ] is converted to case [ enu : case_1_neg ] by the change of variables @xmath471 .
the next technical lemma is proved by straightforward calculation .
[ lem : monotone ] let @xmath467 be a big enough positive number . then on the circle @xmath472
, @xmath439 has three local maxima , @xmath473 , @xmath474 and @xmath475 , and three local minima , around @xmath476 , @xmath477 and @xmath478 . furthermore , on each arc @xmath479 between @xmath480 and @xmath481 , ( @xmath482 and @xmath483 ) , the value of @xmath439 is monotonic as @xmath393 moves along the arc .
now consider the level curves through @xmath399 and @xmath400 , which we denote by @xmath484 and @xmath485 respectively .
note that locally around @xmath441 , @xmath486 is the union of two smooth local level curves , and @xmath486 goes to @xmath32 as it extends along the four ends of the smooth local level curves
. see figure [ fig : l+- ] for a numerical plotting of these level curves .
the plotting demonstrates the results of the following two lemmas .
[ lem:4_crosses ] 1 .
[ enu : level_curves_through_crit_pt_1 ] the branches of @xmath484 go to @xmath447 , @xmath487 , @xmath488 and @xmath448 , and we denote them @xmath489 , @xmath490 , @xmath491 , and @xmath492 respectively .
[ enu : level_curves_through_crit_pt_2 ] the branches of @xmath485 go to @xmath447 , @xmath493 , @xmath494 and @xmath448 , and we denote them @xmath495 , @xmath496 , @xmath497 , and @xmath498 respectively . and @xmath499 .
( they overlap in the bottom part . )
the dotted curve is @xmath500 , and the dashed curves are @xmath485 . here
@xmath501 and @xmath502 . ] and
( they overlap in the bottom part . )
the dotted curve is @xmath500 , and the dashed curves are @xmath485 . here
@xmath501 and @xmath502 . ] and
( they overlap in the bottom part . )
the dotted curve is @xmath500 , and the dashed curves are @xmath485 . here
@xmath501 and @xmath502 . ]
we give the proof to part [ enu : level_curves_through_crit_pt_1 ] , and that to part [ enu : level_curves_through_crit_pt_2 ] is analogous . by the argument in the beginning of the proof of lemma [ lem : level_curve_0 ] , we know that the local level curves through @xmath399 go to infinity in the directions @xmath449 . by inequality
, we know that @xmath484 , the level curve through @xmath399 , does not intersect @xmath443 . since we assume that @xmath399 is in sector @xmath503 , we have that @xmath484 can only go to @xmath32 in directions above @xmath443 , that is , @xmath504 .
recall the notations in lemma [ lem : monotone ] .
let @xmath467 be a large enough positive number , then @xmath484 intersects the circle @xmath505 at four distinct points .
these intersecting points are on @xmath506 , since they are above @xmath443 . by the monotone property stated in lemma [ lem : monotone ] , these intersecting points are on distinct arcs @xmath479 , and then they are around @xmath467 , @xmath507 , @xmath508 and @xmath509 respectively .
thus @xmath484 goes to infinity in the four distinct directions .
[ lem : l+-_intersections ] 1 .
[ enu : lem : l+-_intersections_a ] @xmath489 intersects with the ray @xmath510 at a point , which we denote by @xmath511 ; @xmath490 stays in sector @xmath512 and has the ray as its asymptote , but does not intersect it .
[ enu : lem : l+-_intersections_b ] @xmath492 intersects with the ray @xmath513 at a point , which we denote by @xmath514 ; @xmath491 stays in sector @xmath512 and has the ray as its asymptote , but does not intersect it .
[ enu : lem : l+-_intersections_c ] @xmath495 intersects with the ray @xmath515 at a point , which we denote by @xmath516 ; @xmath496 stays in sector @xmath517 and has the ray as its asymptote , but does not intersect it .
[ enu : lem : l+-_intersections_d ] @xmath498 intersects with the ray @xmath518 at a point , which we denote by @xmath519 ; @xmath497 stays in sector @xmath517 and has the ray as its asymptote , but does not intersect it .
we prove parts [ enu : lem : l+-_intersections_a ] and [ enu : lem : l+-_intersections_b ] , and the proof to parts [ enu : lem : l+-_intersections_c ] and [ enu : lem : l+-_intersections_d ] is similar .
we note that @xmath439 is monotonically decreasing as @xmath393 moves to @xmath32 on either the ray @xmath510 or @xmath520 , which are the two boundaries of @xmath512 .
so @xmath521 intersects either ray at one point at most .
since @xmath489 goes from @xmath399 to @xmath447 , @xmath492 goes from @xmath399 to @xmath448 , and they do not intersect , we have that @xmath489 intersects with the ray @xmath510 , at a point , and @xmath492 intersects with the ray @xmath513 at a point .
the results for @xmath490 and @xmath491 are deduced by their asymptotic property in lemma [ lem:4_crosses ] and the fact that they do not intersect with the two rays .
we consider the images of @xmath523 and their components under the change of variables , which are @xmath524 the construction of @xmath525 is equivalent to the construction of @xmath526 .
the basic ingredient for the construction of the contours are the _ flow curves _ with respect to the gradient field @xmath527 , i.e. , smooth curves with parametrization @xmath444 such that @xmath528 through any point where @xmath527 does not vanish , there is a unique flow curve . but from the critical point @xmath399 or @xmath400 , there are four flow curves connecting to @xmath441 , with two flowing out of @xmath441 , that is , @xmath529 increases along the flow curves away from @xmath441 , and two flowing in @xmath441 , that is , @xmath529 increases along the flow curves towards @xmath441 . generically the flow lines connecting @xmath441 can be extended to @xmath32 , but in some special cases a flow line may connect @xmath399 and @xmath400 . if a flow curve extends to @xmath32 , then it can either go into @xmath32 in directions @xmath530 , @xmath531 , and @xmath310 , or go out of @xmath32 in directions @xmath311 , @xmath312 and @xmath406 . around @xmath441 , the flow curves into and out of @xmath441 alternate with the level curves @xmath532 .
part [ enu : level_curves_through_crit_pt_1 ] of lemma [ lem:4_crosses ] shows that one flow curve that flows in @xmath399 lies between @xmath489 and @xmath490 , and is from @xmath311 , and the other flow curve that flows in @xmath399 lies between @xmath491 and @xmath492 , and is from @xmath312 .
we denote them by @xmath533 and @xmath534 respectively for later use .
part [ enu : level_curves_through_crit_pt_2 ] of lemma [ lem:4_crosses ] shows that one flow curve that flows out of @xmath400 lies between @xmath496 and @xmath497 , and goes to @xmath406 , and the other flow curve that flows out of @xmath400 lies above @xmath495 and @xmath498 .
we denote them by @xmath535 and @xmath500 respectively .
the flow curve @xmath500 may end at @xmath311 , @xmath312 , or @xmath399 , depending on the argument of @xmath384 , but we do not need this piece of information .
below we construct @xmath536 and @xmath537 in the special case that @xmath538 and @xmath539 with @xmath540 .
this construction may seem impractical , since our interest is the limiting case that @xmath541 , or equivalently , @xmath383 .
actually if @xmath542 , @xmath543 may not satisfy the condition in part [ enu : thm : main_asy_1 ] of proposition [ thm : main_asy ] .
but the construction for general @xmath544 and @xmath384 , particularly for large @xmath384 , will be derived by a scaling transform of the special case .
[ [ construction - of - tildegamma0_1-cup - tildegamma0_3-for - tildea-1-and - lvert - tildec - rvert-1 ] ] construction of @xmath545 for @xmath501 and @xmath546 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + recall that @xmath489 intersects with the ray @xmath510 at @xmath511 .
since the flow curve @xmath533 lies above @xmath489 and extends to @xmath311 , it hits the ray @xmath547 at a point , which we denote by @xmath548 . by the property of flow curve , @xmath439 decreases as @xmath393 moves from @xmath399 to @xmath548 along @xmath533 .
a simple calculation shows that as @xmath393 moves to @xmath311 along the ray @xmath549 , @xmath439 is also decreasing .
similarly , @xmath534 hits the ray @xmath550 at a point , which we denote by @xmath551 .
@xmath439 decreases as @xmath393 moves along @xmath534 from @xmath399 to @xmath551 , and furthermore it decreases as @xmath393 moves to @xmath32 along the ray @xmath552 .
we define @xmath545 by the concatenation of ( i ) the ray @xmath552 , ( ii ) the part of @xmath534 between @xmath399 and @xmath551 , ( iii ) the part of @xmath533 between @xmath399 and @xmath548 , and ( iv ) the ray @xmath553 , with the orientation from @xmath312 to @xmath311 . for a numerical plotting of @xmath545 ,
see figure [ fig : gamma_0 ] .
* we have not constructed @xmath554 and @xmath555 individually yet , since the dividing point between them is not given . *
it seems that we can let @xmath545 simply be @xmath556 .
but then it is not easy to show that @xmath533 ( resp .
@xmath534 ) stay in @xmath557 ( resp .
@xmath558 ) , and then it is a problem to verify for @xmath426 and @xmath559 later .
[ [ construction - of - tildegamma2_1-cup - tildegamma2_3-and - tildegamma4_1-cup - tildegamma4_3-for - tildea-1-and - lvert - tildec - rvert-1 ] ] construction of @xmath560 and @xmath499 for @xmath501 and @xmath546 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + first we note that the flow curve @xmath535 stays in @xmath517 .
next we note that the level curves @xmath495 and @xmath498 , the line segment between @xmath516 and @xmath76 , and the line segment between @xmath519 and @xmath76 enclose a region , which we call @xmath561 . on the boundary of @xmath561
, @xmath439 keeps the same on the level curves and decreases as @xmath393 moves above to @xmath76 along either of the two line segments .
the flow curve @xmath500 goes into region @xmath561 .
letting @xmath562 be a small enough positive constant , we take @xmath563 as the point on @xmath500 such that @xmath564 , and denote the part of @xmath500 between @xmath400 and @xmath565 by @xmath566
. then there exists a smooth curve lying in region @xmath561 and connecting @xmath563 and @xmath567 , which we denote by @xmath568 , such that @xmath439 decreases monotonically as @xmath393 moves along @xmath568 from @xmath563 to @xmath567 .
similarly , there exists a smooth curve lying in region @xmath561 and connecting @xmath563 and @xmath569 , which we denote by @xmath570 , such that @xmath439 decreases monotonically as @xmath393 moves along @xmath570 from @xmath563 to @xmath571 . at last , by direct calculation , we find that @xmath439 decreases as @xmath393 moves along the ray @xmath572 to @xmath311 , and analogously that @xmath439 decreases as @xmath393 moves along the ray @xmath573 to @xmath312 .
thus we define @xmath560 by the concatenation of ( i ) the ray @xmath572 , ( ii ) the curve @xmath568 , ( iii ) the curve @xmath566 , and ( iv ) the flow curve @xmath535 , with the orientation from @xmath311 to @xmath406 .
similarly , we define @xmath499 by the concatenation of ( i ) the flow curve @xmath535 , ( ii ) the curve @xmath574 , ( iii ) the curve @xmath570 , and ( iv ) the ray @xmath573 , with the orientation from @xmath406 to @xmath312 . for a numerical plotting of @xmath560 and @xmath499 , see figure [ fig : gamma_24 ] .
note that they have overlap @xmath575 , which explains the overlap in the schematic figure [ fig : contours_part_1 ] .
[ [ construction - of - tildegamma0_2-tildegamma2_2-and - tildegamma4_2-for - tildea-1-and - lvert - tildec - rvert-1 ] ] construction of @xmath576 , @xmath577 and @xmath578 for @xmath501 and @xmath546 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the construction , we define the function @xmath579 and note that the leading coefficient of @xmath580 satisfies @xmath581 . notice that in the integral formulas , integrands on @xmath582 , @xmath583 and @xmath584 , although different , can all be written in the form of @xmath585 , and we have @xmath586 the ray @xmath587 intersects with @xmath588 at a point , which we denote by @xmath589 .
then we define @xmath576 to be the ray @xmath590 .
the following properties can be checked by direct computation : ( i ) @xmath576 is contained in @xmath503 , and ( ii ) @xmath591 let @xmath592 be a small enough positive number such that @xmath593 then the ray @xmath594 intersects with @xmath595 at a point , which we denote by @xmath596 , and the ray @xmath597 intersects @xmath499 at a point , which we denote by @xmath598 .
we define @xmath577 by the ray @xmath599 if @xmath600 , and define @xmath578 by the ray @xmath601 if @xmath602 .
then we have that @xmath603 below we check in the case that @xmath604 , and the case @xmath605 is analogous .
we first note that for all @xmath539 with @xmath606 and for all @xmath607 , @xmath608 so the value @xmath609 for @xmath393 on the ray @xmath610 satisfies @xmath611 on the other hand , @xmath612 since @xmath546 and @xmath600 .
so inequalities and imply in the case that @xmath604 .
[ rem : general_shape_of_gammas ] although our construction depends on the value of @xmath613 , by the compactness argument it is clear that for all @xmath384 that satisfy , there exists @xmath414 such that we can make : 1 . for @xmath614 , @xmath615 , @xmath616 , @xmath617 , and @xmath618 , @xmath619 , @xmath620 .
2 . for @xmath621 , @xmath622 , @xmath623 , @xmath624 , and @xmath625 , @xmath626 , @xmath627 .
3 . for @xmath628 , @xmath629 , @xmath630 , @xmath631 , and @xmath632 , @xmath633 , @xmath634 .
[ [ construction - for - the - contours - with - general - tildea - and - tildec ] ] construction for the contours with general @xmath544 and @xmath384 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + at last we consider the general case that @xmath544 is any positive number between @xmath76 and @xmath635 , and @xmath384 is any number such that @xmath636 .
we first construct the contours @xmath637 , @xmath638 and @xmath639 with respect to the parameters @xmath285 and @xmath640 in place of @xmath641 and @xmath384 , and then scale the contours in @xmath393-plane by the factor @xmath642 , that is , @xmath643 if and only if @xmath644 .
then @xmath545 is still through the point @xmath645 , and @xmath560 and @xmath499 are still through the point @xmath646 . our goal is to construct @xmath523 , and it can be done by a translation of @xmath647 according to .
note that after a translation , the contours @xmath390 may not lie in the same sectors as @xmath648 do .
but as @xmath649 , or equivalently , @xmath650 , the finite translation can be neglected . to be precise , if @xmath410 is large enough , then @xmath651 implies that @xmath613 satisfies , and @xmath652 ( resp .
@xmath653 $ ] ) implies that @xmath600 ( resp .
@xmath602 ) .
thus we derive results by properties stated in remark [ rem : general_shape_of_gammas ] for the contours @xmath654 .
first we compute @xmath655 as @xmath649 with @xmath656 . as discussed in the beginning of this section , this condition is equivalent to and @xmath650 .
we write @xmath657 where @xmath658 @xmath659 is defined in , and @xmath46 is the fundamental solution of that is expressed in @xmath51 and @xmath52 in figure [ fig:2x2phi ] .
we note that @xmath660 that is defined in . by and
, we have that for all @xmath661 , the asymptotic formula holds uniformly
. then we use the asymptotics of @xmath662 to derive that uniformly @xmath663 where @xmath393 depends on @xmath19 by .
similarly , @xmath664 , and by and , we also have that for all @xmath665 , the asymptotic formula holds uniformly . then similar to , we have uniformly @xmath666 we compute the second component of the @xmath78-dimensional vector @xmath667 in detail . the uniform convergence asymptotics and imply that @xmath668 according to the construction in section [ subsubsec : construction_of_contours ] , the contour @xmath545 has the following property that @xmath439 attains its unique maximum on it at @xmath399 , @xmath439 decreases fast as @xmath669 along it , and locally around @xmath399 it is the steepest descent contour for @xmath439 .
thus a standard application of the saddle point method yields @xmath670 hence @xmath671 similarly , we have for the fourth component of @xmath667 @xmath672 for the first and third components of @xmath667 , we can do the same computation , but we only need the estimates as follows @xmath673 next we consider the components of @xmath674 , and give some detail in the estimate of the first component . we note that @xmath675 . by and , we have that for all @xmath676 , the asymptotic formula holds uniformly
. then we use the asymptotics of @xmath677 to derive that uniformly @xmath678 where @xmath679 is defined in and @xmath680 note that the coefficients of @xmath387 and @xmath679 are given in terms of @xmath544 and @xmath384 .
if we denote @xmath681 then @xmath682 by the construction of @xmath683 and , we have that for all @xmath676 , or equivalently @xmath684 , there exists @xmath414 such that @xmath685 as @xmath541 , we have @xmath686 , and then for all @xmath676 , or equivalently @xmath684 , @xmath687 since @xmath688 is independent of @xmath386 and @xmath689 as @xmath669 , we have that as @xmath541 , if @xmath690 , then @xmath691 .
thus for @xmath692 and @xmath690 , there exists @xmath693 such that for large enough @xmath386 @xmath694 on the other hand , if @xmath692 and @xmath695 and @xmath541 , then @xmath679 is dominated by the cubic term , and it is clear that inequality still holds . by the approximation , using and that @xmath696 fast as @xmath669 along @xmath576 , we estimate that @xmath697 where @xmath693 is a constant , which can be taken to be the same as in . by the same method , we obtain the general result @xmath698 plugging , , and into , we derive that @xmath699 and prove part [ enu : thm : main_asy_1 ] of proposition [ thm : main_asy ] .
the proof of parts [ enu : thm : main_asy_2 ] and [ enu : thm : main_asy_3 ] is parallel to that of part [ enu : thm : main_asy_1 ] .
we also take the change of variables and compute the critical point @xmath441 as in .
but now @xmath700 and @xmath701 in the setting of part [ enu : thm : main_asy_2 ] , and @xmath702 and @xmath703 in the setting of part [ enu : thm : main_asy_3 ] .
also we use the method from planar dynamic systems to construct @xmath443 , @xmath486 , and the flow curves , and then @xmath654 , and finally @xmath525 and @xmath526 ( @xmath389 ) .
the results are shown in figures [ fig : contours_part_2 ] and [ fig : contours_part_3 ] .
an obvious @xmath704 rotational symmetry can be ovserved in figures [ fig : contours_part_1 ] , [ fig : contours_part_2 ] and [ fig : contours_part_3 ] , and it is a direct consequence of the symmetry among the settings in the three parts . at last ,
the saddle point analysis is applied , and the critical point @xmath404 yields the result @xmath705 , and the critical point @xmath706 yields the result @xmath707 .
the explicit leading terms of the @xmath708 factors are computed in the way of section [ subsubsec : saddle_point_anal ] .
similar to the proof to parts [ enu : thm : main_asy_1 ] , [ enu : thm : main_asy_2 ] and [ enu : thm : main_asy_3 ] , the essential part of the asymptotic analysis in the proof of parts [ enu : thm : main_asy_4 ] , [ enu : thm : main_asy_5 ] and [ enu : thm : main_asy_6 ] is the integrals on @xmath709 ( @xmath710 ) , where the contours @xmath390 are deformed , in a similar way to the deformation of contours shown in section [ subsec : part_1_of_1.3 ] for part [ enu : thm : main_asy_1 ] .
the integrands on @xmath709 , although various in explicit formulas , all have the asymptotic behavior @xmath711 which is comparable to in remark [ rem : meaning_of_f ] .
thus we take the change of variables , comparable to @xmath712 define @xmath713 and have @xmath714 here and below notations like @xmath393 , @xmath376 , @xmath544 , and @xmath384 , are different from their counterparts in section [ asymptotics ] but serve the same purpose in the proof .
we use the same notations to emphasize the identical use , while we trust that they do not lead to confusion .
then we find the critical points of @xmath387 , and denote them @xmath715 and then let @xmath716 we deform @xmath526 ( @xmath710 ) such that @xmath388 are through either @xmath404 or @xmath706 , satisfy @xmath717 and the deformed contours satisfy conditions analogous to . since the construction of the contours is different from the constructions in parts [ enu : thm : main_asy_1 ] , [ enu : thm : main_asy_2 ] , and [ enu : thm : main_asy_3 ] only in computational detail , we omit it , and only show figures [ fig : contours_part_4 ] , [ fig : contours_part_5 ] and [ fig : contours_part_6 ] to indicate the result of the construction . , @xmath718 and @xmath719 , in the proof of part [ enu : thm : main_asy_6 ] of proposition [ thm : main_asy ] .
@xmath720 and @xmath721 are not labelled , because their major parts overlap . ] , @xmath718 and @xmath719 , in the proof of part [ enu : thm : main_asy_6 ] of proposition [ thm : main_asy ] .
@xmath720 and @xmath721 are not labelled , because their major parts overlap . ] , @xmath718 and @xmath719 , in the proof of part [ enu : thm : main_asy_6 ] of proposition [ thm : main_asy ] .
@xmath720 and @xmath721 are not labelled , because their major parts overlap . ]
at last we apply the saddle point analysis , and find that the critical point @xmath404 yields the result @xmath722 , and the critical point @xmath706 yields the result @xmath723 , and prove the results .
the detailed computation is omitted .
the proof of theorem [ fund_solutions ] follows from combining the results of propositions [ prop : duits - geudens ] , [ diffeq ] , and [ thm : main_asy ] .
we will write the detailed proof of the formula for @xmath724 .
the proofs for @xmath725 are nearly identical , and we leave them to the reader . throughout the proof ,
we refer the reader to figure [ fig : a+dominance ] , which divides the complex plane into 12 sectors , each of size @xmath318 . within each of these sectors
, the asymptotic dominance scheme of the columns of the matrix @xmath113 is indicated .
for example , in the sector @xmath726 the sequence @xmath727 means that as @xmath381 , @xmath728 where we recall that @xmath729 are columns of @xmath238 .
it is easy to check this dominance scheme in each of the sectors from the definitions and the relations .
the rays which separate the different dominance schemes are called the _ stokes rays_. .
the complex plane is separated into sectors of angle @xmath318 . in each sector the relative dominance as @xmath381 of the columns of @xmath114 is indicated , e.g. the sequence 4 , 3 , 1 , 2 in the sector @xmath726 means that @xmath730 as @xmath130 .
the zigzag line indicates the branch cut for @xmath114 . ] .
the complex plane is separated into sectors of angle @xmath318 . in each sector the relative dominance as @xmath381 of the columns of @xmath114 is indicated , e.g. the sequence 4 , 3 , 1 , 2 in the sector @xmath726 means that @xmath730 as @xmath130 .
the zigzag line indicates the branch cut for @xmath114 . ] denote the columns of @xmath724 by @xmath731 , so that @xmath732 according to proposition [ prop : duits - geudens ] , @xmath733 as @xmath381 throughout the sector @xmath734 , so the dominance scheme in figure [ fig : a+dominance ] applies also to the columns of @xmath735 in this sector . notice that @xmath736 overlaps with 5 of the sectors shown in figure [ fig : a+dominance ] : @xmath737 ; @xmath738 ; @xmath739 ; @xmath740 and
@xmath741 we also refer to figure [ fig : prop1_3_summary ] , which summarizes the results of proposition [ thm : main_asy ] . in that figure the asymptotic behavior as @xmath381 of the solutions @xmath185
is matched to that of the functions @xmath742 . in the sector @xmath726
, we have that @xmath743 , and @xmath744 . according to the dominance scheme in this sector ,
this indicates that both @xmath655 and @xmath745 are _ recessive solutions _ to in this sector , i.e. they are not dominant over any other solutions to as @xmath130 in this sector except the trivial one .
recessive solutions are unique up to a constant factor .
since both @xmath655 and @xmath745 match the leading order behavior of @xmath746 in this sector , we must have @xmath747 by considering the recessive solutions in the sectors @xmath748 and @xmath749 we similarly obtain @xmath750 it remains only to find @xmath751 , which is the only column of @xmath735 which is not recessive in one of the sectors which overlap @xmath734 .
we look instead for a sector in which it is the least dominant possible amongst those overlapping @xmath734 .
notice in the sector @xmath752 , @xmath751 dominates @xmath745 , but is dominated by the other columns of @xmath735 . according to the figure [ fig : prop1_3_summary ] , in this sector we have @xmath753 .
since @xmath754 solves it is a linear combination of the rows of @xmath735 , @xmath755 for some constants @xmath756 . in the sector @xmath752 , @xmath753 and @xmath757 , which dominates @xmath758 there , so @xmath759 .
similarly , in the sector @xmath760 , @xmath753 , and @xmath761 , which dominates @xmath758 there , so @xmath762 .
furthermore , in the sector @xmath752 , we have @xmath763 , so @xmath764 by the comparison with the asymptotics of @xmath754 .
using as well , we obtain @xmath765 to find the value of @xmath766 , we can use the linear relation , @xmath767 and consider the asymptotics of @xmath754 in the sector @xmath768 the leading order behavior of each of the functions on the right - hand side of is given in proposition [ thm : main_asy ] ( see also figure [ fig : prop1_3_summary ] ) . inserting these asymptotics into gives , as @xmath130 with @xmath769 @xmath770 according to figure [ fig : a+dominance ] ,
@xmath746 is dominant as @xmath130 in this sector , so becomes @xmath771 or equivalently , in the sector @xmath772 , @xmath773 comparing and , and noting that @xmath743 dominates @xmath774 in the sector @xmath772 , we find that @xmath775 .
thus gives the formula for @xmath751 : @xmath776 combining , , and gives the formula for @xmath735 in theorem [ fund_solutions ] .
the formulas for the rest of the solutions @xmath777 can be obtained in a similar manner . always three out of the four columns of @xmath129 can be identified as solutions to which are recessive in some part of @xmath128 .
these recessive solutions can be identified with one of the functions @xmath237 using proposition [ thm : main_asy ] , or equivalently referencing figure [ fig : prop1_3_summary ] .
there is one column which is never recessive in @xmath128 , but it can be determined using the linear relations in a manner similar to how @xmath778 was determined above . in figure
[ fig : a - dominance ] we include the dominance scheme for the columns of @xmath239 , which should be consulted when considering @xmath779 , @xmath780 and @xmath781 .
in this section , we assume @xmath47 is the hastings mcleod solution to the pii equation . then the solutions @xmath782 to the lax pair that are defined in section [ subsec : flaschka - newell_lax ] are also assumed to be associated with the hastings mcleod solution @xmath47 . these solutions
to are related by the jump conditions which are in turn determined by the parameters @xmath44 . in this section
we assume @xmath58 , associated with the hastings mcleod solution .
recall that for @xmath783 , @xmath784 defined in are vector - valued functions with the parameters @xmath92 .
the vectors @xmath140 also depend on a solution to the pii equation by definition , and we assume it to be the hastings mcleod solution @xmath30 in this section . by the symmetry of equations , and the identity @xmath785 that holds because @xmath786 for the hastings mcleod solution ( see figure [ fig:2x2phi ] ) , we have that @xmath787 it implies that , with functions @xmath788 defined in and , @xmath789 at last , we note that the differential equation implies that @xmath790 the duits guedens critical kernel for the two - matrix model was derived in @xcite , and our proof of theorem [ critical ] is based on the presentation in @xcite .
the critical kernel is described in terms of the tacnode rhp with parameters ( * ? ? ?
* formula ( 2.41 ) ) @xmath791 in ( * ? ? ?
* formulas ( 4.5 ) and ( 4.9 ) ) , two vector - valued functions @xmath792 and @xmath793 , depending on parameters @xmath794 and @xmath159 , are introduced as the linear combinations of the columns of the solution to the tacnode rhp . by the relation between the tacnode rhp and rhp [ rhp:4x4 ]
, we have in our notations @xmath795 the critical kernel @xmath796 has the expression ( * ? ? ?
* theorem 2.9 and formula ( 4.13 ) ) @xmath797,\end{gathered}\ ] ] and for the proof of theorem [ critical ] we also need ( * ? ? ?
* equation ( 4.13 ) ) @xmath798 now we use the integral formulas of @xmath799 and @xmath800 to express the functions @xmath801 and @xmath802 in terms of the entries of @xmath803 which are defined in section [ subsec : flaschka - newell_lax ] .
first we consider @xmath800 . by the definition given by , we have that each component of @xmath655 is expressed by a sum of integrals on @xmath426 , @xmath683 , and @xmath559 .
since @xmath655 is associated to the hastings mcleod solution , and then @xmath58 , by for any component of @xmath655 , the integrand on @xmath683 vanishes , and the integrands on @xmath426 and @xmath559 are identical .
similarly , we have that each component of @xmath804 is expressed by a sum of integrals on @xmath805 , @xmath806 , and @xmath721 , such that the integrand on @xmath806 vanishes and the integrands on @xmath805 and @xmath721 are identical .
furthermore , with @xmath308 as in figure [ fig : sigma_t ] , we can deform @xmath403 into the upper half of @xmath807 and @xmath808 into the lower half of @xmath807 , both with reversed orientation .
thus we can write , after expressing the integrands in by , , and , @xmath809 where the functions @xmath810 and @xmath811 the contour @xmath308 are defined in equation and in figure [ fig : sigma_t ] , respectively . by the same argument
, we have that @xmath812 and @xmath813 have similar but slightly more complicated formulas as integrals on @xmath807 . here
we note that the contour @xmath807 can be replaced by @xmath814 where the definition for @xmath189 and @xmath190 is still given by .
next we use the integral formulas of @xmath815 and @xmath816 to express the functions @xmath817 and @xmath818 .
similar to the argument for @xmath801 and @xmath802 , because @xmath815 and @xmath816 are associated to the hastings mcleod solution , the integrands on @xmath819 and @xmath820 for the integral formula of the @xmath821-th component of @xmath815 are identical to the integrands on @xmath427 and @xmath822 respectively for the integral formula of the @xmath821-th component of @xmath816 , and the integrand on @xmath720 for the integral formula of the @xmath821-th component of @xmath815 is the negative of the integrand on @xmath415 for the integral formula of the @xmath821-th component of @xmath816 . using the contours @xmath390 shown in figure [ fig : six_contours ] , we find that the integrals on @xmath720 and @xmath819 cancel the integrals on @xmath415 and @xmath427 respectively when we compute @xmath823 . hence by , , and , we derive analogous to and , @xmath824 where @xmath825 and @xmath826 are defined in . similarly , @xmath827 and
@xmath828 have similar but slightly more complicated formulas as integrals on the imaginary axis . plugging in , , and into ,
we find @xmath829 next we make the change of variable @xmath830 .
we note that the contour for @xmath91 changes direction under this transform .
hence by , becomes @xmath831 the right - hand side of the above equation can also be expressed as a derivative with respect to @xmath794 . indeed , using we find @xmath832 we also note that in , if we deform the integral contour @xmath807 into @xmath814 , with the definition of @xmath833 and @xmath834 given in when @xmath835 , the integral on the right - hand side does not change .
therefore in we can replace @xmath807 with @xmath814 .
let us denote @xmath836 where @xmath837 as in .
notice that without the deformation of @xmath807 into @xmath814 , the integral in would be singular when the contours cross , and would have to be be regarded as a principal value integral .
the deformation removes the singularity . then and imply that @xmath838 if we can show @xmath839 then follows from .
below we prove and .
we need the asymptotic behavior of @xmath840 , the fundamental solutions to , as @xmath841 , when @xmath47 is the hastings - mcleod solution to , or equivalently , when @xmath58 in .
the result was obtained in ( * ? ? ?
* section 6 ) , and we summarize it below .
suppose @xmath842 .
we define the @xmath18 matrix - valued function @xmath843 , where we follow the notational convention in @xcite and suppress the dependence on @xmath20 , by @xmath844 where @xmath845 are regions of the complex plane as shown in figure [ fig : last_rhp ] , and the power function @xmath846 has a cut on @xmath319 $ ] taking the branch such that @xmath847 as @xmath381 .
then @xmath848 as @xmath541 , and it satisfies the jump condition as shown in figure [ fig : last_rhp ] , cf .
* figure 6.18 ) . , where @xmath849 and @xmath850 .
] then as observed in @xcite , the rhp for @xmath843 converges formally to a rhp on the interval @xmath851 $ ] with jump matrix @xmath852 , and this rhp has a simple solution : @xmath853 the function @xmath854 has a cut on @xmath319 $ ] taking the branch of the fractional power so that @xmath855 as @xmath381 . by constructing local parametrices at @xmath285 and @xmath856 , the convergence can be made rigorous , as discussed in ( * ? ? ?
* section 3 ) . by standard argument
, we derive the following result : [ lem : last ] fix @xmath414 . as @xmath841 , @xmath857
uniformly for all @xmath858 . here
if @xmath386 is on the jump curve , then the uniform convergence holds for both @xmath859 and @xmath860 . by lemma [ lem : last ] and the asymptotics of @xmath861 implied by it , we use the explicit formulas and for @xmath862 and @xmath863 and derive that as @xmath864 , @xmath862 and @xmath863 vanishes exponentially . in a similar way , we have that @xmath827 and @xmath828 also vanishes exponentially .
to estimate @xmath865 , as @xmath864 , it is better to replace the contour @xmath807 by @xmath866 that depends on @xmath867 , which is simply the contour @xmath814 scaled by factor @xmath868 : @xmath869.\end{gathered}\ ] ] then by direct computation based on and , we can find that @xmath870 and @xmath871 vanish exponentially as @xmath864 .
similarly , we can show that @xmath870 and @xmath871 also do .
hence we prove by plugging in the exponential vanishing property stated above to .
similarly , if we replace @xmath814 by @xmath866 when we evaluate the double contour integral in , we find it vanishes exponentially as @xmath864 , and holds .
the detail is left to the reader .
thus theorem [ critical ] is proved .
the multi - time correlation kernel for the tacnode process was first derived in the form of airy resolvents , and it was presented in the most general form in @xcite , where the notation @xmath296 is defined .
we use this notation in theorem [ tacnode ] . in @xcite
, it was shown that the kernel can also be represented in the form of the tacnode rhp . in this section
we prove theorem [ tacnode ] based on the formula of in @xcite .
first we recall that in ( * ? ? ?
* rhp 2.1 in section 2.1.1 ) a riemann
hilbert problem is defined , and it is essentially the same as the tacnode rhp in @xcite that is discussed in section [ subsec : tacnode_rhp ] . in (
* section 2.1.1 ) , a matrix @xmath872 is defined with parameters @xmath92 , and without much difficulty we can check that @xmath873 where @xmath874 is the solution to rhp [ rhp:4x4 ] in sector @xmath512 , associated to the hastings mcleod .
then the vector function @xmath875 defined in ( * ? ? ?
* formula ( 2.9 ) ) , which is the sum of the first and second columns of @xmath872 , become @xmath876 then the functions @xmath877 ( @xmath878 ) , which are denoted in ( * ? ? ? * formula ( 2.26 ) ) and are related to the first two components of @xmath875 by ( * ? ? ?
* formula 4.37 ) , becomes @xmath879 @xmath880 where @xmath881 are components of @xmath882 , @xmath88 is defined in and where @xmath299 is the parameter in the correlation kernel formula @xmath297 .
then by ( * ? ? ? * theorem 2.9 ) , ( noting that our notation @xmath297 means the same as @xmath883 in @xcite if @xmath20 and @xmath300 are related by ( * ? ?
* formula ( 2.15 ) ) , or equivalently ) , @xmath884 where , with @xmath885 as specified in , @xmath886 the vector - valued function @xmath875 is in the same form as @xmath792 defined in with more general parameters . thus similar to and ,
we can write @xmath887 and @xmath888 as integrals on contour @xmath807 , and then have by and @xmath889 d\zeta , \\ \widehat{p}_2(z ; s , \tau ) = { } & \frac{1}{\pi ( 1 + 1/\sqrt{\lambda})^{1/3 } } \int_{\sigma_{\operatorname{tac } } } g(\zeta ; s ) \exp \left [ -\frac{4 i}{3 } \frac{1 - \sqrt{\lambda}}{1 + \sqrt{\lambda } } \zeta^3 + \frac{4\tau}{c^2 } \zeta^2 + \left ( \frac{2iz}{c } + i \frac{1 - \sqrt{\lambda}}{1 + \sqrt{\lambda } } s \right ) \zeta \right ] d\zeta,\end{aligned}\ ] ] where @xmath189 and @xmath190 are defined in . in this section ,
we need a technical lemma : [ lem : tacnode_vanishing ] let @xmath414 and @xmath890 .
then there exists @xmath891 such that for all @xmath892 , and @xmath893 @xmath894 we prove , and the proof of is analogous .
the airy resolvent formulas in @xcite yield ( see ( * ? ? ?
* formulas ( 2.38 ) and ( 2.39 ) ) ) @xmath895 where @xmath896 and the definitions of @xmath897 and @xmath898 are given in ( * ? ? ?
* formulas ( 2.18 ) and ( 2.19 ) ) . for all
@xmath892 , @xmath897 and @xmath898 decays at the speed comparable to the airy function , so that that @xmath899 and @xmath900 are bounded and is proved .
then we can write the kernel @xmath901 as @xmath902 ds,\end{gathered}\ ] ] where the change of order of integrations is justified by lemma [ lem : tacnode_vanishing ] .
next we make the change of variable @xmath903 .
note now that the contour @xmath308 is invariant under this transform .
hence by , becomes @xmath904 ds.\end{gathered}\ ] ] by , we have @xmath905 \\ = \exp \left ( i\frac{(1 - \sqrt{{\lambda}})}{1 + \sqrt{{\lambda } } } s(u - v ) \right ) \left [ f(u ; s)g(v ; s)+\frac{f(u ; s)g(v ; s)}{\sqrt{{\lambda } } } \right ] .
\end{gathered}\ ] ] using this identity , we can write as @xmath906 \,ds.\end{gathered}\ ] ] performing the integration in @xmath794 and using @xmath907 , we derive @xmath908,\end{gathered}\ ] ] given that @xmath909 we thus have that is implied by lemma [ lem : tacnode_vanishing ] .
thus we finish the proof of theorem [ tacnode ] .
in order to present the coefficient matrices of , it is convenient to introduce several notations which were used in @xcite .
below we define several parameters which depend on @xmath912 , @xmath913 , and @xmath159 .
these are the same notations given in ( * ? ? ?
* theorem 6.2 ) up to the rescaling @xmath914 .
the quantities @xmath88 and @xmath89 are the ones which were defined in ; @xmath23 and @xmath90 , are the pii solution and its derivative ; @xmath91 is the pii hamiltonian defined in ; and all painlev functions are evaluated at the point @xmath29 which is defined in .
other notations in this appenix are independent of the rest of the paper .
in particular the parameters @xmath275 and @xmath147 given below are _ not _ the same ones as in .
we use these symbols to match the notation of @xcite and we trust it will not cause confusion . define the parameters @xmath915 , \quad \tilde{b } = \frac{{\gamma}}{cr_1 \sqrt{r_1 r_2}}\left[\left(s_1 ^ 2+\frac{2r_1 ^ 2\tau}{r_1 ^ 2+r_2 ^ 2}\right)q-\frac{uq+q'}{c}\right ] , \\ { \beta}= \frac{{\gamma}}{cr_2 \sqrt{r_1 r_2}}\left[\left(s_2 ^ 2-\frac{2r_2 ^ 2\tau}{r_1 ^ 2+r_2 ^ 2}\right)q-\frac{uq+q'}{c}\right ] , \quad \tilde{{\beta}}= \frac{1}{cr_1 \sqrt{r_1 r_2}{\gamma}}\left[\left(s_1 ^ 2-\frac{2r_1 ^ 2\tau}{r_1 ^ 2+r_2 ^ 2}\right)q-\frac{uq+q'}{c}\right ] , \\\end{gathered}\ ] ] as well as @xmath916 also introduce the notations @xmath917+b\left(- c -\frac{r_2}{r_1 } \tilde{c } + \frac{(r_1 ^ 2+r_2 ^ 2)\tau}{r_1}\right ) \\
-\frac{2 s_2 d}{r_1},\end{gathered}\ ] ] @xmath918+\tilde{b}\left(- \frac{r_1}{r_2}c - \tilde{c } + \frac{(r_1 ^ 2+r_2 ^ 2)\tau}{r_2}\right ) \\ -\tilde{d}^2 d+\frac{r_1}{r_2 } c^2 \tilde{d } -\frac{2 s_1 \tilde{d}}{r_2}.\end{gathered}\ ] ] then the matrices @xmath910 , @xmath911 , and @xmath267 in are given below .
note that besides the scaling of @xmath159 , the solution to the lax system in @xcite differs from ( the hastings mcleod case of ) the solution to lax system by a scalar factor @xmath919 , so our lax system is slightly different from ( * ? ? ?
* formulas ( 6.21)(6.24 ) ) in @xmath82 and @xmath267 .
@xmath920 athanassios s. fokas , alexander r. its , andrei a. kapaev , and victor yu .
novokshenov . , volume 128 of _ mathematical surveys and monographs_. american mathematical society , providence , ri , 2006 .
the riemann - hilbert approach .
r. garnier .
sur des quations diffrentielles du troisime ordre do nt lintgrale gnrale est uniforme et sur une classe dquations nouvelles dordre suprieur do nt lintgrale gnrale a ses points critiques fixes . , 29:1126 , 1912 .
j. harnad , c. a. tracy , and h. widom .
hamiltonian structure of equations appearing in random matrices . in _ low - dimensional topology and quantum field theory ( cambridge , 1992 )
_ , volume 315 of _ nato adv .
_ , pages 231245 .
plenum , new york , 1993 . | we consider two lax systems for the homogeneous painlev ii equation : one of size @xmath0 studied by flaschka and newell in the early 1980 s , and one of size @xmath1 introduced by delvaux kuijlaars zhang and duits
geudens in the early 2010 s .
we prove that solutions to the @xmath1 system can be derived from those to the @xmath0 system via an integral transform , and consequently relate the stokes multipliers for the two systems . as corollaries we are able to express two kernels for determinantal processes as contour integrals involving the flaschka newell lax system : the tacnode kernel arising in models of nonintersecting paths , and a critical kernel arising in a two - matrix model . |
the study of low - mass electron - positron pairs in heavy ion collisions is a powerful tool to investigate properties of the newly discovered strongly coupled quark gluon plasma .
results obtained by the na45 and na60 experiments at cern @xcite show an excess of pairs produced in the mass region below 1gev / c@xmath3 . the phenix experiment at rhic measured the low mass dilepton region @xcite , however
a detailed study of dileptons is very difficult without suppression of the combinatorial background coming from electron - positron pairs with small opening angle .
their primary source is the decays of @xmath4 mesons and @xmath5-conversions .
the main purpose of the phenix upgrade with the hbd detector is the rejection of close @xmath6 pairs by two orders of magnitude .
an extensive simulation done for the hbd @xcite demonstrated that the only way to suppress the dilepton combinatorial background without loosing the signal requires electron identification .
the @xmath7 rejection factor must be 100 with an electron detection efficiency of 90% .
the hbd must cover solid angle @xmath820% larger than the nominal phenix acceptance .
detectors based on cherenkov radiation are the most effective choice to meet such requirements .
cherenkov detector consists of a radiator in which particles shall follow straight trajectories .
this competes with the high resolution tracking which benefits from the increase of the magnetic field .
the phenix detector @xcite has two magnetic field coils which allow to increase the magnetic field for better tracking or compensate it within the radius of 50 - 60 cm from the interaction point and keeping the initial angles of all tracks .
an important element of the classic ring imaging cherenkov detector is the focusing mirrors located at the end of the radiator . in phenix configuration
the mirror can not be used because of the geometrical constrains .
therefore it was decided to replace the mirrors with the light sensitive detectors placing them in the path of all particles produced in the collisions .
the hbd detection unit thus must be sensitive to uv light and blind to all hadrons traversing it .
it must also keep its radiation budget below 3% of the radiation length and fit into @xmath9 cm . a crucial decision for the cherenkov counter
is the choice of the photo - sensor . amid a variety of choices
the only considered were those which fit the radiation length budget , i.e. solid film and gaseous / aerosol sensors .
the second option , however , requires to separate radiator and amplification into two different gaseous volumes .
it significantly increases the radiation length and besides that the window serves as an auxiliary source of cherenkov radiation .
the first option is preferable in spite of the same gas must be used as the cherenkov radiator and the electron multiplication media .
the most widely used film photocathode is the @xmath1 reaching 80% quantum efficiency ( q.e . ) in extreme ultra - violet ( e.u.v . )
part of the spectra , see fig .
[ fig : qe ] . to match the photocathode q.e .
one has to use the gas which is transparent in the e.u.v . that limits the choice of gases to mixtures of carbofluorides and noble gases having the deepest e.u.v . cut - off wavelength . among them
the pure @xmath2 transparent up to 11.5 ev and with the refraction index of n=1.00062 is considered the primary choice .
based on the r&d carried out during the development stage @xcite the electrons can be effectively extracted into @xmath2 atmosphere with up to 85% efficeincy compare to vacuum ( full symbols in fig .
[ fig : qe ] ) .
the same parameter in pure @xmath10 of never exceeds 60% .
photocathode in the vacuum ( open symbols ) and in the pure @xmath2 atmosphere ( full symbols).[fig : qe ] ] the main challenge is to build a detector sensitive to the smallest possible charge produced by the cherenkov photons and not sensitive to the much larger ionization coming from hadrons traversing it .
the difference between the two processes is in the localization of the primary charges . in the first case they appear on the surface of the photocathode and in the second inside the gas volume .
the detector concept utilizing primary charge localization was suggested by y. giomataris and g. charpak @xcite in 1991 .
electrons extracted from the photocathode surface undergo full amplification in a parallel plate detector whereas electrons produced inside the detector sensitive volume are only partially amplified .
a prototype of such detector was built and studied by the group of t. hemmick at stony brook university in 1996 @xcite .
the major problem with this approach are the avalanche photons shining back to the photocathode .
since the gas used in the detector is the same gas as used in the radiator no admixtures can be used to quench the feedback .
the advent of the new detector , the gas electron multiplier ( gem ) , invented by f.sauli in 1997 @xcite made it possible to overcome this difficulty . among other features of the gem is that the photon feedback is blocked by geometry of the detector . by evaporating a thin layer of @xmath1 onto the gem surface one can convert a gem into highly efficient semi - transparent photocathode .
it was demonstrated @xcite that the electric field required for electron amplification inside the gem holes is sufficient to extract electrons form any point of the surface of the gem and direct them into the nearest hole .
electrons produced by the ionizing particles in the gas volume above the photocathode can be removed by an inversely biased drift field making the detector insensitive to ionizing radiation .
this scheme is shown in fig [ fig : gems ] .
, title="fig : " ] , title="fig : " ] three gems are used in the detector to provide the gas amplification of @xmath11 @xcite . without a focusing mirror the cherenkov photons on the detector
populate a circle not a ring thus the single electron detection does not improve the pattern recognition . in the hbd
the signal is collected by hexagonal pads of a size slightly smaller than the size of the circle ] .
this maximizes the signal collected on a single pad and provides extra rejection power to the hadrons .
ionization is always localized in one single pad while primary electrons producing cherenkov light spread it over at least two adjacent pads .
the separation between a single electron and a close electron pair is done by the analyzing the amplitudes . according to the simulation
the number of photoelectrons produced by a primary electron is 36 .
such signal can be reliably distinguished from twice that number produced by a dilepton pair with a small opening angle .
the hbd detector is shown in the right panel of fig . [
fig : gems ] @xcite .
the detector consists of two half - cylinder volumes made of honeycomb fiberglass sandwich panels .
the top and bottom sectors , sitting outside the phenix fiducial acceptance are used for services and the six inner sectors are covered by a single piece kapton film with 1152 pads printed on it .
the film also serves as an additional gas seal .
particles enter the detector trough a 0.12 mm aluminized mylar window . 12 gas amplification modules as described above , are located on the inner side of each hbd vessel .
each module is 27@xmath1223 cm@xmath3 .
gems used in them are subdivided into 28 hv modules to reduce energy stored in them in case of electric discharge .
all gems in a module are powered with a single hv power supply through a resistive chain .
the hbd will be operational in phenix in 2007 physics run .
authors acknowledge support from the department of energy , nsf ( u.s.a . ) and us - israel bsf .
work of the speaker is supported by the goldhaber fellowship at bnl with funds provided by brookhaven science associates .
a. agakichev , et al , _ phys .
letters _ * 75 * , ( 1995 ) 1272 ; _ phys . letters _ * b422 * , ( 1998 ) 405 .
m. floris
et al , _ nucl - ex_/0606023 a. toia et al , _ nucl - ex_/0510006 z. fraenkel , et al , www.phenix.bnl.gov/www/tpchbd/weizmann@[email protected] k. adcox , et al ,
_ nim _ * a499 * , ( 2003 ) 469 - 479 z. fraenkel , et al ,
_ nim _ * a546 * , ( 2005 ) 466 - 480 y. giomataris and g. charpak , _ nim _ * a310 * , ( 1991 ) 589 - 595 r. pisani , et al , _ nim _ * a400 * , ( 1997 ) 243 - 254 f. sauli , et al , _
nim _ * a386 * , ( 1997 ) 531 - 534 a. kozlov , et al , _ nim _ * a523 * , ( 2004 ) 345 i. tserruya , et al , _ nim _ * a563 * , ( 2006 ) 333 | the phenix collaboration has designed a conceptually new hadron blind detector ( hbd ) for electron identification in high density hadron environment . the hbd will identify low momentum electron - positron pairs to reduce the combinatorial background in the mass region below 1 gev / c@xmath0 .
the hbd shall be installed in phenix during the 2007 physics run .
the hbd is a windowless proximity focusing cherenkov detector with a radiator length of 50 cm , @xmath1 photocathode and three layers of gas electron multipliers ( gem ) for gas amplification .
pure @xmath2 serves both as a radiator and as a detector gas .
the radiation budget of the device is less than 3% of a radiation length .
address = brookhaven national laboratory , upton , ny , 11973 , usa |
a solution of the einstein equation that connects distant points of space
time was introduced by @xcite .
this `` einstein rosen bridge '' was the first solution to later be referred to as a wormhole .
initially , this type of solution was just a trivial or teaching example of mathematical physics .
however , @xcite proved that some wormholes are `` traversable '' ; i.e. , space and time travel can be achieved by passing through the wormholes .
they also showed that the existence of a wormhole requires exotic matter that violates the null energy condition .
although they are very exotic , the existence of wormholes has not been ruled out in theory .
inspired by the morris
thorne paper , there have been a number of theoretical works ( see @xcite and references therein ) on wormholes .
the curious natures of wormholes , such as time travel , energy conditions , space time foams , and growth of a wormhole in an accelerating universe have been studied .
although there have been enthusiastic theoretical studies , studies searching for real evidence of the existence of wormholes are scarce .
only a few attempts have been made to show the existence or nonexistence of wormholes . a possible observational method that has been proposed to detect or exclude the existence of wormholes is the application of optical gravitational lensing .
the gravitational lensing of wormholes was pioneered by @xcite , who inferred that some wormholes show `` negative mass '' lensing .
they showed that the light curve of the negative - mass lensing event of a distant star has singular double peaks .
several authors subsequently conducted theoretical studies on detectability @xcite .
another gravitational lensing method employing gamma rays was proposed by @xcite , who postulated that the singular negative - mass lensing of distant active galactic nuclei causes a sharp spike of gamma rays and may be observed as double - peaked gamma - ray bursts .
they analyzed baste data and set a limit for the density of the negative - mass objects .
there have been several recent works @xcite on the gravitational lensing of wormholes as structures of space time .
such studies are expected to unveil lensing properties directly from the space time structure .
one study @xcite calculated the deflection angle of light due to the ellis wormhole , whose asymptotic mass at infinity is zero .
the massless wormhole is particularly interesting because it is expected to have unique gravitational lensing effects .
the ellis wormhole is expressed by the line element @xmath3 where @xmath4 is the throat radius of the wormhole .
this type of wormhole was first introduced by @xcite as a massless scalar field .
later , @xcite studied this wormhole and proved it to be traversable .
the dynamical feature was studied by @xcite , who showed that gaussian perturbation causes either explode to an inflationary universe or collapse to a black hole .
@xcite showed that the tchyon condensate can be a source for the ellis geometry . in this paper
, we derive the light curve of lensing by the ellis wormhole and discuss its detectability . in section 2 ,
we discuss gravitational lensing by the ellis wormhole in the weak - field limit .
the light curves of wormhole events are discussed in section 3 .
the validity of the weak - field limit is discussed in section 4 .
the optical depth and event rate are discussed in section 5 .
the results are summarized in section 6 .
magnification of the apparent brightness of a distant star by the gravitational lensing effect of another star was predicted by @xcite .
this kind of lensing effect is called `` microlensing '' because the images produced by the gravitational lensing are very close to each other and are difficult for the observer to resolve .
the observable effect is the changing apparent brightness of the source star only .
this effect was discovered in 1993 @xcite and has been used to detect astronomical objects that do not emit observable signals ( such as visible light , radio waves , and x rays ) or are too faint to observe .
microlensing has successfully been applied to detect extrasolar planets @xcite and brown dwarfs @xcite .
microlensing is also used to search for unseen black holes @xcite and massive compact halo objects @xcite , a candidate for dark matter .
the gravity of a star is well expressed by the schwarzschild metric .
the gravitational microlensing of the schwarzschild metric @xcite has been studied in the weak - field limit . in this section
, we simply follow the method used for schwarzschild lensing .
figure [ fig1 ] shows the relation between the source star , the lens ( wormhole ) , and the observer .
the ellis wormhole is known to be a massless wormhole , which means that the asymptotic mass at infinity is zero .
however , this wormhole deflects light by gravitational lensing @xcite because of its curved space time structure .
the deflection angle @xmath5 of the ellis wormhole was derived by @xcite to be @xmath6 where @xmath7 is the closest approach of the light . in the weak - field limit ( @xmath8 ) ,
the deflection angle becomes @xmath9 the angle between the lens ( wormhole ) and the source @xmath10 can then be written as @xmath11 where @xmath12 , @xmath13 , @xmath14 , and @xmath15 are the distances from the observer to the lens , from the observer to the source , and from the lens to the source , and the impact parameter of the light , respectively . in the asymptotic limit , schwarzschild lensing and massive janis
winnicour ( jnw ) wormhole lensing @xcite have the same leading term of @xmath16 .
therefore , the lensing property of the jnw wormhole is approximately the same as that of schwarzschild lensing and is difficult to distinguish .
as shown in equation ( [ eqn : defl ] ) , the deflection angle of the ellis wormhole does not have the term of @xmath16 and starts from @xmath17 .
this is due to the massless nature of the ellis wormhole and indicates the possibility of observational discrimination from the ordinary gravitational lensing effect . in the weak - field limit
, @xmath15 is approximately equal to the closest approach @xmath7 . for the ellis wormhole , @xmath18 .
we thus obtain @xmath19 the light passing through the other side of the lens may also form images . however , equation ( [ eqn : proj ] ) represents deflection in the wrong direction at @xmath20 .
thus , we must change the sign of the deflection angle : @xmath21 it would be useful to note that a single equation is suitable both for @xmath22 and @xmath20 images in the schwarzschild lensing .
however , such treatment is applicable only when the deflection angle is an odd function of @xmath7 .
if the source and lens are completely aligned along the line of sight , the image is expected to be circular ( an einstein ring ) .
the einstein radius @xmath23 , which is defined as the radius of the circular image on the lens plane , is obtained from equation ( [ eqn : proj ] ) with @xmath24 as @xmath25{\frac{\pi}{4}\frac{d_l d_{ls}}{d_s } a^2}. \label{eqn : re}\ ] ] the image positions can then be calculated from @xmath26 and @xmath27 where @xmath28 is the angle between the image and lens , and @xmath29 is the angular einstein radius . using reduced parameters @xmath30 and @xmath31 , equations ( [ eqn : imgeq ] ) and ( [ eqn : imgeq2 ] ) become simple cubic formulas : @xmath32 and @xmath33 as the discriminant of equation ( [ eqn : poly ] ) is @xmath34 , equation ( [ eqn : poly ] ) has two conjugate complex solutions and a real solution : @xmath35 with , @xmath36{\frac{\hat{\beta}^3}{27 } + \frac{1}{2 } \pm \sqrt{\frac{1}{4}\left(1 + \frac{2 \hat{\beta}^3}{27}\right)^2 - \frac{\hat{\beta}^6}{27 ^ 2}}}.\ ] ] the real positive solution corresponds to the physical image .
the discriminant of equation ( [ eqn : poly2 ] ) is @xmath37 .
thus it has a real solution if @xmath38{27/4}$ ] : @xmath39 where , @xmath40{\frac{\hat{\beta}^3}{27 } - \frac{1}{2 } \pm \sqrt{\frac{1}{4}\left(1 - \frac{2 \hat{\beta}^3}{27}\right)^2 - \frac{\hat{\beta}^6}{27 ^ 2 } } } , \ ] ] with @xmath41 .
this solution corresponds to a physical image inside the einstein ring . for @xmath42{27/4}$ ] , equation ( [ eqn : poly2 ] ) has three real solutions .
however , two of them are not physical because they do not satisfy @xmath43 .
only the solution @xmath44 corresponds to a physical image inside the einstein ring .
figure [ fig2 ] shows the calculated images for source stars at various positions on a straight line ( source trajectory ) .
the motion of the images are similar to those of the schwarzschild lensing .
table [ tbl-1 ] shows the einstein radii and angular einstein radii for a bulge star ( @xmath45 and @xmath46 are assumed ) and a star in the large magellanic cloud ( lmc , @xmath47 and @xmath48 are assumed ) for various throat radii .
the detection of a lens for which the einstein radius is smaller than the star radius ( @xmath49 ) is very difficult because most of the features of the gravitational lensing are smeared out by the finite - source effect .
thus , detecting a wormhole with a throat radius less than @xmath50 from the galactic gravitational lensing of a star is very difficult .
the light curve of schwarzschild lensing was derived by @xcite .
the same method of derivation can be used for wormholes .
the magnification of the brightness @xmath51 is @xmath52 where @xmath53 and @xmath54 are magnification of the outer and inner images , @xmath55 and @xmath56 correspond to outer and inner images , respectively .
the relation between the lens and source trajectory in the sky is shown in figure [ fig3 ] .
the time dependence of @xmath57 is @xmath58 where @xmath59 is the impact parameter of the source trajectory and @xmath60 is the time of closest approach .
@xmath61 is the einstein radius crossing time given by @xmath62 where @xmath63 is the transverse velocity of the lens relative to the source and observer .
the light curves obtained from equations ( [ eqn : a ] ) and ( [ eqn : hbeta ] ) are shown as thick red lines in figure [ fig4 ] . the light curves corresponding to schwarzschild lensing are shown as thin green lines for comparison .
the magnifications by the ellis wormhole are generally less than those of schwarzschild lensing .
the light curve of the ellis wormhole for @xmath64 shows characteristic gutters on both sides of the peak immediately outside the einstein ring crossing times ( @xmath65 ) .
the depth of the gutters is about 4% from the baseline .
amazingly , the star becomes fainter than normal in terms of apparent brightness in the gutters .
this means that the ellis wormhole lensing has off - center divergence . in conventional gravitational lensing theory
@xcite , the convergence of light is expressed by a convolution of the surface mass density .
thus , we need to introduce negative mass to describe divergent lensing by the ellis wormhole . however , negative mass is not a physical enitity .
as the lensing by the ellis wormhole is convergent at the center , lensing at some other place must be divergent because the wormhole has zero asymptotic mass . for @xmath66 , the light curve of the wormhole
has a basin at @xmath60 and no peak . using these features , discrimination from schwarzschild lensing
can be achieved .
equations ( [ eqn : re ] ) and ( [ eqn : te ] ) indicate that physical parameters ( @xmath12 , @xmath4 , and @xmath63 ) are degenerate in @xmath61 and can not be derived by fitting the light - curve data .
this situation is the same as that for schwarzschild lensing . to obtain or constrain these values , observations of the finite - source effect @xcite or parallax @xcite are necessary .
the detectability of the magnification of the star brightness depends on the timescale .
the einstein radius crossing time @xmath61 depends on the transverse velocity @xmath63 .
there is no reliable estimate of @xmath63 for wormholes .
here we assume that the velocity of the wormhole is approximately equal to the rotation velocity of stars ( @xmath67 ) if it is bound to the galaxy .
if the wormhole is not bound to our galaxy , the transverse velocity would be much higher .
we assume @xmath68 @xcite for the unbound wormhole .
table 2 shows the einstein radius crossing times of the ellis wormhole lensings for the galactic bulge and lmc in both bound and unbound scenarios . as the frequencies of current microlensing observations are limited to once every few hours , an event for which the timescale is less than one day is difficult to detect .
to find very long timescale events ( @xmath69 ) , long - term monitoring of events is necessary .
the realistic period of observation is @xmath70 .
thus , the realistic size of the throat that we can search for is limited to @xmath71 both for the galactic bulge and lmc if wormholes are bound to our galaxy . if wormholes are unbound , the detection is limited to @xmath72 .
first , we consider the outer image . in the previous section , we applied the weak - field approximation to the impact parameter @xmath15 and the deflection angle @xmath5 .
as previously mentioned , the impact parameter @xmath15 is written as @xmath73 the condition to neglect the second term is @xmath74 . as the image is always outside the einstein ring , @xmath75 from the deflection angle , we obtain a similar relation from equation ( [ eqn : defl ] ) : @xmath76 the values of @xmath4 and @xmath23 in table [ tbl-1 ] show that the weak - field approximation is suitable for @xmath77 in the galactic microlensing .
more generally , @xmath78 is derived from equation ( [ eqn : re ] ) for @xmath79 .
this means that @xmath23 is much greater than @xmath4 if @xmath80 .
thus , the weak - field approximation is suitable if the throat radius is negligibly small compared with the source distance . for the inner image
, the higher - order effect is expected to be greater than that for the outer image . however , the contribution of the inner image to the total brightness is small and decreases quickly with @xmath57 ( @xmath81 for @xmath82 and @xmath83 for @xmath84 ) . on the other hand ,
the absolute value of the corresponding @xmath85 does not decrease as quickly ( @xmath86 for @xmath87 and @xmath88 for @xmath89 ) .
thus the contribution of the higher order effect of the second image to the total brightness is expected to be small .
another possibility of deviation from the weak - field approximation is the contribution of relativistic images .
recently , gravitational lensing in the strong - field limit @xcite has been studied for lensing by black holes . in this limit , light rays are strongly bent and wound close to the photon sphere . as a result ,
a number of relativistic images appear around the photon sphere .
however , it has been shown that there is no photon sphere @xcite in ellis wormhole lensing .
therefore , there is no contribution of relativistic images to the magnification in ellis wormhole lensing .
we thus conclude that the weak - field hypothesis is a good approximation unless the throat radius is comparable to the galactic distance .
the probability of a microlensing event to occur for a star is expressed by the optical depth @xmath90 : @xmath91 where @xmath92 is the number density of wormholes as a function of the line of sight .
here we simply assume that @xmath92 is constant ( @xmath93 ) : @xmath94 ^{2/3}d d_l , \\ & = & \sqrt[3]{\frac{\pi^5}{2 ^ 4 } } n a^{4/3 } d_s^{5/3 } \int_0 ^ 1 \left[x ( 1- x)\right]^{2/3 } dx , \\ &
\approx & 0.785 n a^{4/3 } d_s^{5/3}.\end{aligned}\ ] ] the event rate expected for a source star @xmath95 is calculated as @xmath96{2\pi } n v_t d_s^{4/3 } a^{2/3 } \int_0 ^ 1 \sqrt[3]{x ( 1 - x ) } dx , \\ & \approx & 0.978 n v_t a^{2/3 } d_s^{4/3}. \label{eqn : gm3}\end{aligned}\ ] ] there is no reliable prediction of the number density of wormholes .
several authors @xcite have speculated that wormholes are very common in the universe , at least as abundant as stars .
even if we accept such speculation , there are still large uncertainties in the value of @xmath97 because the distribution of wormholes is not specified . here , we introduce two possibilities .
one is that wormholes are bound to the galaxy and the number density is approximately equal to the local stellar density .
the other possibility is that wormholes are not bound to the galaxy and are approximately uniformly distributed throughout the universe . for the bound hypothesis , we use @xmath98 , where @xmath99 is the local stellar density in the solar neighborhood , @xmath100 , and @xmath101 is the average mass of stars .
we use @xmath102 , a typical mass of an m dwarf ; _
i.e. _ , the dominant stellar component in the galaxy . for the unbound hypothesis
, we assumed that the number density of the wormholes is the same as the average stellar density of the universe .
the stellar density of the universe is estimated assuming that the fraction of baryonic matter accounted for by star is the same as that of the solar neighborhood
. then we obtain @xmath103 , where @xmath104 is the critical density , @xmath105 is the baryon density of the universe divided by the critical density , and @xmath100 and @xmath106 are the local star and local baryon densities , respectively . using these values , we calculated the optical depths and event rates for bulge and lmc lensings .
table [ tbl-3 ] presents the results for the bulge lensings . in an ordinary schwarzschild microlensing survey , observations are made of more than 10 million stars .
thus , we can expect approximately @xmath107 events in a year .
however , the situation is different in a wormhole search . as mentioned previously ,
the magnification of wormhole lensing is less than that of schwarzschild lensing , and a remarkable feature of wormhole lensing is the decreasing brightness around the einstein radius crossing times .
past microlensing surveys have mainly searched for stars that increase in brightness .
the stars monitored are those with magnitudes down to the limiting magnitude or less .
however , we need to find stars that decrease in brightness in the wormhole search .
to do so , we need to watch brighter stars .
therefore , far fewer stars can be monitored than in an ordinary microlensing survey .
furthermore , the detection efficiency of the wormhole is thought to be less than that for schwarzschild lensing because of the low magnification .
here we assume that the effective number of stars monitored to find a wormhole is @xmath108 . to expect more than one event in a survey of several years
, @xmath95 must be greater than @xmath109 .
the values in table [ tbl-3 ] indicate that the detection of wormholes with @xmath110 is expected in the microlensing survey of the galactic bulge in the case of the bound model .
the results of the optical depths and the event rates for lmc lensing are presented in table [ tbl-4 ] . on the basis of the same discussion for bulge lensing
, we expect @xmath111 to find a wormhole .
the event rates expected for lmc lensing are greater than those for bulge lensing .
we expect detection of a wormhole event if @xmath112 for the bound model .
if no candidate is found , we can set upper limits of @xmath95 and/or @xmath90 as functions of @xmath61 . to convert these values to physical parameters ( @xmath97 and
@xmath4 ) requires the distribution of @xmath63 . right now
, there is no reliable model of the distribution except for using the bound or unbound hypothesis . on the other hand ,
the event rates for the unbound model are too small for the events to be detected . in past microlensing surveys @xcite ,
large amounts of data have already been collected for both the bulge and lmc fields .
monitoring more than @xmath108 stars for about @xmath113 can be achieved by simply reanalyzing the past data .
thus , discovery of wormholes can be expected if their population density is as high as the local stellar density and @xmath114 .
such wormholes of astronomical size are large enough for humans to pass through .
thus , they would be of interest to people discussing the possibility of space time travel .
if no candidate is found , the possibility of a rich population of large - throat wormholes bound to the galaxy can be ruled out .
such a limit , however , may not affect existing wormhole theories because there is no prediction of the abundance .
however , theoretical studies on wormholes are still in progress .
the limit imposed by observation is expected to affect future wormhole theories . on the other hand ,
the discovery of unbound wormholes is very difficult even if their population density is comparable to that of ordinary stars . to discover such wormholes ,
the monitoring of a much larger number of stars in distant galaxies would be necessary .
for example , @xmath115 and @xmath116 for the m101 microlensing survey ( @xmath117 ) if the throat radius is @xmath2 . to carry out such a microlensing survey , observation from space is necessary because the resolving of a large number of stars in a distant galaxy is impossible through ground observations .
only the ellis wormhole has been discussed in this paper .
there are several other types of wormholes @xcite for which deflection angles have been derived .
these wormholes are expected to have different light curves . to detect those wormholes , calculations of their light curves are necessary .
the method used in this paper can be employed only when we know the analytic solutions of the image positions .
if no analytic solution is found , the calculation must be made numerically .
the gravitational lensing of the ellis wormhole is solved in the weak - field limit .
the image positions are calculated as real solutions of simple cubic formulas .
one image appears on the source star side and outside the einstein ring .
the other image appears on the other side and inside the einstein ring .
a simple estimation shows that the weak - field hypothesis is a good approximation for galactic microlensing if the throat radius is less than @xmath0 .
the light curve derived has characteristic gutters immediately outside the einstein ring crossing times .
optical depths and event rates for bulge and lmc lensings are calculated for simple bound and unbound hypotheses .
the results show that the bound wormholes can be detected by reanalyzing past data if the throat radius is between @xmath118 and @xmath2 and the number density is approximately equal to the local stellar density .
if the wormholes are unbound and approximately uniformly distributed in the universe with average stellar density , detection of the wormholes is impossible using past microlensing data . to detect unbound wormholes ,
a microlensing survey of distant galaxies from space is necessary .
we would like to thank professor hideki asada of hirosaki university , professor matt visser of victoria university , and professor tomohiro harada of rikkyo university for discussions and their suggestions concerning this study .
we thank professor philip yock of auckland university for polishing our manuscript .
finally , we would like to thank an anonymous referee who pointed out the existence of a second image .
alcock , c. , et al .
1993 , , 365 , 621 alcock , c. , et al .
1995 , , 454 , l125 alcock , c. , et al .
2000 , , 542 , 281 alcock , c. , et al .
2001 , , 550 , l169 aubourg , e. et al .
1993 , , 365 , 623 beaulieu j .- p . ,
2006 , , 439 , 437 bennett , d. p. et al .
2002 , , 579 , 639 bond , i. a. 2004 , , 606 , l155 bogdanov , m. b. & cherepashchuk , a. m. 2008 , ap&ss 317 , 181 calchi novati , s. , mancini , l. , scarpetta , g. , & wyrzykowski , .
2009 , , 400 , 1625 chetouani , l. & clment , g. 1984 , gen .
16 , 111 clment , g. 1984 , j. theor .
23 , 335 cramer , j. g. , forward , r. l. , morris , m. s. , visser , m. , benford , g. , & landis , g. a. 1995 , , 51 , 3317 das , a. , & kar , s. 2005 , class .
grav . , 22 , 3045 dey , t. k. & sen , s. 2008 , mod . phys
a 23 , 953 einstein , a. , & rosen , n. 1935 , phys .
48 , 73 einstein , a. 1936 , science 84 , 506 ellis , h. g. 1973 , j. math .
, 14 , 104 gould , a. et al .
2009 , , 698 , l147 krasnikov , s. 2000 , , 62 , 084028 liebes , s. 1964 , phys .
, 133 , 835 lobo f. s. n. 2007 , arxiv:0710.4474v1 [ gr - qc ] mao , s. et al .
2002 , , 329 , 349 morris , m. s. , & thorne , k. s. 1988 , am .
j. ph . , 56 , 395 nemiroff , r. j. & wickramasinghe , w. a. d. t. 1994 , , 424 , l21 nandi , k. k. , zhang , y .- z .
, & zakharov , a. v. 2006 , , 74 , 024020 paczyski , b. 1986 , , 304 , 1 perlick , v. 2004 , , 69 , 064017 poindexter , s. , et al . , 2005 , , 633 , 914 rahaman , f. , kalam , m. , & chakraborty , s. 2007 , chin .
45 , 518 refsdal , s. 1964 , , 128 , 295 safonova , m. , diego , f. t. , & romero , g. e. 2002 , , 65 , 023001 schneider , p. , ehlers , j. , & falco , e. e. , 1992 , gravitational lenses , springer - verlag new york , isbn 0 - 387 - 97097 - 3 , p. 175 shatsuki , a. a. 2004 , astr .
48 , 525 shinkai , h. & hayward , s. a. 2002 , , 66 , 044005 sumi , t. , et al . , 2003 , 591 , 204 tisserand , p. , et al . 2007 , a & a , 469 387 torres , d. f. , romero , g. e. , & anchordoqui , l. a. 1998 , , 58 , 123001 udalski a. , et al . , 1993 ,
aca 43 , 289 virbhadra , k. s. & ellis , g. f. r. 2000 , 62 , 084003 visser , m. `` lorentzuan wormholes : from einstein to hawking '' , american institute of physics , new york , 1995 wyrzykowski , .
2009 , , 397 , 1228 ( top left ) , @xmath119 ( top right ) , @xmath120 ( bottom left ) , and @xmath121 ( bottom right ) .
thick red lines are the light curves for wormholes .
thin green lines are corresponding light curves for schwarzschild lenses.[fig4],title="fig : " ] ( top left ) , @xmath119 ( top right ) , @xmath120 ( bottom left ) , and @xmath121 ( bottom right ) .
thick red lines are the light curves for wormholes .
thin green lines are corresponding light curves for schwarzschild lenses.[fig4],title="fig : " ] ( top left ) , @xmath119 ( top right ) , @xmath120 ( bottom left ) , and @xmath121 ( bottom right ) .
thick red lines are the light curves for wormholes .
thin green lines are corresponding light curves for schwarzschild lenses.[fig4],title="fig : " ] ( top left ) , @xmath119 ( top right ) , @xmath120 ( bottom left ) , and @xmath121 ( bottom right ) .
thick red lines are the light curves for wormholes .
thin green lines are corresponding light curves for schwarzschild lenses.[fig4],title="fig : " ] | a method to calculate light curves of the gravitational microlensing of the ellis wormhole is derived in the weak - field limit . in this limit , lensing by the wormhole produces one image outside the einstein ring and one other image inside .
the weak - field hypothesis is a good approximation in galactic lensing if the throat radius is less than @xmath0 .
the light curves calculated have gutters of approximately 4% immediately outside the einstein ring crossing times .
the magnification of the ellis wormhole lensing is generally less than that of schwarzschild lensing .
the optical depths and event rates are calculated for the galactic bulge and large magellanic cloud fields according to bound and unbound hypotheses .
if the wormholes have throat radii between @xmath1 and @xmath2 , are bound to the galaxy , and have a number density that is approximately that of ordinary stars , detection can be achieved by reanalyzing past data .
if the wormholes are unbound , detection using past data is impossible . |
the advent of reliable redshift - independent distance estimators lead to an enormous growth of activity in the field of measuring and interpreting the peculiar velocities of galaxies .
the velocity field can in particular be fruitfully investigated by means of perturbation analysis .
one important result is the , @xmath0-dependent , velocity - density relationship that follows from linear theory ( see e.g. peebles 1980 ) .
moreover , taking advantage of the fact that perturbation theory predicts that the rotational part of the velocity field vanishes , bertschinger & dekel ( 1989 ) developed the non - parametric potent method in which the local cosmological velocity field is reconstructed from the measured line - of - sight velocities ( bertschinger et al .
1990 ) .
then , via the velocity - density relationship it is possible to estimate the value of @xmath1 , where it is assumed that the bias of the galaxies can be simply represented by a linear bias factor @xmath2 ( see the review paper of dekel 1994 and references therein ) .
there are however other methods that have been proposed that uses _ intrinsic _ properties of the large scale velocity field to estimate @xmath0 .
for example , nusser & dekel ( 1993 ) proposed to use a reconstruction method assuming gaussian initial conditions to constrain @xmath0 , while dekel & rees ( 1994 ) use voids to achieve the same goal
. another approach has been proposed by bernardeau et al .
( 1995 ) and bernardeau ( 1994a ) based on the use of statistical properties of the divergence of the locally smoothed velocity field .
preliminary comparisons of the analytical predictions with numerical simulations ( bernardeau 1994b , juszkiewicz et al .
1995 , okas et al .
1995 ) yielded encouraging results .
however , such a comparison is complicated due to the fact that the velocities are only known at , non - uniformly distributed , particle locations . here , i report recent results obtained by bernardeau & van de weygaert ( 1996 ) and van de weygaert et al .
( 1996 ) addressing specifically the issue of the discrete nature of the velocity sampling .
to start with , let me remind a few analytical results obtained from perturbation theory applied to the large - scale velocity field .
i consider the statistical properties of the one - point _ volume
_ averaged velocity divergence , @xmath3 , ( in units of the hubble constant ) when the average is made with a top - hat window function . although the first analytical results that have been obtained dealt with the values of high order moments of its distribution function ( bernardeau et al .
1995 , bernardeau 1994a ) it is more convenient to consider its global shape . particularly interesting is the @xmath4 case ( where @xmath5 is the index of the power spectrum ) for which there is a simple analytical fit for the pdf ( bernardeau 1994b ) , @xmath6/\kappa^{1/2}+ [ \lambda-1]/\lambda^{1/2})^{-3/2 } \over \kappa^{3/4 } ( 2\pi)^{1/2 } \sigma_{\theta } } \exp\left[-{\theta^2\over 2\lambda\sigma_{\theta}^2}\right ] { \rm d}\theta,\ ] ] with @xmath7 where @xmath8 is the variance of the distribution .
the resulting shape of the pdf is shown in fig .
it is worth noting that the @xmath0 dependence shows up in the shape and position of the peak ( given by @xmath9 indicated on the figure ) and by the position of the large @xmath3 cutoff ( @xmath10 on the figure ) .
a similar property to the later one was also found by dekel & rees ( 1994 ) using the zeldovich approximation .
it allows indeed to constraint @xmath0 from the largest expanding void . in principle
all these features can be used to constrain @xmath0 .
before trying to apply these ideas to the data we have extensively checked these features in numerical simulations . in usual numerical analysis , the velocity field is defined with a momentum average of the closest particles on grid points .
this method , however , yields very poor results when a subsequent _ volume _ average is required .
the main reason is that the two smoothing scales , grid size and smoothing radius , can not be very different from each other .
the problem is in fact to define properly the velocity _ field _ ( the velocity at any location in space ) from the velocities of a given set of discrete and sparse points .
this is what the methods proposed by van de weygaert and i are designed for . in the first method we propose the local velocity to be given by the _ velocity of the closest particle_. for defining the velocity in the whole space one
has then to divide space in cells , each containing a particle ( of a simulation for instance ) , and so that any point inside the cell is closer to it than to any other particle .
this partition is called the _ voronoi _ tessellation . in fig 2 .
i present a 2d sketch of such a partition : the solid lines form the voronoi tessellation of the filled circles representing the particles .
then , from the initial assumption that the velocity is constant in the voronoi cells , the velocity gradients ( in particular the divergence ) are localized on the walls .
they have actually a surface density given by ( @xmath11 is the peculiar velocity ) @xmath12 where @xmath13 is the unit vector normal to the wall and going outward of the cell ( see fig .
the local smoothed velocity divergences are then just given by the sum of the fraction of all walls that are within a given sphere of radius @xmath14 ( thick solid line in the figure ) multiplied by the value of the divergence on each wall , @xmath15 in the delaunay method the local velocity is supposed to be given by a _
linear combination of the velocities of the four neighbors_. if in 1d it is easy to identify the closest neighbors , this is no longer the case in 2d or 3d .
the solution is once again provided by the voronoi tessellation or rather its dual , the _ delaunay _ tessellation .
this is the triangulation in which the particles are connected together when they share a wall in the voronoi tessellation ( dashed lines in fig .
because of the properties of the voronoi tessellation , the delaunay triangulation satisfies a criterion of compactness : tetrahedra that are defined in such a way ( or triangles in 2d ) are as less elongated as possible .
this is a crucial property for doing a subsequent linear interpolation in the tetrahedra , since it ensures that the linear interpolation will be as good as possible .
so having identified the four neighbors @xmath16 of a point @xmath17 its velocity is assumed to be given by @xmath18 where @xmath19 are the barycentric weights of the points @xmath16 at the position @xmath17 , @xmath20 the velocity gradients are then uniform in the tetrahedra and the local smoothed divergence is given by a sum over all the tetrahedra that intersect a given sphere ( gray area in fig .
2 ) . in practice to use these methods we have to select only a fraction of the points provided by the @xmath21-body codes .
this is done in such a way that the largest voids retain as many particles as possible .
for details see bernardeau and van de weygaert ( 1995 ) .
the tessellations are then built using the codes developed by van de weygaert ( 1991 ) .
so far we have applied these methods to two different numerical simulations .
one kindly provided by h. couchman ( 1991 ) with cdm initial condition for @xmath22 and a pm simulation with @xmath23 and a power law spectrum of index @xmath4 ( van de weygaert et al .
1996 ) here i just present a very significant figure showing the scatter plots of the local divergences measured in 8000 different locations ( fig .
3 ) with the various available methods .
when the delaunay method is compared to the previous grid method the correlation is very noisy and there is even a systematic error in the variance . when the voronoi and the delaunay methods are compared to each other , no such features are seen and there is a perfect correlation between the two estimations .
this gives us a good confidence in our methods .
the resulting pdf - s are presented in fig .
4 . they show a remarkable agreement between the numerical estimations and the predictions of eq .
moreover the left panel demonstrates that the statistics of the velocity divergence is indeed sensitive to @xmath0 ( comparison of the dashed line with the solid line ) .
an interesting remark to make is that , in principle , it is not only possible to measure @xmath0 but it must also be possible to test the gravitational instability scenario . the pdf given in ( 1 ) is indeed a two - parameter family of curves . if the actual distribution fails to reproduce one of these curves it is not compatible with the gravitational instability scenario with gaussian initial conditions
. that would be the case , for instance , if the observed distribution is skewed the other way around .
bernardeau , f. , 1994a , apj , 433 , 1 bernardeau , f. , 1994b , a&a , 291 , 697 bernardeau , f. , juszkiewicz , r. , dekel , a. , bouchet , f.r . , 1995 , mnras , 274 , 20 bernardeau , f. , van de weygaert , r. , 1996 , mnras in press .
bertschinger , e. , dekel , a. , 1989 , apj , 336 , l5 bertschinger , e. , dekel , a. , faber , s.m . , dressler , a. , burstein , d. , 1990 , apj , 364 , 370 couchman , h.m.p .
, 1991 , apj , 368 , l23 dekel , a. , 1994 , araa , 32 , 371 dekel , a. , rees , m. , 1994 , apj , 422 , l1 juszkiewicz , r. , weinberg , d.h .
, amsterdamski , p. , chodorowski , m. , bouchet , 1995 , apj , 442 , 39 okas , e.l . ,
juszkiewicz , r. , weinberg , d.h . ,
bouchet , f.r . , 1995 , mnras , in press nusser , a. , dekel , a. , 1993 , apj , 405 , 437 peebles , p.j.e . , 1980 , the large - scale structure of the universe , princeton univ .
press van de weygaert , r. , 1991 , ph.d .
thesis , leiden university van de weygaert , r. , bernardeau , f. , hivon , e. , bouchet , f. , 1996 , in preparation | a lot of predictions for the statistical properties of the cosmic velocity field at large - scale have been obtained recently using perturbation theory . in this contribution
i report the outcomes of a set of numerical tests that aim to check these results . using voronoi and delaunay tessellations for defining the velocity field by interpolation between the particle velocities in numerical simulations
, we have been able to get reliable estimates of the local velocity gradients .
thus , we have been able to show that the properties of the velocity divergence are in very good agreement with the analytical results . in particular
we have confirmed the @xmath0 dependence expected for the shape of its distribution function . |
in this paper we present experimental data for and a theoretical analysis of the decay rates of fluctuations in a fluid layer between two horizontal plates that are maintained at two different temperatures .
the size @xmath0 of the layer in the horizontal directions is much larger than the distance @xmath1 between the plates .
by now it has been well established that the presence of a uniform and stationary temperature gradient @xmath2 in the fluid layer induces hydrodynamic fluctuations that are long ranged in space .
an important dimensionless parameter that governs the nature of the thermal nonequilibrium fluctuations is the rayleigh number @xmath3 where @xmath4 is the gravitational acceleration and where @xmath5 is the isobaric thermal expansion coefficient , @xmath6 the thermal diffusivity , and @xmath7 the kinematic viscosity of the fluid . the rayleigh number is commonly taken to be positive when the fluid layer is heated from below .
states in which the fluid layer is heated from above then correspond to @xmath8 .
the fluid layer in the presence of a temperature gradient remains in a quiescent state for all @xmath9 less than a critical value @xmath10 for the onset of rayleigh - bnard ( rb ) convection , including all negative values of @xmath9 .
the present paper is concerned with thermal fluctuations in such a fluid layer for positive values of @xmath9 below @xmath10 . at sufficiently large wave numbers
@xmath11 the intensity of nonequilibrium fluctuations is proportional to @xmath12 both for negative and positive @xmath9 @xcite . for smaller values of @xmath11
it was shown theoretically @xcite and confirmed experimentally @xcite that the increase of the fluctuation intensity with decreasing @xmath11 is quenched in the presence of gravity , yielding a constant value in the limit of small @xmath11 .
if the presence of top and bottom boundaries is taken into account , one finds that the intensity of the fluctuations vanishes as @xmath13 @xcite at small @xmath11 .
hence , the nonequilibrium structure factor @xmath14 is predicted to exhibit a crossover from a @xmath15 dependence for larger @xmath11 to a @xmath13 dependence in the limit @xmath16 , leading to a maximum at an intermediate wave number @xmath17 which has a value near @xmath18 . as the rb instability is approached from below
, linear theory predicts that @xmath19 , as well as the total fluctuation power [ _ i.e. _ , the integral of @xmath14 ] diverges , in agreement with the asymptotic prediction obtained by zaitsev and shliomis @xcite and by swift and hohenberg @xcite .
the predicted increase as the rb instability is approached was verified quantitatively by experiments using shadowgraph measurements @xcite .
as the fluctuations become large it was predicted @xcite and confirmed experimentally @xcite that linear theory breaks down and that the fluctuation amplitudes saturate due to nonlinear interactions .
the present paper is concerned not with the intensity , but with the dynamics of the nonequilibrium fluctuations .
one experimental technique to probe the time dependence of the fluctuations is provided by dynamic rayleigh light scattering .
dynamic light scattering experiments in fluid layers with negative rayleigh numbers have shown the existence of two modes : a heat mode with a decay rate equal to @xmath21 associated with temperature fluctuations , and a viscous mode with a decay rate equal to @xmath22 associated with transverse momentum fluctuations @xcite .
thus , for large @xmath11 where gravity and boundary effects are negligible the coupling between the heat mode and the viscous mode causes an enhancement of the amplitude of nonequilibrium fluctuations , but does not affect the decay rates , in accordance with the original predictions of kirkpatrick _ et al_. @xcite
. however , for smaller @xmath11 ( corresponding to macroscopic wavelengths ) , gravity and boundary effects induce a coupling between the decay rates of the viscous and heat modes , as originally suggested by lekkerkerker and boon @xcite . for certain negative values of the rayleigh number the modes can even become propagating @xcite . on the other hand , for positive @xmath9 near the rb instability
the nonequilibrium structure factor is dominated by a very slow mode with a decay rate which vanishes as @xmath23 @xcite .
the phenomenon of critical slowing down of the fluctuations as the rb instability is approached from below has been observed experimentally , but the results obtained so far are qualitative . using forced rayleigh scattering allain _
et al_. @xcite measured the decay of imposed horizontal spatially periodic temperature profiles with various wave numbers @xmath11 .
the decay time of these imposed deviations from the steady state increased as the temperature gradient approached the value associated with @xmath10 . furthermore , the decay times became larger for periodic temperature profiles with @xmath11 closer to @xmath24 .
critical slowing down of nonequilibrium fluctuations has also been observed by sawada with an acoustic method @xcite . in this experiment ,
first a deterministic rb pattern was established and then the cell was turned over so as to change the sign of @xmath9 .
the experimental observations were interpreted with an amplitude equation that did not include any wave - number dependence .
the single decay time extracted from the data did indeed increase as the onset of convection was approached , but numerical agreement with the amplitude equation was poor . using neutron scattering , riste and coworkers
@xcite observed the critical slowing down of nonequilibrium fluctuations close to the onset of convection in a liquid crystal .
quantitative experimental studies have been performed showing the slowing down of the dynamics of deterministic patterns as the rb instability is approached from _ above _
@xcite . in the present paper
we present a quantitative theoretical and experimental study of the decay rates of fluctuations for positive @xmath9 but below the rb instability .
the results of both experiment and theory reveal the critical slowing down as the rb instability is approached from below .
theoretically , the decay rates can be determined , in principle , from a linear stability analysis of the deterministic oberbeck - boussinesq ( ob ) equations against perturbations , as has been done traditionally @xcite .
however , here we derive these decay rates by solving the linearized stochastic ob equations , obtained by supplementing the deterministic ob equations with random dissipative fluxes @xcite . because of onsager s regression hypothesis , the two procedures should yield the same result , and indeed they do .
however , it is important to note that a solution of the stochastic ob equations is required to obtain the correct amplitudes of the nonequilibrium fluctuations .
for instance , it turns out that the wave number @xmath17 corresponding to the maximum enhancement of the intensity of the fluctuations as a function of @xmath11 differs , at least for @xmath25 , from the value of @xmath11 at which the decay rate has a minimum @xcite .
we present in this paper the results of experimental shadowgraph measurements @xcite of the fluctuations in a thin horizontal layer of sulfur hexafluoride heated from below but with @xmath25 .
we show that the decay rates of the fluctuations can be obtained from an analysis of images acquired at various exposure times , and compare the experimental results with the theoretical predictions .
the data as well as the theory reveal the critical slowing down of the fluctuations as the onset of convection is approached .
to our knowledge , previous shadowgraph experiments were used primarily to obtain the fluctuation intensities @xcite .
exceptions are studies of the dynamics of deterministic patterns above the rb instability that used image sequences at constant time intervals @xcite
. measurements of the structure factor with the shadowgraph method can probe fluctuations at the small wave numbers which reveal the influence of the finite geometry .
these wave numbers generally are not accessible to light scattering
. we shall proceed as follows . in sect .
[ sec : sdynamic ] we derive the decay rates of the nonequilibrium fluctuations in a fluid layer between two horizontal rigid boundaries . to obtain analytic expressions
we determine the dynamic structure factor in a first - order galerkin approximation .
this section is a sequel to a previous publication @xcite , in which the static nonequilibrium structure factor was considered . in sect .
[ sec : shadow ] we then use the dynamic structure factor to calculate the dependence of shadowgraph signals on the exposure - time interval @xmath26 . in sect .
[ sec : exper ] we describe the experimental conditions and procedures , including the collection and analysis of the shadowgraph images .
specifically , we show how the decay rate @xmath27 of the fluctuations below the rb instability can be determined experimentally from shadowgraph images obtained with two different exposure times .
a comparison of the experimental results obtained for @xmath27 with our theoretical predictions is presented in sect .
[ sec : results ] .
finally , our conclusions are summarized in sect .
[ sec : summary ] .
as was done previously @xcite , we determine the nonequilibrium structure factor by solving the linearized stochastic ob equations for the temperature and velocity fluctuations .
it should be noted that the use of the linearized ob equations implies that the critical slowing down of the fluctuations near the rb instability is treated in a mean - field approximation in which the decay rate will vanish at @xmath28 when the wave number @xmath11 equals a critical wave number @xmath24 @xcite .
it is known from theory @xcite and experiment @xcite that nonlinear effects will cause a saturation of the fluctuating amplitudes when they become very large .
the effect of nonlinear terms on the decay rates depends on the nature of the bifurcation in the presence of fluctuations . in the rb case
the fluctuations change the bifurcation , which is supercritical in the absence of fluctuations @xcite , to subcritical @xcite . in that case
one would expect the decay rates to remain finite at the bifurcation ; however , this issue is beyond the scope of the present paper . for the case of a fluid layer between two plates with stress - free ( _ i.e_. , slip ) boundary conditions the intensity and the decay rates of the nonequilibrium fluctuations have been obtained in a previous publication @xcite .
the advantage of stress - free boundary conditions is that they permit an exact analytic solution of the problem . here ,
instead , we consider the more realistic case of a fluid layer between two rigid boundaries , corresponding to two perfectly conducting walls for the temperature and with no - slip boundary conditions for the local fluid velocity .
while this case does not permit an exact analytic solution , a good approximation can be obtained by representing the dependence of the temperature and velocity fluctuations upon the vertical @xmath29 coordinate by galerkin polynomials @xcite .
specifically , two of us have determined the static nonequilibrium structure factor in a first - order galerkin approximation @xcite .
a first - order galerkin - polynomial solution appears to provide a good approximation below the convection threshold for the total intensity of the fluctuations @xcite .
we thus extend here the first - order galerkin treatment considered in ref .
@xcite , in the expectation that it will also provide a good approximation for the decay rates and the actual amplitudes of the two coupled hydrodynamic modes .
the decay rates of the hydrodynamic modes can be readily obtained by finding the roots in @xmath30 of the determinant of the matrix @xmath31 , defined by eq .
( 15 ) in ref .
there are two coupled hydrodynamic modes whose decay rates @xmath32 are given by @xmath33 ^ 2 + \frac{27 r\sigma\tilde{q}^2}{7(\tilde{q}^2 + 12)(\tilde{q}^2 + 10)^2}}\right\ } , \nonumber\end{aligned}\ ] ] where @xmath34 is the prandtl number , and @xmath35 is the dimensionless horizontal wave number of the fluctuations .
we note that in the previous publication @xcite the symbol @xmath36 was used to emphasize that it represents the magnitude of a two - dimensional wave vector @xmath37 in the horizontal @xmath38-plane . here
we drop the subscript @xmath39 , since in the present paper the wave vectors will always be two - dimensional in the horizontal plane . in eq .
( [ unoa ] ) , @xmath40 represents the decay rate of a slower heat - like mode which approaches @xmath21 for large values of @xmath11 , while @xmath41 represents the decay rate of a faster viscous mode approaching @xmath42 for large @xmath11 .
the advantage of the galerkin approximation is that one can specify the decay rates @xmath32 explicitly as a function of @xmath43 .
we note that in the first - order galerkin approximation considered here , we find only two decay rates , instead of a series of decay rates that would be obtained if higher orders were considered in the galerkin expansion . for studying the situation below the rb instability , where fluctuations decay to zero
, consideration of the two primary decay rates will be adequate .
as discussed more in detail later , for these eq . ( [ unoa ] ) is a good approximation .
we are interested in the time correlation function for the density fluctuations at constant pressure which is directly related to the time - dependent autocorrelation function @xmath44 of the temperature fluctuations by @xcite @xmath45 here @xmath46 and @xmath47 is the average fluid density . the time correlation function @xmath48 is just the inverse frequency fourier transform of the dynamic structure factor @xmath49 defined by eq .
( 9 ) in ref .
@xcite . following the steps described in ref .
@xcite we find after some long algebraic calculations that the time correlation function @xmath48 can be expressed as the sum of two exponentials : @xmath50 } \right .
\\ & + \left .
\tilde{a}_-(\tilde{q})\exp{\left[-\gamma_-(\tilde{q})~\delta t\right ] } \right\ } \left(\frac{z}{d}-\frac{z^2}{d^2}\right)\left(\frac{z^\prime}{d}-\frac{z^{\prime 2}}{d^2}\right).\nonumber \end{aligned}\ ] ] here the coefficient @xmath51 represents the intensity of the fluctuations of the fluid in a local thermodynamic equilibrium state at the average temperature @xmath52 ( see eq .
( 2 ) in ref .
the galerkin polynomials appear in eq .
( [ dynamic ] ) because they have been used to represent the dependence of the temperature fluctuations on the vertical coordinate @xmath29 .
we introduce the dimensionless decay rates @xmath53 where @xmath54 is the vertical thermal relaxation time .
the dimensionless amplitudes @xmath55 in eq .
( [ dynamic ] ) are then given by @xmath56- \dfrac{27(\tilde{q}^4 + 24\tilde{q}^2 + 504)\tilde{q}^2}{28(\tilde{q}^2 + 12)^2}~\sigma ( \tilde{s}^0_\mathrm{ne}-\sigma r ) } { \tilde{\gamma}_\pm(\tilde{q})~ [ \tilde{\gamma}_+^2(\tilde{q } ) -\tilde{\gamma}_-^2(\tilde{q})]}.\ ] ] in eq .
( [ amplitude ] ) , @xmath57 denotes the strength of the enhancement of the static nonequilibrium structure factor of the fluid in the absence of any boundary conditions : @xmath58 here @xmath59 is the isobaric specific heat per unit mass @xcite . in eq .
( [ strength ] ) , as everywhere else in the present paper , all thermophysical properties are to be evaluated at the average temperature @xmath52 .
we note that the amplitudes @xmath55 depend on the prandtl number @xmath60 and on the rayleigh number @xmath9 not only explicitly in accordance with eq .
( [ amplitude ] ) , but also implicitly through expression ( [ unoa ] ) for the decay rates @xmath32 .
as noted in the introduction , the decay rates of the fluctuations can also be obtained from an analysis of deviations from steady state on the basis of the deterministic ob equations , _
i.e. _ , the ob equations without random noise terms .
hence , eq . ( [ unoa ] ) for @xmath32 is implicit in the standard calculation of the convection threshold within the same galerkin approximation employed here @xcite .
however , to obtain the correct amplitudes @xmath55 , it is necessary to solve the stochastic ob equations for the fluctuating fields .
expression ( [ unoa ] ) for the decay rates , as well as eq .
( [ amplitude ] ) for the amplitudes , are rather complicated .
they simplify considerably for @xmath61 . in that limit the decay rate of the slower mode @xmath62
reduces to @xmath63 while the decay rate @xmath64 of the faster mode becomes proportional to @xmath60 and is so large that the first exponential term in eq .
( [ dynamic ] ) can be neglected .
the amplitude @xmath65 of the remaining exponential term reduces , in the same @xmath61 limit , to a simpler expression to be used later , in eq .
( [ final ] ) .
it is interesting to note that the limit for large prandtl numbers is approached rather fast .
for instance , at @xmath66 , the difference between the actual @xmath40 given by eq .
( [ unoa ] ) and the value deduced from the asymptotic expression ( [ uno ] ) is always smaller than 3% for any value of @xmath43 , with the larger deviations of about 3% at @xmath43 values close to @xmath67 . from eq .
( [ uno ] ) we find that @xmath62 approaches 10 when @xmath68 , independently of the values of the rayleigh or prandtl number .
the value 10 obtained from our first - order galerkin approximation has to be compared with @xmath69 , obtained from the exact theory for the limiting value at @xmath68 of this decay rate @xcite .
the analysis in the previous paper @xcite was specifically devoted to the static structure factor , @xmath70 , which may be obtained by setting @xmath71 in eq .
( [ dynamic ] ) or , equivalently , by integrating the dynamic structure factor @xmath49 over the entire range of frequencies @xmath30 @xcite .
however , as discussed in more detail in ref .
@xcite , the structure factor @xmath14 that is actually measured in small - angle light - scattering or in zero - collecting - time shadowgraph experiments is the one obtained after integration of the full static structure factor over @xmath29 and @xmath72 @xcite , so that : @xmath73 where @xmath74 represents the normalized enhancement of the intensity of nonequilibrium fluctuations in the first - order galerkin approximation , as given by eq .
( 20 ) in ref .
we note that , in accordance with eq .
( [ static ] ) , the static structure factor @xmath14 is expressed as the sum of an equilibrium and a nonequilibrium contribution .
however , the dynamic structure factor and its equivalent , the time dependent correlation function given by eq .
( [ dynamic ] ) , can no longer be written as a sum of equilibrium and nonequilibrium contributions .
this difference between the static and the dynamic structure factor results from the coupling between hydrodynamic modes due to the presence of gravity and boundaries .
for the same reason , the nonequilibrium intensity enhancement @xmath75 no longer appears as a simple multiplicative factor in the expression for the dynamic structure factor .
the observation that the nonequilibrium dynamic structure factor is no longer the sum of an equilibrium and a nonequilibrium contribution already pertains to the dynamic structure factor of the bulk " fluid ( _ i.e_. , without considering boundary conditions ) , where mixing of the modes is still caused by gravity effects @xcite .
the shadowgraph method provides a powerful tool for visualizing flow patterns in rayleigh - bnard convection @xcite . of special interest for the present paper
is that the shadowgraph method can also be used to measure fluctuations in quiescent nonequilibrium fluids at very small wave numbers @xcite . in shadowgraph experiments
an extended uniform monochromatic light source is employed to illuminate the fluid layer .
many shadowgraph images of a plane perpendicular to the temperature gradient are obtained with a charge - coupled - device ( ccd ) detector that registers the spatial intensity distribution @xmath76 as a function of the two - dimensional position in the imaging plane @xmath77 and the nonzero exposure time @xmath26 used by the detector to average photons for a single picture .
an effective shadowgraph signal @xmath78 is then defined as : @xmath79 where @xmath80 is a blank intensity distribution in the absence of any thermally excited fluctuations . in practice , @xmath80 is calculated as an average over many original shadowgraph images , so that fluctuation effects cancel . from a series of experimental shadowgraph signals
@xmath78 , the experimental shadowgraph structure factor @xmath81 is defined as the modulus squared of the two - dimensional fourier transform of the shadowgraph signals , averaged over all the signals in the series : @xmath82 .
the physical meaning of the shadowgraph structure factor @xmath81 in the past has been based on the assumption that the shadowgraph images are taken instantaneously . in the limit
@xmath83 @xmath81 has been related to the static structure factor of the fluid @xmath84 by a relation of the form @xcite @xmath85 in eq .
( [ signal ] ) , @xmath86 is an optical transfer function that contains various properties , such as the wave number of the incident light , the temperature derivative of the refractive index of the fluid , and details of the experimental optical arrangement . for the present work a specification of @xmath86
is not needed since it will be eliminated in the treatment of the experimental data in section [ sec : exper ] . the quantity @xmath14 in eq .
( [ signal ] ) equals the static structure factor of the nonequilibrium fluid , as defined in eq .
( [ static ] ) . to determine the dependence of the experimental shadowgraph structure factor @xmath87 on the exposure time @xmath26 we need to extend the physical optics treatment performed by trainoff and cannell @xcite .
we then find that eq .
( [ signal ] ) is to be generalized to : @xmath88 where @xmath86 is the same optical transfer function as in eq .
( [ signal ] ) and where @xmath89 is a new exposure - time - dependent structure factor , which is related to the autocorrelation function of temperature fluctuations by @xmath90 ( see eq .
[ final ] below ) between the theoretical structure factors at @xmath91 ( eq .
( [ finitetime ] ) ) , and at @xmath92 ( eq . ( [ dos ] ) ) as a function of @xmath43 , for three different collecting times .
the solid curve corresponds to @xmath93 ms , the dashed curve to @xmath94 ms , and the dotted curve to @xmath95 ms .
the prandtl number is @xmath96 and @xmath97 s.,title="fig:",width=283 ] 0.1 in in principle , @xmath89 can be evaluated by substituting eq .
( [ deff ] ) with eq .
( [ dynamic ] ) for the correlation function of the density fluctuations into the right - hand side of eq .
( [ integra1 ] ) . however , for the experimental results to be presented , it is sufficient to evaluate @xmath89 for large values of the prandtl number @xmath60 , in which case the exponential contribution with decay rate @xmath98 in eq .
( [ dynamic ] ) can be neglected , as discussed before . retaining only the contribution from the slower mode with amplitude @xmath99 in eq .
( [ dynamic ] ) and performing the integrations , we deduce from eq . ( [ integra1 ] ) @xmath100 } { 18 \tilde{\tau}^2~ \tilde{\gamma}^2_-(\tilde{q})},\ ] ] where @xmath101 is a dimensionless exposure time , defined by @xmath102 with @xmath103 the vertical relaxation time introduced in eq .
( [ deftv ] ) . in the limit @xmath83 ,
( [ finitetime ] ) reduces to @xmath104 , so that it equals the static structure factor measured in small - angle light - scattering or in shadowgraph experiments in the zero collecting time approximation , as given by eq .
( [ static ] ) .
it is interesting to note that , due to the nonzero collecting time , @xmath26 , even in thermal equilibrium ( @xmath92 ) the shadowgraph measurements present some structure " . from eq .
( [ finitetime ] ) , we find that this equilibrium structure , as a function of the dimensionless collecting time , @xmath101 , is given by @xmath105}{(\tilde{q}^2 + 10)^2}.\ ] ] it can be readily checked that , in the limit @xmath83 , eq . ( [ dos ] ) reduces to the structureless constant @xmath106 , in agreement with eq .
( [ static ] ) .
this value is 17% lower than the actual value @xmath51 , due to the use of a first - order galerkin approximation @xcite . to gain insight into the effect of the exposure time on shadowgraph measurements ,
we show in fig .
[ f : uno ] the difference between the nonequilibrium structure factor as obtained from eq .
( [ finitetime ] ) and the equilibrium structure factor ( @xmath92 ) as obtained from eq .
( [ dos ] ) , as a function of @xmath43 , for three different collecting times .
corresponding to some of the experimental results to be discussed below , we used @xmath91 , a vertical relaxation time @xmath107 of 0.74 s , and a prandtl number @xmath60 equal to 34 .
we evaluated the difference for @xmath108 , @xmath109 , and @xmath95 ms .
the structure factors @xmath89 and @xmath110 have been normalized by dividing them by the product @xmath111 . from fig .
[ f : uno ] , we arrive at the following conclusions : ( i ) the main effect of a nonzero collecting time is to lower the height of the measured @xmath89 ; this is expected since fluctuations cancel out when larger exposure times are used .
( ii ) an additional effect of a nonzero collecting time , which can be observed in fig .
[ f : uno ] , is a displacement of the maximum in @xmath89 to lower @xmath43 values .
this effect is mainly due to the subtraction of the equilibrium structure given by eq .
( [ dos ] ) , which decreases with increasing @xmath43 .
( iii ) another interesting feature , we infer from fig . [
f : uno ] , is that the dependence on @xmath112 for low @xmath43 is preserved , while the dependence on @xmath113 for large @xmath43 is destroyed for the effective structure factor .
actually , at large @xmath11 , there exists a crossover from a @xmath15 dependence to a @xmath114 dependence , for nonzero values of @xmath26 , as will be shown below in eq .
( [ g32 ] ) .
a simpler formula for @xmath115 can be obtained by introducing into eq .
( [ finitetime ] ) the limiting value of @xmath99 for large prandtl numbers .
we then obtain @xmath116 } { \tau^2~\tilde{\gamma}^2_-(\tilde{q})}\\ \hfill\times \frac{27 \tilde{q}^2}{28 ( \tilde{q}^2 + 10)(\tilde{q}^4 + 24\tilde{q}^2 + 504)-27 r \tilde{q}^2},\end{gathered}\ ] ] where for @xmath62 eq .
( [ uno ] ) for large @xmath60 should be used . as discussed after eq .
( [ uno ] ) , for prandtl numbers near 15 the asymptotic limit ( [ final ] ) is already closely approached .
equation ( [ final ] ) reduces , in the limit @xmath83 , to the expressions obtained in ref .
it is interesting to note that , in the limit @xmath117 , eq . ( [ final ] ) reduces to : @xmath118 showing , as mentioned above , that a nonzero collecting time changes the asymptotic behavior from @xmath15 to @xmath114 .
note that eq .
( [ g32 ] ) is evidently not valid when @xmath93 in which case one needs the asymptotic expression given by eq .
( 30 ) in ref .
to visualize the thermal fluctuations with the shadowgraph method one needs to perform the experiments with a fluid in which the thermal noise will be large @xcite .
this goal can be accomplished by selecting a fluid in the vicinity of its critical point @xcite . the measurements reported here
were obtained for sulfur hexafluoride .
the apparatus and experimental procedures have been described in detail elsewhere @xcite . here
we describe only those aspects which are specific to the experiment with a fluid near its critical point . for the details of the cell construction we refer to fig .
9 of ref .
initially we used a diamond - machined aluminum bottom plate which could be positioned with piezo - electric elements .
the bottom - plate thermistors were embedded approximately 0.64 cm below its top surface .
even though to the naked eye this plate had a near - perfect mirror finish , the tool marks from the diamond machining imposed a preferred direction on the fluctuations below the onset of convection .
thus we placed an optically flat sapphire of thickness 0.318 cm on top of the aluminum plate .
a thin silver film was evaporated on the top surface of this sapphire to provide a mirror for the shadowgraphy .
initially we used a sapphire of thickness 0.952 cm for the cell top .
in addition a sapphire of thickness 1.90 cm was in the optical path of the shadowgraphy and provided the top window of the pressure vessel . in combination with the very small cell thickness used in these experiments , it turned out that the optical anisotropy of these randomly oriented sapphires introduced an anisotropy of the shadowgraph images which obscured the rotational symmetry of the fluctuations . in order to minimize this effect
, we replaced the pressure window by a fused - quartz window and used an optically flat sapphire of 0.318 cm thickness for the cell top . under these conditions we found that the fluctuations below the onset of convection yielded a structure factor which was nearly invariant under rotation , as can be seen from fig .
[ fig : images ] below .
the cell spacing was fixed by a porous paper sidewall with an inner ( outer ) diameter of 2.5 ( 3.5 ) cm .
since the top sapphire was supported along its perimeter which had a diameter of 10 cm ( _ i.e. _ , considerably larger than the cell wall ) , the force exerted on the cell top by the bottom plate and the wall caused a slight bowing of the initially flat top . over the entire sample diameter
this yielded a radial cell - spacing variation corresponding to about one circular fringe when illuminated with an expanded parallel he - ne laser beam .
this variation of the thickness by about 0.3 @xmath119 m assured that convection would start in the cell center , rather than being nucleated inhomogeneously near the cell wall . assuming a parabolic radial profile for the cell spacing
, we estimate that the spacing was uniform to much better than 0.1% over the [email protected] @xmath121 area near the cell center which was actually used for the shadowgraph images .
the actual sample thickness was measured interferometrically @xcite and found to be 34.3 @xmath119 m .
the thermodynamic properties of sulfur hexafluoride in the critical region can be calculated from a equation developed by wyczalkowska and sengers @xcite . for the viscosity we used a fit of data from refs .
@xcite and @xcite to a smooth function .
this approach neglects a small anomaly of the viscosity at the critical point .
we used fits of smooth functions to the conductivity data from refs .
@xcite .
the measurements were made at constant pressure @xmath123 and at constant mean sample temperature @xmath124 .
the mean temperature was adjusted so that the density @xmath125 was the critical density @xmath126 kg m@xmath127 . the imposed temperature difference @xmath128 caused a density variation of the sample along an isobar .
this is illustrated in fig .
[ fig : phase_dia ] for the conditions of the present experiments , namely for @xmath129 bar and @xmath130c . at @xmath131
, we have @xmath132 , @xmath133 @xmath121/s , and @xmath134 s. .
the dashed line indicates the coexistence curve separating liquid and vapor .
the vertical dotted line is the critical isochore .
the solid circle is the critical point with @xmath135c , @xmath136 bar , @xmath137 g/@xmath138 .
the solid line represents the isobar @xmath139 bar used in our measurements , and the heavy section of this line shows the temperature - density range on the isobar which is spanned by the sample with thickness @xmath140 m when @xmath141c.,title="fig:",width=264 ] 0.1 in along the isobar @xmath139 bars .
the heavy part of the line terminated by two open circles indicates the temperature range spanned by the sample with @xmath142 m and @xmath143c.,width=283 ] much of the theoretical work on rbc has been done in the ob approximation @xcite , which assumes that the fluid properties do not vary over the imposed temperature interval , except for the density where it provides the buoyant force @xcite .
non - ob effects have been considered by a number of investigators , and most systematically by busse @xcite . at leading order
, they break the reflection symmetry of the system about the horizontal midplane , and at the onset of convection they yield a transcritical bifurcation to a hexagonal pattern @xcite , instead of the roll pattern of pure ob convection .
when the mean temperature corresponds to the critical isochore , @xmath131 , this effect is of modest size .
although several properties contribute , we illustrate this by showing the isobaric thermal expansion coefficient in fig .
[ fig : alpha ] along the isobar of fig .
[ fig : phase_dia ] as an example .
one can approximate @xmath5 as a sum of two contributions , one of which is anti - symmetric and the other one symmetric about the mean temperature ( and thus approximately also about the horizontal midplane ) of the sample .
only the anti - symmetric part is considered in the theory @xcite .
its smallness near the onset of convection is seen from the similar values of @xmath5 at the top and bottom of the cell ( open circles in the figure ) . a quantitative calculation of the parameter @xmath144 introduced by busse @xcite ( see eqs .
( 13 ) in ref .
@xcite ) yields @xmath145 .
this value indicates that , within our experimental resolution , only rolls should be seen near onset .
this is indeed the case @xcite .
however , the variation of @xmath5 which is symmetric about the midplane is quite large .
it does not break the reflection symmetry about that plane and thus permits the existence of _ rolls _ at onset rather than requiring a hexagonal plan form .
behavior similar to that of @xmath5 is found for the specific heat @xmath59 and for the rayleigh number . at present
there is no theoretical treatment of these higher - order non - ob effects .
thus we proceed empirically by examining the variation of @xmath9 as a function of vertical position or local sample temperature .
if we neglect the temperature dependence of the thermal conductivity and assume that the local temperature in the fluid layer still varies linearly as a function of @xmath29 , we obtain the rayleigh number profiles shown in fig .
[ fig : rayleigh ] .
the curve at the top is for the experimental value @xmath146c corresponding to the onset of convection .
the other curves , from top to bottom , correspond to @xmath1470.860 , 0.699 , 0.430 , and 0.269 . for @xmath148c
one sees that the local @xmath149 far exceeds the value @xmath150 for the uniform system . on the other hand , near the top and bottom of the sample
the local @xmath9 is well below the onset of convection for the homogeneous system . the data in fig .
[ fig : rayleigh ] show that , though we find rolls above threshold , non - ob effects can not be completely neglected in our experiments .
we shall return to the influence of non - ob effects on the comparison between experiment and theory in sect .
[ sec : results ] .
( see eq .
( [ eq : r ] ) ) along the isobar @xmath139 bar for @xmath151 m . from top to bottom , the curves are for @xmath152 , 0.860 , 0.699 , 0.430 , and 0.269 with @xmath153c.,title="fig:",width=283 ] 0.1 in the bath temperature @xmath154 and the bottom - plate temperature @xmath155 were adjusted so as to hold the mean sample temperature constant .
this temperature was chosen so that the mean density corresponded to the critical density @xmath156 .
because of the small sample thickness the thermal resistance of the sample was comparable to that of the top and bottom confining plates .
this fact required a special procedure to assure that the sample was indeed at the temperature corresponding to @xmath156 . before a run at a given fixed pressure
was started , we measured the power of shadowgraph images of the fluctuations at a fixed imposed @xmath157 as a function of the mean temperature @xmath158 of the system consisting of the bottom plates , the sample , and the top plate as determined by the bath temperature and the bottom - plate thermistor .
this temperature difference is the sum of those across the bottom aluminum plate @xmath159 , across the boundary between the aluminum plate and the bottom sapphire @xmath160 , across the bottom sapphire @xmath161 , across the sample @xmath162 , across the top sapphire @xmath163 , and across a boundary layer above the top sapphire in the water bath @xmath164 . from estimates of the thermal resistances of these sections
we find @xmath165 .
thus , at the onset of convection , we measured @xmath166c and deduced @xmath153c .
( top row ) and @xmath167 k ( bottom row ) .
the exposure time was 500 ms.,title="fig:",width=283 ] 0.1 in [ fig : images ] at each @xmath128 , three shadowgraph - image sequences @xmath168 with @xmath169 , and @xmath170 were acquired . for each sequence , a different exposure time
@xmath171 was used , namely @xmath172 = 0.500 s , @xmath173 = 0.350 s , and @xmath174 = 0.200 s. the time interval @xmath175 between the images was typically one or two seconds , which was large enough for the images to be nearly uncorrelated .
the images of each sequence were averaged to provide a background image @xmath176 , as discussed in sect .
[ sec : shadow ] .
then , for each image of the sequence , a dimensionless shadowgraph signal @xmath177 was computed , in accordance with eq .
( [ shadowsignal ] ) .
the mean ( over @xmath178 ) value of a typical @xmath179 was within the range @xmath180 , indicating adequate stability of the light intensity and image - acquisition system .
next , the two - dimensional fourier transform of each shadowgraph signal in the sequence was computed , and the modulus square calculated , obtaining series of @xmath181 for further analysis .
typical shadowgraph signals and the squares of the moduli of their fourier transforms are shown in fig .
[ fig : images ] .
the nearly - uniform angular distribution of the transforms illustrates the rotational invariance of the rayleigh - bnard system . as explained in sect .
[ sec : shadow ] , due to the rotational invariance of rb convection in the horizontal plane , the modulus squared fourier transformed shadowgraph signals have rotational symmetry , and for an infinitely extended sample they would depend only on the modulus @xmath11 of the wave vector @xmath37 .
however , the finite spatial extent of the images leads to random angular fluctuations of @xmath181 . to reduce these fluctuations , we performed azimuthal averages @xmath182 over thin rings in fourier space ( the angular average is denoted by the overline and depends only on @xmath11 and @xmath171 ) .
now for each shadowgraph measurement @xmath183 the integral @xmath184 is the total power and , by parseval s theorem , has to be equal to the variance of the original @xmath177 .
we used eq .
( [ eq : p ] ) as a check of consistency for the entire procedure of taking the fourier transform , of calculating the modulus squared and the azimuthal average over thin rings , and of assigning to each ring a @xmath11 value .
finally , we averaged over the @xmath185 individual @xmath182 of each @xmath171 series , to obtain the experimental shadowgraph structure factor @xmath186 . as a function of the dimensionless wave number @xmath43 .
these results are for @xmath129 bar , @xmath187c , and @xmath188c .
the solid squares ( solid circles ) are for @xmath189 s ( @xmath190 s ) .
the open symbols are the corresponding background measurements for @xmath191.,title="fig:",width=283 ] 0.1 in for the remainder of this paper we shall use the experimentally determined cell spacing @xmath192 m to scale @xmath11 according to @xmath193 .
two examples of the product @xmath194 , both for @xmath139 bar , @xmath195c , and @xmath196c , are shown in fig .
[ fig : s_s ] .
the solid squares are for @xmath197 s , whereas the solid circles are for @xmath198 s. as discussed in the previous section , the relationship ( [ signal2 ] ) between the experimental shadowgraph structure factor @xmath199 and the corresponding fluid structure factor @xmath200 involves the optical transfer function @xmath201 .
nevertheless , we observe some qualitative agreement of the experimental results for @xmath202 in fig .
[ fig : s_s ] and the theoretical results for @xmath203 shown in fig .
[ f : uno ] .
the expected maximum near @xmath204 is present , and the decrease as @xmath43 vanishes is consistent with the predicted @xmath112 dependence .
the longer averaging of the random fluctuations diminishes @xmath205 , which was one of the features noted after fig .
[ f : uno ] for @xmath203 .
the experimental structure factor @xmath206 depends on the rayleigh number and , hence , on @xmath128 . for @xmath191 , instrumental ( mostly camera ) noise
is expected to be the dominant contribution , _
i.e_. , the contribution from the equilibrium fluctuations is negligible and the theoretical equilibrium structure , given by eq .
( [ dos ] ) , is expected to be unobservable in our experiments .
examples of the shadowgraph structure factors for @xmath191 are shown by the open symbols in fig .
[ fig : s_s ] .
one sees that @xmath207 is well represented by a straight line , corresponding to white noise . at large @xmath43 ,
@xmath208 merges smoothly into the data for the @xmath209 with the same @xmath171 , as one would expect .
( see eq .
( [ eq : ratio ] ) ) based on @xmath210 s and @xmath198 s. from top to bottom the data are for @xmath211 and @xmath212c .
the onset of convection occurred at @xmath213c.,title="fig:",width=283 ] 0.1 in a detailed comparison between the experimental and the theoretical structure factors requires knowledge of the optical transfer function , @xmath201 , which depends on the details of the experimental optical arrangement .
it involves , for instance , the size of the pinhole and the focal length of the lens used to make the parallel " beam , and the spectral width of the light source.@xcite in the present work the spatial structures to be determined ( the fluctuation wavelengths ) had length scales [ typically @xmath214 that are one or two orders of magnitude smaller than those of more conventional rb experiments .
for this reason , we found it difficult to obtain @xmath215 for our instrument with sufficient accuracy to avoid significant distortion of @xmath216 .
we circumvented the difficulty of the optical transfer function by deriving the _ dynamic _ properties of the fluctuations from ratios of @xmath199 with different values of @xmath171 . to account for the instrumental white noise , before taking such ratios , we subtracted the measured @xmath217 in the absence of a temperature gradient from @xmath218 to yield @xmath219 after this background subtraction , we formed for @xmath220 the ratio @xmath221 where @xmath172 is the longest exposure time @xmath198 s. the ratios @xmath222 obtained for @xmath210 s as a function of @xmath43 are shown in fig .
[ fig : ratio ] for four values of the temperature difference @xmath128 .
the shadowgraph transfer function @xmath215 cancels and is no longer contained in @xmath223 .
in addition , for the ratios @xmath222 , it is irrelevant whether we use the definition of the structure factor considered in sect .
[ sec : sdynamic ] , or the shadowgraph definition displayed in fig .
[ fig : s_s ] , the latter including a factor @xmath224 .
thus , we are allowed to use eq .
( [ finitetime ] ) for @xmath225 , so that @xmath226 one sees that @xmath223 depends only on @xmath227 and @xmath228 . at each @xmath43 and @xmath128 ,
only the decay rate @xmath229 is unknown and thus can be determined from the experimental value of @xmath230 . , in units of inverse seconds ,
obtained from the values of @xmath230 shown in fig .
[ fig : ratio ] ( see eq .
( [ eq : calr ] ) ) .
the symbols are the same as in fig .
[ fig : ratio ] .
the top ( dash - double - dotted ) curve is the prediction for @xmath191 .
the remaining curves are the theoretical predictions ( see eq .
( [ uno ] ) ) for the values of @xmath128 of the data ( see fig . [
fig : ratio]).,title="fig:",width=283 ] 0.1 in as a function of @xmath231 , using the values @xmath232 s and @xmath233 from the least - squares fit described in the text .
the data and symbols correspond to those in fig .
[ fig : gamma0 ] .
the curves indicate the corresponding theoretical results obtained from eq .
( [ uno ] ) by using the scale factors of the rayleigh number @xmath234 from the least - squares fit.,title="fig:",width=302 ] 0.1 in obtained from the least - squares fit described in the text .
solid symbols correspond to @xmath229 based on exposure times @xmath198 s and @xmath210 s. for the open symbols @xmath171 was 0.350 s. the solid curve is the polynomial @xmath235 which passes through the point ( 1,1 ) .
the dashed curve represents @xmath236 and was calculated from values of @xmath9 like those shown in fig .
[ fig : rayleigh].,title="fig:",width=283 ] 0.1 in
in fig . [ fig : gamma0 ] we show the decay rate @xmath229 as a function of @xmath43 .
the symbols represent the experimental values deduced from the data for @xmath222 displayed in fig .
[ fig : ratio ] by solving eq .
( [ eq : calr ] ) . from top to bottom , the data sets are for @xmath237 0.269 , 0.430 , 0.698 , and 0.861 .
the curves represent the theoretical values calculated from eq .
( [ uno ] ) with @xmath238 s ( the value derived from the fluid properties at the mean temperature and at the critical density ) .
the topmost ( dash - double - dotted ) line is for equilibrium : @xmath239 ( _ i.e. _ , @xmath240 ) .
the remaining four curves are for the values of @xmath241 of the data sets , if we adopt the boussinesq estimate @xmath242 with @xmath243 ( the value obtained from the galerkin approximation used in this paper @xcite ) for the rayleigh number in eq .
( [ unoa ] ) .
one sees that the predictions , based on the boussinesq approximation , do not agree very well with the experiment .
however , both theory and experiment reveal clearly a minimum of @xmath229 near @xmath244 which becomes more pronounced as @xmath128 approaches @xmath245 . for the larger values of @xmath246 the experimental values of @xmath229 have a maximum near @xmath247 .
we expect that this is due to nonlinear effects in the physical system which lead to second - harmonic generation .
this phenomenon is not contained in the theory .
we believe that the major disagreement between the measurements and the calculation is due to symmetric non - ob effects discussed in sect .
[ sec : nob ] .
since there is no quantitative theory , we proceeded empirically and explored the possibility that non - ob effects can be accommodated to a large extent by multiplying the experimental @xmath241 used to estimate @xmath9 by an adjustable scale factor .
we introduced an adjustable parameter @xmath234 which was allowed to be different for each @xmath248 and used @xmath249 for the rayleigh number .
in addition , we introduced a single adjustable scale factor for all data sets which adjusted the length scale of the experiment so as to yield a corrected wave number @xmath250 to be used in the fit of the theory to the data .
we expect @xmath251 to compensate for experimental errors in the cell spacing and in the spacing between the pixels of the images , to be within a few percent of unity , and to be the same for the runs at all @xmath128 .
finally , we treated the vertical relaxation time @xmath103 as an adjustable parameter .
we note that @xmath252 depends on the cell spacing , and thus any error in the length scale will lead to an error in the time scale @xmath103 .
one also might expect @xmath103 to depend on @xmath128 because quadratic non - ob effects would be larger at larger @xmath128 ; but it turned out that a single value for @xmath103 for the runs at all @xmath128 was sufficient to describe all the data .
we carried out a simultaneous least - squares fit of eq .
( [ eq : calr ] ) to a group of 11 data sets of @xmath229 , like those in fig .
[ fig : gamma0 ] , based on @xmath198 s and @xmath210 s. each data set was for a different @xmath253 , and collectively they spanned the range @xmath254 .
a separate fit was done to the second group of 9 available data sets based on @xmath172 = 0.500 s and @xmath255 s. for each group we simultaneously adjusted @xmath256 , and the eleven or nine @xmath234 .
we obtained the same result @xmath257 from both groups .
the fit also gave @xmath258 ( @xmath259 ) s for the group based on @xmath260 s. qualitatively consistent with the expected influence of the symmetric non - ob effects , the fitted value of @xmath103 is somewhat smaller than the value 0.738 s estimated for @xmath131 and the experimental value @xmath261 m .
as said above , part of this difference is attributable to the error in @xmath1 indicated by the result obtained for @xmath251 . in fig .
[ fig : gamma1 ] we show the results for the product @xmath262 for the four examples displayed in fig . [
fig : gamma0 ] as a function of @xmath263 , together with the corresponding predictions generated by using the values of @xmath234 from the fit .
except for the second - harmonic contribution at large @xmath43 and @xmath128 , the adjusted theory agrees quite well with the data .
the values obtained for @xmath234 are given in fig .
[ fig : paras ] .
the two groups ( @xmath210 s and @xmath255 s ) agree very well with each other , showing that consistent results are obtained with different exposure times .
also shown is a fit to the data sets with @xmath210 s in which the individual @xmath234 were replaced by a quadratic function which was forced to pass through @xmath264 at @xmath265 .
this fit gave @xmath266 this two - parameter representation of @xmath267 fits the data equally well .
one sees that @xmath267 is largest at the smallest @xmath128 , indicating that the estimate @xmath268 ( see eq .
( [ eq : rob ] ) ) becomes worse as @xmath128 decreases .
this is to be expected because the approximation for @xmath267 must approach unity as @xmath128 approaches @xmath269 in order for @xmath270 to vanish . as a simplest empirical attempt to include non - ob effects in the comparison between experiment and theory , we define , in analogy to eq .
( [ eq : rob ] ) , a non - boussinesq rayleigh number @xmath271 where as before @xmath243 and where the angular bracket indicates an average over the spanned temperature ( and thus density ) range along the isobar .
the averaged critical rayleigh number @xmath272 is equal to @xmath273 for @xmath274 .
it turns out that @xmath275 for our experiment .
this approximation corresponds to a re - defined @xmath276 we note that by definition @xmath277 and thus @xmath278 as it should be .
equation ( [ eq : f_r , nob ] ) is plotted in fig .
( [ fig : paras ] ) as a dashed line .
one sees that this simplest non - ob model accounts very well for the experimental data of @xmath279 .
of course , it would be very helpful to have a proper theory ( rather than an empirical model ) of this interesting effect .
in this paper we have reported on experimental and theoretical studies of the dynamics of thermal fluctuations below the onset of rayleigh - bnard convection in a thin horizontal fluid layer bounded by two rigid walls and heated from below .
starting from the fluctuating linearized boussinesq equations , we derived theoretical expressions for the dynamic structure factor and the decay rates and amplitudes of the hydrodynamic modes that characterize the dynamics of the fluctuations .
the dynamic structure factor is dominated by a slow mode with a decay rate that vanishes as the rayleigh number @xmath9 becomes equal to its critical value @xmath10 for the onset of convection .
we used the shadowgraph method to determine the ratio @xmath230 of shadowgraph structure factors obtained with different camera exposure times . from the theoretical results for the dynamic structure factor
, we derived a relationship between this ratio and the decay rates of the fluctuations . using this result and the experimental values of @xmath230
, we obtained experimental decay - rate data for a wide range of temperature gradients below the onset of rayleigh - bnard convection in sulfur hexafluoride near its critical point .
quantitative agreement between the experimental decay rates and the theoretical prediction could be obtained when allowance was made for some experimental uncertainty in the small spacing between the plates and an empirical estimate was employed for symmetric deviations from the oberbeck - boussinesq approximation which are expected in a fluid with its mean density on the critical isochore .
the research of j. oh and g. ahlers was supported by u.s . national science foundation grant dmr02 - 43336 .
g. ahlers and j. m. ortiz de zrate acknowledge support through a grant under the del amo joint program of the university of california and the universidad complutense de madrid .
the research at the university of maryland was supported by the chemical sciences , geosciences and biosciences division of the office of basic energy of the u.s .
department of energy under grant no .
de - fg-02 - 95er14509 . | we present experimental data and their theoretical interpretation for the decay rates of temperature fluctuations in a thin layer of a fluid heated from below and confined between parallel horizontal plates .
the measurements were made with the mean temperature of the layer corresponding to the critical isochore of sulfur hexafluoride above but near the critical point where fluctuations are exceptionally strong .
they cover a wide range of temperature gradients below the onset of rayleigh - bnard convection , and span wave numbers on both sides of the critical value for this onset .
the decay rates were determined from experimental shadowgraph images of the fluctuations at several camera exposure times .
we present a theoretical expression for an exposure - time - dependent structure factor which is needed for the data analysis . as the onset of convection
is approached , the data reveal the critical slowing - down associated with the bifurcation .
theoretical predictions for the decay rates as a function of the wave number and temperature gradient are presented and compared with the experimental data .
quantitative agreement is obtained if allowance is made for some uncertainty in the small spacing between the plates , and when an empirical estimate is employed for the influence of symmetric deviations from the oberbeck - boussinesq approximation which are to be expected in a fluid with its density at the mean temperature located on the critical isochore . 0.2 in |
odd - odd @xmath0 nuclei are special many - body systems which are very suitable for the test of isospin symmetry @xcite .
the reason is that they are most symmetric with respect to the proton - neutron degree of freedom and that yrast states with different isospin quantum numbers coexist at low energy @xcite .
this allows @xmath2-ray spectroscopy of isovector @xmath12 transitions and it makes odd - odd @xmath0 nuclei important for testing isospin symmetry @xcite .
furthermore , these nuclei play a decisive role in the determination of the @xmath13 part of effective interactions , _ e.g. _ , @xcite , and they are of great interest for the understanding of weak processes , enhancement mechanisms of electromagnetic transitions , as well as for problems of nuclear astrophysics @xcite .
however , up to very recent time more or less comprehensive information was available only for the odd - odd @xmath0 nuclei in the @xmath14 , @xmath15 shells and for one nucleus from the @xmath9-shell : @xmath16sc .
recent progress in both experimental and theoretical directions brought new valuable data for the heavy odd - odd @xmath0 nuclei @xmath17v @xcite , @xmath18mn @xcite , and @xmath19co @xcite in the lower part of the @xmath9-shell ( @xmath20-shell ) and even for some nuclei of the upper part of the @xmath9-shell , like @xmath21br @xcite . while some understanding of the key problems of the low - energy structure of @xmath20 nuclei seems to be obtained and regularities similar to the ones appropriate for the @xmath22-shell are revealed there are still many uncertainties for low - spin structure of these nuclei with mass numbers @xmath23 @xcite .
the first odd - odd @xmath0 nucleus of this region , which may help to draw confident conclusions on the situation in the mass region above @xmath11ni is @xmath1cu .
but the experimental data available for the low - energy level scheme of @xmath1cu @xcite is quite sparse and the theoretical full-@xmath9 shell model treatment of this nucleus is a tough computational problem .
early attempts to understand the low - spin level scheme of @xmath1cu were , therefore , limited to shell model calculations with the inert @xmath11ni core .
this approach unsatisfactorily required substantial changes in the values of single particle energies with respect to the ones for the @xmath20 nuclei and too large effective quadrupole charges @xcite .
this paper presents new experimental data for @xmath1cu which was investigated with the @xmath1ni(p , n@xmath2)@xmath1cu fusion evaporation reaction up to an excitation energy of 3.5 mev with the cologne osiris cube @xmath2-array .
we could significantly extend the hitherto known low spin level scheme of @xmath1cu @xcite , identify many new transitions , and establish their multipole character and relative intensities .
the new experimental results are accompanied by full @xmath9-shell model calculations with the new gxpf1 residual interaction @xcite universal for the whole @xmath9-shell and help to verify experimental assignments .
the data and the gxpf1 results are compared to schematic shell model calculations with the @xmath11ni core and a residual surface delta interaction ( sdi ) .
the consequences of the softness of the @xmath11ni core on the spectra are pointed out .
excited states of @xmath1cu were populated in the @xmath1ni(p , n@xmath2)@xmath1cu fusion evaporation reaction with a beam energy of 14 mev provided by the cologne fn - tandem accelerator .
the target was a 1 mg/@xmath24 thick highly enriched self - supporting @xmath1ni foil .
five compton - suppressed ge - detectors and one compton - suppressed euroball cluster detectors @xcite were used in the cologne osiris cube - spectrometer .
two of the ge - detectors were mounted in forward direction at an angle @xmath25 with respect to the beam axis .
another two were mounted in the backward direction at an angle of @xmath26 with respect to the beam axis .
the fifth ge - detector and the euroball cluster detector were placed at an angle @xmath27 below and above the beam line , respectively . about @xmath28 @xmath3-coincidence events
were recorded .
single @xmath2 spectra and @xmath3 coincidence spectra of the depopulating photon cascades in @xmath1cu were measured with high energy resolution . as an example of the data , fig .
[ fig1 ] shows the @xmath2 spectrum observed in coincidence with the decay of the @xmath29 , @xmath13 state to the @xmath30 , @xmath13 ground state of @xmath1cu .
the low - spin level scheme of @xmath1cu was constructed from the @xmath3 coincidence relations .
it is displayed in fig .
we observed 17 levels and 31 @xmath2 transitions in this nucleus .
with respect to earlier spectroscopic work @xcite , 17 @xmath2 transitions and five levels are new . in order to assign spin and parity quantum numbers we analyzed the @xmath3 angular correlation information and the signs of linear polarizations using the euroball cluster detector as a compton polarimeter .
the angular correlation pattern is determined by the spin quantum numbers of the levels involved in a cascade , by the gaussian width @xmath31 of the @xmath32-substate distribution of the initial level and by the multipole character of the corresponding @xmath2 radiation .
the gaussian width @xmath31 @xcite and multipole mixing ratio @xmath33 have been deduced from @xmath34 minimization @xcite .
the sign convention following krane , steffen and wheeler @xcite has been used for the determination of @xmath33 .
the analysis of the @xmath3 angular correlations resulted in five new unambiguous spin assignments for the levels at 444 kev ( @xmath29 ) , 1052 kev ( @xmath35 ) , 2066 kev ( @xmath36 ) , 2751 kev ( @xmath37 ) and 3423 kev ( @xmath38 ) .
the spin assignments for the levels at 2066 kev ( @xmath36 ) and 3423 kev ( @xmath38 ) are based on the spin and parity assignment @xmath39 of the level at 444 kev .
the assignment for the level at 1653 kev ( @xmath40 ) has been confirmed in the present experiment .
as an example for the assignments we show in fig .
[ fig3 ] the experimental values of the relative @xmath3 coincidence intensities of the 1103 1648 kev cascade for the angular correlation groups of our spectrometer together with the values fitted for two different spin hypotheses .
the number of different correlation groups results from the geometry of the cologne - coincidence - cube - spectrometer @xcite .
the 1103 1648 kev cascade connects the level at 1648 kev , which could be assigned @xmath29 via the angular correlation of the 1204 444 kev cascade , with the @xmath41 ground state .
it is evident from the figure , that a spin quantum number @xmath42 for the level at 2751 kev can not reproduce the data ( @xmath43 ) for any value of the possible octupole / quadrupole mixing ratio @xmath33 of the assumed @xmath44 transition .
in contrast to this the fitted values are in good accordance with the experimental ones ( @xmath45 ) for a spin quantum number @xmath37 for the level at 2751 kev . for the correlation analysis
we treat the parameter , @xmath31 , which describes the gaussian width of the @xmath32-substate distribution , as a free parameter . aside from the spin quantum numbers of the excited states , the measured @xmath3 angular correlations also give valuable information on the multipole mixing ratios of the @xmath2 transitions involved ( see table [ table1 ] ) . for seven levels with known spin values
, we could also deduce the parity .
this assignment was based on the electric or magnetic character of the depopulating @xmath2-transitions . to determine this character ,
the cluster detector was used as a compton polarimeter .
the sum of two coincident detector signals , which stem from the compton scattering of an initial @xmath2-quantum in one segment of the cluster and the subsequent absorption in another segment , carries the full energy information of the initial @xmath2-ray .
the geometry of the compton scattering process depends on the polarization of the initial @xmath2-ray with respect to the beam axis .
therefore observable asymmetries of the compton scattering process allow to measure the @xmath2-polarizations and the radiation character .
the seven large volume ge - crystals of the cluster form a non - orthogonal polarimeter . numerical simulations @xcite as well as recent experiments
@xcite have shown , that the cluster detector is an efficient compton polarimeter .
the in - set in fig .
[ fig5 ] shows the configuration of the cluster with respect to the beam axis in the present experiment .
this configuration leads to three different scattering planes for the compton scattering of @xmath2-rays between adjacent segments of the cluster . in our experiment
these scattering planes enclosed angles of @xmath46 , @xmath47 , and @xmath48 with the reaction plane , respectively .
the sum energy of the two coincident signals was sorted in two different spectra , @xmath49 and @xmath50 depending on to which scattering plane the involved pair of segments corresponds .
these spectra were used to obtain proper spectra for compton scattering intensity differences and sums , namely , @xmath51 and @xmath52 .
the experimental asymmetry is defined as @xcite @xmath53 where @xmath54 denotes the positively defined polarization sensitivity of the cluster and @xmath55 is the linear polarization of the incoming photon with respect to the given geometry . since the sign of the linear polarization , @xmath56 , determines the character of the electromagnetic radiation in case of pure multipolarity , we can conclude this character with eq .
( [ expasym ] ) from the sign of the experimental asymmetry @xmath57 .
[ fig5 ] shows the difference spectrum @xmath58 .
@xmath59 has positive values for all energies .
a summary of the energy levels with certain spin and parity values and with their depopulating @xmath2 transitions and branching ratios is given in table [ table1 ] .
the assignment of the isospin quantum number @xmath60 is done by comparing the energies of the levels to the energies of the corresponding states of the @xmath60 isobaric partner nucleus @xmath1ni . from isospin symmetry
we expect , that the excitation energies of analogue states are close in isobaric partners .
the two lowest excited states in @xmath1ni are the @xmath61 state at 1454 kev and the @xmath62 state at 2460 kev . from the excitation energy of the @xmath63 state in @xmath1ni and from the difference of excitation energies ( 1450 kev ) of the @xmath64 state ( 203 kev ) and the @xmath4 state ( 1653 kev ) in @xmath1cu
one can assign the isospin quantum number @xmath60 to the @xmath4 state in @xmath1cu .
furthermore the large @xmath65 matrix elements of the 601.4 kev and 1208.8 kev transitions to the @xmath66 and @xmath67 , @xmath13 states , respectively , and the predominantly isovector character of the @xmath65 transition operator support the @xmath60 assignment . assuming positive parity for the @xmath37 state of @xmath1cu at 2751 kev excitation energy , it can be tentatively identified as the ias of the @xmath62 , @xmath60 state of @xmath1ni . similar to the case of the @xmath4 state , this assignment is again based on the comparison of the excitation energy difference ( 2548 kev ) to the @xmath64 state at 203 kev with the excitation energy of the @xmath62 state of @xmath1ni and on the @xmath2-decay pattern .
the decays of that @xmath68 state to the @xmath69 and @xmath70 @xmath13 states is characterized by very small quadrupole / dipole mixing ratios .
this fact supports the @xmath60 assignment for the level at 2751 kev .
although small quadrupole / dipole mixing ratios were also expected for a @xmath60 state with negative parity we can discard the possibility of a @xmath60 @xmath71 state because there are no negative parity states in @xmath1ni in this energy region . in a previous study of @xmath1cu @xcite a level at 2690(20 ) kev was identified as the @xmath60 , @xmath72 state from particle spectroscopy in ( @xmath73he,@xmath74 ) charge exchange reactions .
its @xmath60 assignment was done in ref .
@xcite on the basis of the close energy match to the @xmath75 state of the isobaric partner nucleus @xmath1ni .
this state was not observed in the present experiment and its decay properties are not known . due to the comparably large uncertainty for the excitation energies deduced from that particle
spectroscopy the values of energies from the previous and the present paper for the assigned @xmath7 states at about 2.7 mev agree within three standard deviations .
therefore , one might think that the uncertainty in excitation energy of 2690(20 ) kev claimed by the authors of ref .
@xcite could have been too optimistic for that particular level and their @xmath60 , @xmath72 state would coincide with the @xmath60 , @xmath72 state at 2751 kev proposed above .
it is , however , also possible that there exists a doublet of @xmath7 states , one with isospin quantum number @xmath60 and the other with @xmath13 as it was recently observed in the neighboring odd - odd @xmath0 nucleus @xmath19co @xcite .
the latter hypothesis is supported by the shell model results discussed in the next section .
one of the first successful and very important results of the nuclear shell model was an understanding of the origin of the @xmath76 magic number .
thus , the nucleus @xmath11ni has the properties of a doubly magic inert core in a simple approach of spherical shell model .
this assumes that the low - energy structure of @xmath77 nuclei with several nucleons above @xmath11ni may be described within the shell model in small configurational spaces .
@xmath1cu is one of such nuclei and shell model calculations with a @xmath11ni core have been performed for this nucleus already in the late 1960 s @xcite .
however , it was realized that the excitations of @xmath11ni core are important for the structure of the nuclei with @xmath77 @xcite .
the recent information on @xmath11ni @xcite establishes a rather high degree of softness of @xmath11ni .
core excitations are important and can be described by modern large - valence shell calculations .
in fact , it has been found that a recent effective interaction suitable for mid-@xmath9-shell nuclei produces a significant amount of @xmath11ni core excitations in neighbor nuclei of about 30 - 40% @xcite .
one way to identify the impact of the core excitations on the structure of @xmath1cu is to compare the predictions of modern large - scale shell model calculations with an effective interaction adjusted for the full pf - shell which include the core excitations with small space shell model calculations with an inert @xmath11ni core and a schematic interaction suitable for the smaller @xmath10 space that do not contain the core excitations .
we have , therefore , performed two sets of shell model calculations for @xmath1cu .
the first one uses @xmath11ni as the core and a residual surface delta interaction ( sdi ) @xcite with a parameterization similar to the one for @xmath19co @xcite .
single particle energies were extracted further from the spectrum of @xmath78ni ( see table [ param ] ) .
the resulted excitation energies are compared to the experimental spectra in fig .
[ sdispectra ] .
we note that there is a good agreement for the two lowest states of each spin value @xmath79 , except for the @xmath67 and the @xmath80 state for which the calculated energies are 0.5 mev higher than the experimental ones .
we have calculated also @xmath81 and @xmath82 values between the low - lying states which are shown in table [ 4 ] .
the second set of calculations was performed by the tokyo group with the new effective gxpf1 interaction @xcite .
this interaction ( 195 two - body matrix elements and 4 single particle energies ) was determined partly from a fit to 699 experimental binding energies and level energies from 87 nuclei with @xmath83 and @xmath84 .
the starting point for the fitting procedure was a realistic g - matrix interaction with core - polarization corrections based on the bonn - c potential .
thus , for the first time a universal effective interaction for the whole @xmath9-shell is determined .
the calculation with the gxpf1 interaction was performed in the full @xmath9 shell with up to 6-particle excitations from the @xmath20 orbital to the @xmath85 , @xmath86 and @xmath87 orbitals .
results for the excitation energies are compared to the experimental data in fig .
[ gxpf1spectra ] .
one may note quite good reproduction of the experimental data .
in contrast to the calculations with the inert @xmath11ni core ( see fig . [ sdispectra ] ) there are also states with @xmath88 which are entirely due to the breaking of the @xmath11ni core .
they are coming mainly from a coupling of the one - neutron - one - proton states to the first excited @xmath89 state at 2.7 mev in the @xmath11ni core .
the energies of the @xmath90 and @xmath91 states are perfectly reproduced indicating that the core excitations are correctly taken into account .
furthermore there is a much better agreement for the @xmath92 and the @xmath93 states .
it has to be emphasized also a very good reproduction of the excitation energies and an ordering of the @xmath94 and the @xmath95 states which form isospin doublet .
the electromagnetic transition strengths and lifetimes calculated with gxpf1 are compared also to the available experimental data in table [ 4 ] .
it is interesting to compare the two sets of calculations .
the excitation energies of the yrast low - lying states with @xmath96 are reproduced excellently by the gxpf1 and acceptable for the sdi interaction .
the mean level deviations are 41 kev and 83 kev , respectively . furthermore the single particle energies ( s.p.e . )
used for the sdi with the @xmath11ni core and the effective s.p.e .
from gxpf1 for the @xmath11ni core are rather similar .
however , switching to electromagnetic transition strengths we find many differences ( see columns th-1 and th-2a , th-2b of table [ 4 ] ) . first , we note that to reproduce the experimental @xmath97 value in the calculations with the @xmath11ni core we have to increase the sum of the effective quadrupole charges @xmath98 by a factor of 2 as compared to the gxpf1 charges .
this causes also other @xmath99 @xmath100 transitions to become enhanced some of them even exceeding the corresponding large @xmath82 values from the gxpf1 calculations , _
e.g. _ like the @xmath100 decays of the @xmath66 state .
second , favored isovector @xmath101 @xmath65 transitions are of special interest , while isoscalar @xmath65 s are strongly suppressed and usually carry less information on the structure of the wave functions . in the simple quasideuteron picture @xcite one expects a strong enhancement of @xmath102 transitions ( up to 7.3 @xmath103 with spin - quenching of 0.7 ) for @xmath1cu because of the firm presence of the @xmath85 orbital
. indeed , in the calculations with the @xmath11ni - core , which closer matches the quasi - deuteron scheme , the summed @xmath81 strength for the lowest two @xmath104 states amounts to 5.7 @xmath103 .
the inclusion of core excitations reduces this sum to 2.5 @xmath103 .
the distribution of this @xmath65 strength among these two lowest @xmath104 states is different for the two sets of calculations , too .
the @xmath105 values are rather similar in both calculations , but the @xmath106 values differ by a factor of 5 even for the quenching of 0.7 for the sdi .
since the calculations with the ni core are in a very small configurational space one expects larger @xmath81 values as compared to the large scale shell model .
this is correct for the @xmath107 value but not for the @xmath108 value .
furthermore , the comparison of the @xmath100 strengths for the @xmath109 and @xmath110 transitions yields an apparent inversion of the @xmath111 and @xmath66 states in the sdi calculation with respect to the gxpf1 results .
the latter yields almost the same @xmath82 ratio for these two transitions as the experimental one .
the @xmath112 values allow to draw the same conclusion .
this inversion of the @xmath104 states and the considerable reduction of the @xmath65 strengths are caused by the core excitations .
most interesting are , however , the isovector @xmath65 strengths for the @xmath113 and @xmath114 transitions .
their ratio also indicates the inversion of the @xmath13 @xmath104 states : for the gxpf1 calculations the stronger transition goes to the @xmath66 state , while for the sdi interaction it is the transition to the @xmath111 state .
the latter should be almost completely forbidden according to the gxpf1 result .
the @xmath65 strengths for the isovector @xmath115 , @xmath116 , @xmath117 , and @xmath118 transitions indicate that many @xmath81 values even from the gxpf1 calculations are significantly stronger than the isovector @xmath119 or @xmath113 transitions by two to four orders of magnitude .
a suppression of an isovector @xmath65 transition by four orders of magnitude could indicate the presence of a powerful selection rule being at work .
we propose that this hindrance of the @xmath113 transition is a consequence of a @xmath120-phonon @xcite selection rule applied here to @xmath65 transitions in the shell model .
the reasoning if this interpretation is sketched in fig .
[ fig : qphon ] . in the shell model calculation with the gxpf1 interaction the @xmath60 @xmath4 state is most dominantly a complex one - quadrupole phonon excitation of the @xmath60 @xmath121 state , _
i.e. _ , to a good approximation @xmath122 where @xmath120 denotes the isoscalar quadrupole operator
. the @xmath13 @xmath111 state s wave function is instead generated to a large extent from the action of a part @xmath123 of the isovector @xmath65 transition operator on the @xmath60 @xmath121 state , @xmath124 .
consequently , the @xmath125 transition represents a two - step process .
the one - body @xmath65 transition operator can not simultaneously annihilate the @xmath126 @xmath120-phonon and cause the @xmath127 @xmath65 transition .
therefore , the @xmath113 @xmath65 transition is strongly hindered which is well confirmed by the data .
this interpretation is supported by the strong @xmath128 , @xmath129 @xmath65 transition , which is allowed in the @xmath120-phonon scheme if we consider the @xmath67 state as a @xmath120-phonon excitation of the @xmath111 state . for this @xmath128 transition
one @xmath120-phonon excitation is present in both , the initial and the final state , and acts as a spectator .
indeed , the @xmath81 values for the @xmath128 and @xmath127 transitions calculated with the gxpf1 interaction are close .
it is of interest to analyze these observations from the viewpoint of symmetries discussed in @xcite .
in summary we have investigated the low spin states of the odd - odd @xmath0 nucleus @xmath1cu with the @xmath1ni(p , n@xmath2)@xmath1cu fusion evaporation reaction . in the present experiment 17 low spin states were observed .
five of them and 17 new @xmath2-ray transitions were observed for the first time .
numerous multipole mixing ratios and branching ratios were determined and 5 new spin assignments were made .
the new data helps to understand the role of core excitations for the low - spin structure of @xmath1cu .
we have performed shell model calculations for the low - lying states of @xmath1cu with the sdi residual interaction with a @xmath11ni core and with the new gxpf1 interaction which is universal for the whole pf - shell .
comparison of the experimental excitation energies to the corresponding experimental quantities shows that both calculations yield good agreement for the yrast states with @xmath96 .
however the results of the two calculations differ considerably for electromagnetic transition strengths and the agreement with experiment is much better for the full-@xmath9 shell calculations . in particular
, we note that the @xmath82 values for isoscalar transitions are enhanced by a factor of 4 and the isovector @xmath81 values are reduced by factor 5 - 10 for the full calculation as compared to the @xmath10 space .
big changes in the electromagnetic transition strengths indicate the important role of @xmath11ni excitations for the structure of the low - spin states of @xmath1cu .
the apparent hindrance of the @xmath12 isovector @xmath125 @xmath65 transition is well reproduced by the gxpf1 interaction and can be interpreted as the manifestation of a @xmath120-phonon selection rule for @xmath65 transitions in the shell model .
another interesting result is the suggested existence of a @xmath13 and @xmath60 doublet of @xmath7 states at @xmath1302.7 mev .
the comparison of data with the calculations favor the @xmath7 state at 2.751 mev to have isospin @xmath60
. it would be interesting to find @xmath2-transitions from the nearby @xmath131 state predicted by the shell model and ambiguously suggested in @xcite at @xmath1302.69(2 ) mev as the lowest @xmath132 state of @xmath1cu .
the identification and study of this isospin doublet may offer valuable information on the isospin breaking for nuclei along the @xmath0 line above @xmath11ni .
the authors want to thank in particular a. fitzler , s. kasemann , and h. tiesler for help in data taking .
we also thank a. dewald , j. eberth , a. gelberg , j. jolie , r.v .
jolos , d. rudolph , d. weihaar , and k.o .
zell for helpful discussions .
this work was supported in part by the dfg under support no .
pi 393/1 - 2 , no .
br -799 /10 - 2 , the u.s .
national science foundation grant no .
phy-0070911 and grant - in - aid for specially promoted research ( 13002001 ) from the ministry of education , science , sport , culture and technology of japan .
we mention that this work is originated in a jsps - dfg joint project .
the large - scale numerical calculations were performed on parallel computers at the center for nuclear study ( cns ) at the university of tokyo supported by the grant above mentioned and also by the cns - riken joint project for large - scale nuclear structure calculations .
s. m. lenzi , d. r. napoli , c. a. ur , d. bazzacco , f. brandolini , j. a. cameron , e. caurier , g. de angelis , m. de poli , e. farnea , a. gadea , s. hankonen , s. lunardi , g. martinez - pinedo , zs .
podolyak , a. poves , c. rossi alvarez , j. sanchez - solano , h. somacal , phys .
c*60 * , 021303 ( 1999 ) .
f. brandolini , n. h. medina , r. v. ribas , s. m. lenzi , a. gadea , c. a. ur , d. bazzacco , r. menegazzo , p. pavan , c. rossialvarez , a. algora - pineda , g. de angelis , m. de poli , e. farnea , n. marginean , t. martinez , d. r. napoli , m. ionescu - bujor , a. iordachescu , j. a. cameron , s. kasemann , i. schneider , j. m. espino , j. sanchez - solano , phys .
c*64 * , 044307 ( 2001 ) . c. e. svensson , s. m. lenzi , d. r. napoli , a. poves , c.a .
ur , d. bazzacco , f. brandolini , j. a. cameron , g. de angelis , a. gadea , d. s. haslip , s. lunardi , e. e. maqueda , g. martnez - pinedo , m. a. nagarajan , c. rossi - alvarez , a. vitturi , and j. c. waddington , phys .
c * 58 * , r2621 ( 1998 ) . c. d. oleary , m. a. bentley , s. m. lenzi , g. martinez - pinedo , d. d. warner , a. m. bruce , j. a. cameron , m. p. carpenter , c. n. davids , p. fallon , l. frankland , w. gelletly , r. v. f.j anssens , d. t. joss , c. j. lister , p. h. regan , p. reiter , b. rubio , d. seweryniak , c. e. svensson , s. m. vincent , s. j. williams , phys.lett .
* 525b * , 49 ( 2002 ) . g. de angelis , t. martinez , a. gadea , n. marginean , e. farnea , e. maglione , s. lenzi , w. gelletly , c. a. ur , d. r. napoli , th .
kroell , s. lunardi , b. rubio , m. axiotis , d. bazzacco , a. m. bizzeti sona , p. g. bizzeti , p. bednarczyk , a. bracco , f. brandolini , f. camera , d. curien , m. de poli , o. dorvaux , j. eberth , h. grawe , r. menegazzo , g. nardelli , j. nyberg , p. pavan , b. quintana , c. rossi alvarez , p. spolaore , t. steinhardt , i. stefanescu , o. thelen , r. venturelli , eur .
j. a 12 , 51 ( 2001 ) .
d. g. jenkins , n. s. kelsall , c. j. lister , d. p. balamuth , m. p. carpenter , t. a. sienko , s. m. fischer , r. m. clark , p. fallon , a. gorgen , a. o. macchiavelli , c. e. svensson , r. wadsworth , w. reviol , d. g. sarantites , g. c. ball , j. rikovska stone , o. juillet , p. van isacker , a. v. afanasjev , s. frauendorf , phys.rev .
c*65 * , 064307 ( 2002 ) .
s. skoda , b. fiedler , f. becker , j. eberth , s. freund , t. steinhardt , o. stuch , o. thelen , h.g .
thomas , l. kubler , j. reif , h. schnare , r. schwengner , t. servene , g. winter , v. fischer , a. jungclaus , d. kast , k.p .
lieb , c. teich , c. ender , t. hrtlein , f. kock , d. schwalm , and p. baumann , phys .
c * 58 * , r5 ( 1998 ) .
d. rudolph , c.j .
gross , j.a .
sheikh , d.d .
warner , i. g. bearden , r. a. cunningham , d. foltescu , w. gelletly , f. hannachi , a. harder , t. d. johnson , a. jungclaus , m. k. kabadiyski , d. kast , k. p. lieb , h. a. roth , t. shizuma , j. simpson , . skeppstedt , b. j. varley , and m. weiszflog , phys .
lett . * 76 * , 376 ( 1996 ) .
vincent , p.h .
regan , d.d .
warner , r.a .
bark , d. blumenthal , m.p .
carpenter , c.n .
davids , w. gelletly , r.v.f .
janssens , c.d .
oleary , c.j .
lister , j. simpson , d. seweryniak , t. saitoh , j. schwartz , s. trmnen , o. juillet , f. nowacki , and p. van isacker , phys .
b * 437 * , 264 ( 1996 ) .
d. rudolph , c. baktash , m. j. brinkman , e. caurier , d. j. dean , m. devlin , j. dobaczewski , p .- h .
heenen , h .- q .
jin , d. r. lafosse , w. nazarewicz , f. nowacki , a. poves , l. l. riedinger , d. g. sarantites , w. satula , c .- h.yu , phys.rev.lett .
82 , 3763 ( 1999 ) .
a. o. macchiavelli , p. fallon , r. m. clark , m. cromaz , m. a. deleplanque , r. m. diamond , g. j. lane , i. y. lee , f. s. stephens , c. e. svensson , k. vetter , d. ward , phys .
480b , 1 ( 2000 ) .
t. mizusaki , t. otsuka , m. honma , b. a. brown , phys .
c*63 * , 044306 ( 2001 ) .
y. fujita , h. fujita , t. adachi , g. p. a. berg , e. caurier , h. fujimura , k. hara , k. hatanaka , z. janas , j. kamiya , t. kawabata , k. langanke , g. martinez - pinedo , t. noro , e. roeckl , y. shimbara , t. shinada , s. y. van der werf , m. yoshifuku , m. yosoi , r. g. t. zegers , eur .
j. a 13 , 411 ( 2002 ) .
g.kraus , p.egelhof , c.fischer , h.geissel , a.himmler , f.nickel , g.munzenberg , w.schwab , a.weiss , j.friese , a.gillitzer , h.j.korner , m.peter , w.henning , j.p.schiffer , b.a.brown , j.v.kratz , l.chulkov , m.golovkov , a.ogloblin , phys.scr .
t56 , 114 ( 1995 ) . | low - spin states in the odd - odd @xmath0 nucleus @xmath1cu were investigated with the @xmath1ni(p , n@xmath2)@xmath1cu fusion evaporation reaction at the fn - tandem accelerator in cologne .
@xmath3-coincidences , @xmath3-angular correlations , and signs of @xmath2-ray polarizations were measured .
seventeen low spin states below 3.6 mev and 17 new transitions were observed .
ten multipole mixing ratios and 17 @xmath2-branching ratios were determined for the first time .
new detailed spectroscopic information on the @xmath4 state , the isobaric analogue state ( ias ) of the @xmath5 state of @xmath1ni , makes @xmath1cu the heaviest odd - odd @xmath0 nucleus with known b(e2;@xmath6 ) value .
the @xmath7 state at 2.751 mev , observed here for the first time , is identified as the ias of the @xmath8 state in @xmath1ni .
the new data are compared to full @xmath9-shell model calculations with the novel gxpf1 residual interaction and to calculations within a @xmath10 configurational space with a residual surface delta interaction .
the role of the @xmath11ni core excitations for the low - spin structure in @xmath1cu is discussed . |
the plinth of cryptography is built upon the properties of confusion and diffusion as stated by shannon in 1949 @xcite , which can be linked to the main characteristics of chaotic systems : ergodicity and sensitivity to control parameters and initial conditions .
the connection between the basic coordinates of cryptography and chaotic systems has paved the research on chaotic cryptography @xcite .
a lot of different methods have been proposed in the field of chaos - based cryptography , but most of them show very serious security flaws ( * ? ? ?
* chapters 8 and 9 ) .
a very important family of chaotic cryptosystems is the one inheriting the characteristics of the substitution - permutation networks ( spns ) , as it is explained in @xcite .
this kind of architecture is not secure unless the avalanche criterion is satisfied @xcite . as matter of fact
, the inclusion of chaotic systems in this kind of architecture does not guarantee security and the assessment of the avalanche property should be thoroughly carried out @xcite . in @xcite
a chaotic cryptosystem is proposed to encrypt colour images through the permutation of their columns and rows , along with a substitution procedure based on the logistic map . from a general point of view , this cryptosystem can be interpreted as one round of a spn .
this kind of architecture present a very low level of confusion and security pitfalls if the substitution stage can be rewritten as a way to change the plaintext according to a keystream which is independent of the plaintext . as we will discuss along this paper
, this is the case of the cryptosystem described in @xcite .
the rest of the paper is organized as follows . in sec .
[ sec : description ] it is described the cryptosystem under examination . with the aim of underlining the shortcomings of this encryption scheme , we discuss in sec . [
sec : weaknesses ] some limitations with respect to the dynamical system bearing encryption , to the key space and , finally , in regards to the diffusion property of the cryptosystem .
the analysis is complemented by remarking the vulnerability of the cryptosystem against a chosen - plaintext attack . in this concern , we explain along sec . [
sec : probl - deriv - from ] how to elude the security laying on the encryption architecture selected in @xcite . finally , in sec .
[ sec : conclusions ] we summarize and discuss the results of the cryptanalysis .
the encryption procedure defined in @xcite is applied on colour plain - images of size @xmath0 and coded in rgb format .
the plain colour image is treated as a matrix @xmath1 of size @xmath2 , whereas the cipher - image is given by @xmath3 also of size @xmath2 . for the sake of clarity ,
we have first modified the notation used in @xcite and second divided the encryption method into four stages ( see fig .
[ fig : encryption ] ) : 1 .
_ rows permutations_. + the colour plain - image @xmath1 is transformed into a gray - scale image @xmath4 of size @xmath5 , just by incorporating the rows of the green and blue components after the rows of the red one .
let @xmath6 be a permutation matrix that transforms @xmath4 into @xmath7 by shuffling its rows in a pseudo - random way , through the iteration of the logistic map for control parameter equals to @xmath8 and initial condition given by @xmath9 .
the logistic map is defined by the iteration function @xmath10 and the orbit @xmath11 can be generated from a given initial condition @xmath12 by doing @xmath13 .
columns permutations_. + the matrix @xmath7 is converted into a matrix @xmath14 of size @xmath15 , by combining horizontally ( one after the other ) the three groups of @xmath16 rows that define @xmath7 . for each row of @xmath14
, the pixels are permuted according to the corresponding row of a permutation matrix @xmath17 .
the resulting matrix is noted as @xmath18 .
again , this permutation matrix is obtained by iterating the logistic map in this case with control parameter @xmath19 and initial condition @xmath20 .
selection of the next pixel to encrypt_. + once pixels have been shuffled , substitution is performed using a keystream and selecting the pixel to encrypt based on a pseudo - random sequence @xmath21 , with @xmath22 .
the sequence @xmath23 determines if the next pixel to encrypt proceeds from either the first n columns ( @xmath24 ) , the second group of n columns ( @xmath25 ) , or the third set of columns ( for @xmath26 ) of @xmath27 . in case
all pixels of a band have been already selected , the pixel to encrypt is chosen from the next colour band ( after the blue pixels , the next ones are the red ) . consequently , a vector @xmath28 of length @xmath29 is obtained by reading each colour component of @xmath30 from the first row and from the left to the right , according to the selection vector @xmath31 .
substitution stage_. + finally , the output of the previous step is masked using a keystream @xmath32 .
the update rule is given by @xmath33 the resulting cipher - image @xmath3 is derived from @xmath34 using @xmath31 , i.e. , by grouping the pixels of @xmath35 into colour components in the reversed order that they were grabbed from @xmath30 to build up @xmath28 . according to @xcite ,
the secret key of the cryptosystem consists of the set of values @xmath36 , which are used to compute two orbits of the logistic map ( eq . ) .
those orbits are the core of the procedures to generate the permutation matrices @xmath6 and @xmath37 , the pseudo - random sequence @xmath31 , and the keystream @xmath38 .
as we discuss below , the cryptanalysis of the cryptosystem can be carried out independently of those generation procedures . for a more detailed description of any of those procedures or other design details
, please refer to sec .
2.1 of @xcite .
as result of our previous work on the field of chaos - based cryptography @xcite , we can conclude that the most critical problems in chaotic cryptography are linked to three aspects : the selection of the chaotic system , the choice of an encryption architecture , and the implementation of the cryptosystem . in the specific scenario depicted by @xcite ,
there exist some problems that we have previously highlighted in regards to both the selection of the chaotic system and the encryption architecture @xcite .
those problems inform about a non exhaustive description of the cryptosystem , but also about security breaches .
the drawbacks of the cryptosystem definition are derived in sec .
[ sec : non - exha - defin ] by studying the key space of the cryptosystem on account of the dynamical properties of the underlying chaotic map , and in sec .
[ sec : low - sens - change ] through the discussion of the diffusion property of the encryption architecture .
the security analysis is the core of sec .
[ sec : probl - deriv - from ] .
one major concern in chaotic cryptography is on designing cryptosystems in such a way that the underlying dynamical systems evolves chaotically ( * ? ? ?
* rule 5 ) . in the case of the logistic map ( and other maps ) , this resorts to the evaluation of the lyapunov exponent in order to guarantee chaoticity ( see fig . [
fig : lyapunov ] ) . as a matter of fact , after the myrberg - feigenbaum point ( @xmath39 ) it can not be asserted that the logistic map is always chaotic due to the existence of a dense set of periodic windows ( i.e. , of values of @xmath40 implying _ regular _ and _ non stochastic _ behavior @xcite ) . .
the selection of @xmath8 and @xmath19 should be performed guaranteeing chaoticity , i.e. , positive values for the lyapunov exponent . ] additionally , in @xcite the use of the logistic map relies not only on its positive rate of local divergence , but also on its topological properties .
certainly , the permutation of columns and rows is conducted by the ordering of chaotic orbits of the logistic map of length @xmath29 and @xmath41 , respectively . in this sense
, we should assess whether the number of possible permutations on the values of those orbits is at least equal to the number of possible initial conditions .
the number of initial conditions is given by the inverse of the _ machine epsilon _
@xcite , which is @xmath42 for double precision floating - point arithmetic . on the other hand ,
the number of possible permutations on a given orbit of length @xmath43 is @xmath44 . in the case of deterministic dynamical systems
, this upper value is not reached due to the existence of a set of _ forbidden _ permutations @xcite .
if we restrict our discussion to dynamical systems with iteration function @xmath45 defined as a scalar , then the cardinality of the set of possible permutations of an orbit is upper bounded by @xmath46 @xcite , where @xmath47 is the topological entropy of the map @xmath45 @xcite . for unimodal maps
the topological entropy can be easily computed according to the theory of applied symbolic dynamics @xcite and , in some cases , it is even possible to give a closed analytical form @xcite . in fig .
[ fig : top_log ] we show the topological entropy of the logistic map with respect to the control parameter . according to the scope depicted by the permutation phases of the cryptosystem defined in @xcite , the control parameter
should be selected in such a way that @xmath48 is greater than @xmath49 .
if we consider that the smallest value for @xmath16 and @xmath50 is 128 , then the previous restriction is satisfied for @xmath40 above @xmath51 .
this fact implies a reduction of the key space as defined in @xcite and , although it is not a large shortening , it indeed informs about the needs of using not only the lyapunov exponent but also the topological entropy as core of the selection of the keys of the cryptosystem .
finally , another problem when defining the key space of the cryptosystem arises from the symmetry of the iteration function of the logistic map . as it is commented in @xcite , the fact that eq .
( [ eq : logistic ] ) satisfies @xmath52 implies that @xmath9 and @xmath53 are equivalent sub - keys for decryption .
the same applies to @xmath20 and @xmath54 . in the context of cryptography
a minor change in the input of a cryptosystem should imply a major change in the corresponding output ( * ? ? ?
* rule 9 ) . in this respect , if we take into account two images @xmath55 and @xmath56 with only one different pixel , then the associated cipher - images should be very different . to assess this property for the cryptosystem in @xcite
, we have encrypted the images in fig .
[ fig : low_sensitivity ] using as key @xmath57 , @xmath58 , @xmath59 , and @xmath60 .
the differential cipher - image is equal to zero for a meaningful set of pixels , which informs about the limitations of the diffusion property of the cryptosystem given in @xcite regarding changes in the plain - image .
according to @xcite , the security assessment of any cryptosystem must be carried out ( at least ) with respect to four basic attacks : * ciphertext - only attack : the cryptanalysis only knows the result of encryption . *
known - plaintext attack : several pairs of plaintext and ciphertext are accessible for the cryptographer .
* chosen - plaintext attack : the attacker gains access to the encryption machine and performs cryptanalysis by selecting adequate plaintexts . *
chosen - ciphertext attack : the decryption machine can be used by the cryptanalyst , which chooses ciphertexts in order to extract information about the secret parameters of the cryptosystem . in this section we show that the cryptosystem described in @xcite does not exclude the successful application of a chosen - plaintext attack . as it has been pointed out in sec .
[ sec : description ] , the encryption scheme consists of two classes of procedures : permutation and substitution of pixels .
the main weakness of the proposal is a consequence of the independence between the shuffling stages and the last stage , i.e. , the one concerning the substitution of pixels .
this fact can be exploited by means of the following _ divide - and - conquer attack _ , using as bottom - line chosen plain - images which are neutral elements with respect to row / column permutations @xcite . in this sense , if one encrypts a plain - image with all pixels equal to the same value , then the output of the shuffling procedures is the same plain - image .
moreover , if the plain - image is selected forcing all rows / columns being equal , then encryption only shuffles columns / rows . in correspondence to the previous comments , we can mount an attack based on a chosen plain - image with all pixels equals to zero .
let @xmath1 be a colour image with all pixels equal to zero , which implies that @xmath61 for @xmath62 .
taking into account eq .
( [ eq:1 ] ) , we have @xmath63 for @xmath64 . from the previous equation we can find the value of @xmath65 just by subtracting @xmath66 from @xmath67 .
if we want to apply the recovered @xmath38 to get any @xmath28 from the corresponding @xmath35 , then @xmath31 must be obtained .
this commitment can be accomplished using a second _ constant value plain - image_. for instance , we can use a chosen plain - image with all pixels equal to one .
this being the case , we have @xmath68 for @xmath62 and @xmath69 for @xmath70 .
let us focus on the image given as the difference between the cipher - images obtained from @xmath71 and @xmath72 respectively .
since the difference between eq .
( [ eq:3 ] ) and eq .
( [ eq:4 ] ) is equal to @xmath73 , the components of @xmath31 are determined by looking for the pixel with value @xmath73 in each colour band of that difference image .
if that pixel belongs to the red component , then @xmath74 ; if it is one of the green pixels , then @xmath75 ; finally , @xmath76 leads to a pixel in the blue band .
once the substitution keystream @xmath38 and the selection vector @xmath31 have been obtained , it is possible to reconstruct the input of the shuffling procedures according to @xcite .
this new goal is going to be achieved by using @xmath77 chosen plain - images . in this paper
we restrict our analysis to images of the same size as those used in @xcite , i.e. , images of size @xmath78 and , consequently , four chosen - plain images are required to elude the permutation - only phase . in order to validate our cryptanalysis , we have configured an encryption machine by selecting the key defined by the set @xmath57 , @xmath58 , @xmath59 , and @xmath60 . upon the assumption of having access to the encryption machine , we encrypt an image equal to zero and an image with all pixels equal to 1 .
the cryptanalysis described in sec .
[ sec : break - conf - stage ] is applied , and thus the keystream @xmath38 and the pseudo - random sequence @xmath31 are recovered . as it is commented in similar
cryptanalysis works @xcite , the recovering of those sequences is equivalent to getting the secret key .
nevertheless , the complete cryptanalysis of the cryptosystem in @xcite demands to infer a permutation matrix representing the composition of the permutation procedures lead by @xmath6 and @xmath37 .
this goal can be achieved by using plain - images with all rows / columns equals . to illustrate the cryptanalysis we are going to extract the original positions of the pixels of the first row of @xmath79 .
first , we encrypt a plain - image with each colour component determined by @xmath80 if we consider the vector @xmath81 of length 768 given by the concatenation of the first row of red , green , and blue component of the cipher - image , it is easy to verify that it contains only three values .
the values corresponding to the selected secret key are 93 , 203 , and 223 , which indicates that the first row of the cipher image comes from either the row 93 , 203 , 223 of either of the colour components of the plain image . in order to establish the colour band of each of the three candidates for row permutation
, we encrypt a plain - image with red component with all pixels equal to zero , green band being 1 , and blue component being 2 .
then , we look for the occurrences of 0 , 1 , and 2 in the first row of each colour component of the cipher - image . the intersection of this new vector of indexes of occurrence with the previous one enables to conclude that @xmath81 contains the row 93 of the blue band of the plain - image , the row 203 of the red component of the plain - image , and the row 223 of the red component of the plain - image . after identifying the source of the first row of @xmath79
, we need to label each pixel of the rows identified as sources of that row .
this aim is fulfilled if we encrypt a colour image with its three colour components equal to @xmath82 afterwards , we look for the occurrences of @xmath83 through the vector @xmath81 .
the indexes of occurrence are given by the set @xmath84 .
let us begin with @xmath85 , which is @xmath86 for the selected key .
the set @xmath85 implies that either of the referred pixels comes from the first pixel of either the row 93 of the blue component , the red row number 203 , or the row 223 of the red band of the plain - image . to select the proper value among the three candidates for the three identified pixels , we encrypt an plain - image such that the row 93 of the blue component is @xmath87 the red row number 203 @xmath88 and the row 223 of the red band
is defined as @xmath89 again , we look for @xmath90 through @xmath81 and we obtain the indexes of occurrence 235 , 356 , and 556 . only 356 is included in the previous set @xmath85 , and as a result we have that the first pixel of the row 93 of the blue component of @xmath1 goes to the pixel 100 ( @xmath91 ) of the first row of the green component of @xmath79 . if we proceed in the same fashion with @xmath84 for @xmath92 , then we obtain the permutations for all the pixels of the row 93 of the blue band of the plain - image .
the same applies to the row 203 ( 223 ) of the red band , but taking into account that the first pixel of the row is now labeled by 255 ( 254 ) .
if one applies the previous methodology for all the rows of the cipher - image , then the permutation matrix can be inferred . in this sense
, we have applied the cryptanalysis based on the six chosen plain - images to an encryption machine with secret key @xmath57 , @xmath58 , @xmath59 , and @xmath60 .
the cryptanalysis allows to get @xmath31 , @xmath38 , and the permutation matrix , which is equivalent to obtain the secret key . to verify this assertion we have encrypted an image ( the result is in fig .
[ fig : chosen](a ) ) , applied the cryptanalysis , and decrypted the cipher - image using the outputs of the cryptanalysis .
the decrypted image is the one in fig .
[ fig : chosen](b ) , which coincides with the original plain - image .
in this paper we have studied in detail a recent proposal in the area of chaos - based cryptography .
we have underlined some problems related to the dynamical properties of the system sustaining encryption , and we have also pinpointed some flaws related to the encryption architecture .
the goal of our work was not simply to show the problems of a given chaotic cryptosystem , but to highlight the possibility of creating secure proposals to encrypt information using chaos . in this flavour ,
our recommendation is on the line of the set of rules given in @xcite .
this work was supported by the uam projects of teaching innovation and the spanish government projects tin2010 - 19607 and bfu2009 - 08473 .
the work of david arroyo was supported by a juan de la cierva fellowship from the ministerio de ciencia e innovacin of spain .
d. arroyo , framework for the analysis and design of encryption strategies based on discrete - time chaotic dynamical systems , ph.d .
thesis , etsia of the polytechnic university of madrid , madrid , spain , avalaible online at http://digital.csic.es/handle/10261/15668 ( july 2009 ) .
d. arroyo , r. rhouma , g. alvarez , s. li , v. fernandez , on the security of a new image encryption scheme based on chaotic map lattices , chaos : an interdisciplinary journal of nonlinear science 18 ( 2008 ) 033112 , 7 pages .
d. arroyo , g. alvarez , v. fernandez , on the inadequacy of the logistic map for cryptographic applications , in : l. hernandez , a. martin ( eds . ) , x reunin espaola sobre criptologa y seguridad de la informacin ( x recsi ) , universidad de salamanca , salamanca , spain , 2008 , pp . 7782 , ( isbn 9788469151587 ) .
d. arroyo , j. m. amig , s. li , g. alvarez , on the inadequacy of unimodal maps for cryptographic applications , in : j. d. ferrer , a. m. ballest , j. c. roca , a. s. gmez ( eds . ) , xi reunin espaola sobre criptologa y seguridad de la informacin ( xi recsi ) , universitat rovira i virgili , tarragona , spain , 2010 , pp . 3742 , isbn 9788469333044 . s. li , c. li , g. chen , n. g. bourbakis , k .-
lo , a general quantitative cryptanalysis of permutation - only multimedia ciphers against plaintext attacks , signal processing : image communication 23 ( 3 ) ( 2008 ) 212 223 . | the interleaving of chaos and cryptography has been the aim of a large set of works since the beginning of the nineties .
many encryption proposals have been introduced to improve conventional cryptography .
however , many of those proposals possess serious problems according to the basic requirements for the secure exchange of information . in this paper
we highlight some of the main problems of chaotic cryptography by means of the analysis of a very recent chaotic cryptosystem based on a one round substitution permutation network .
more specifically , we show that it is not possible to avoid the security problems of that encryption architecture just by including a chaotic system as core of the derived encryption system .
image encryption , substitution permutation networks , permutation - only ciphers , unimodal maps , chosen - plaintext attack . |
the so called proton spin crisis is still calling for a theoretical explanation in the context of qcd .
the problem may be stated as follows .
let @xmath0 be the singlet axial current @xmath1 its on shell nucleon matrix element may be written as @xmath2 u(\mbox{\boldmath{$p$ } } ' , s ' ) , \label{form1}\end{aligned}\ ] ] where @xmath3 . in the naive parton model @xmath4 can be interpreted as the fraction of the nucleon spin carried by the quarks , and so is expected to be close to 1 . on the contrary , experimental determinations @xcite lead to a value @xmath5 , which is also in contrast with the ozi expectation @xmath6 @xcite .
it is the small experimental value of @xmath7 which is referred to as the proton spin crisis .
a numerical simulation on the lattice is the ideal nonperturbative tool to obtain an estimate of @xmath7 from first principles .
a direct computation of the matrix element of eq . ( [ form1 ] ) involves the evaluation of a disconnected diagram , characterized by a very noisy signal @xcite .
we shall adopt a different method @xcite , which makes use of the axial anomaly @xmath8 @xmath9 is the number of flavours and @xmath10 is the topological charge density .
the on shell nucleon matrix element of @xmath11 can be written , using eq .
( [ anom ] ) , as @xmath12 where @xmath13 is the nucleon mass and @xmath14 , whence @xmath15 .
the last relation does not hold in the quenched approximation , since @xmath16 develops a pole at @xmath17 in that case @xcite .
therefore , @xmath7 can be calculated on the lattice by evaluating the nucleon matrix element of the topological charge density @xmath11 in full qcd . however , a necessary condition to do this is to have the correct sampling of topological modes . here
we will give some preliminary results obtained using the standard hmc algorithm with four flavours of dynamical staggered flavours and will show that the feasibility of the determination is affected by the bad sampling of topological modes obtained with the hmc algorithm @xcite .
we have simulated four degenerate flavours of staggered fermions at @xmath18 on a @xmath19 lattice , using the standard hmc algorithm described in ref . @xcite and the wilson action for the pure gauge sector .
a trajectory length @xmath20 was used , with time step @xmath21 .
the value @xmath22 was chosen for the fermion mass . in order to evaluate the matrix element of eq .
( [ form2 ] ) we need a lattice regularization @xmath23 of the topological charge density .
the renormalization properties of @xmath23 are nontrivial @xcite : it mixes with the continuum @xmath11 , @xmath24 as well as with @xmath25 .
however , it can be shown that the last two operators can be safely neglected in a first approximation @xcite , so that the following relation holds @xmath26 we use the improved smeared operators @xmath23 proposed in ref .
@xcite , and in particular the operator defined at the second smearing level .
the value of @xmath27 can be obtained non perturbatively using the so called heating method @xcite : the short range fluctuations responsible for the renormalization are thermalized by applying a few updating steps on a discretized classical configuration of charge 1 .
the thermalization shows up as a plateau in @xmath28 plotted against the number of heating steps performed , as is shown in fig . 1 .
@xmath27 is then measured on this plateau .
we have obtained @xmath29 with the values of the parameters used in our simulation .
0.3 cm the next step is to compute the correlation function @xmath30 where @xmath31 and @xmath32 is the spin projection operator .
for @xmath33 we have taken the wall source , after performing a coulomb gauge fixing on the @xmath34 time slice by means of a steepest descent algorithm .
for @xmath35 we have taken the usual local baryon operator @xcite @xmath36 the following behaviour is expected for @xmath37 @xcite @xmath38 where @xmath39 and @xmath40 can be determined by a fit to the full baryon propagator @xmath41 @xcite . @xmath42 and @xmath43 are the axial baryonic charge and mass of the @xmath44(1405 ) particle respectively @xcite .
we have collected a sample of 230 configurations , each separated by 15 hmc trajectories , using 700 hours of a 25 gflops ape / quadrics machine . in fig .
2 we show the topological charge distribution in our sample .
it is clear that the correct thermalization for the topological charge has not yet been achieved : @xmath45 and the distribution is not symmetric under @xmath46 . the time history of the topological charge is shown in fig .
3 , where the inefficiency of the hmc algorithm to change the topological sector is clearly visible . despite this deficiency in our ensemble of configurations ,
we have attempted an estimate of @xmath7 and @xmath47 . from a fit to the full baryon propagator we have obtained @xmath48 we have thereafter measured the correlation function of eq .
( [ correl ] ) .
the zero momentum transfer limit has been approximated by using @xmath49 ( @xmath50 mev on our lattice ) .
an average has been performed over the three possible spatial directions for @xmath51 and over the two possible polarizations .
the error analysis has been done by a jackknife method .
however , the error quoted in eq .
( [ results ] ) does not take account of the systematic error coming from the bad sampling of topological modes . by looking at the effective number of topological sectors explored during the simulation ( fig .
3 ) , we roughly estimate the effective error to be up to 45 times the one quoted in eq .
( [ results ] ) .
therefore our data can only be regarded as preliminary and no conclusive results can be produced unless a correct sampling of topological modes is obtained .
this goal seems to be hardly feasible for the standard hmc algorithm : the possibility of achieving it by using different algorithms is currently under investigation .
9 emc , j. ashman et al .
b 328 ( 1989 ) 1 .
smc , d. adams et al .
b 329 ( 1994 ) 399 .
e143 , k. abe et al . ,
phys , rev .
74 ( 1995 ) 346 .
shore , g. veneziano , nucl .
b 381 ( 1992 ) 23 . j.e .
mandula , m.c .
ogilvie , phys , lett .
b 312 ( 1993 ) 327 .
m. fukugita , y. kuramashi , m. okawa , a. ukawa , phys . rev
. lett . 75 ( 1995 ) 2092 .
s.j . dong , j.f .
laga , k.f .
liu , phys .
lett . 75 ( 1995 ) 2096 .
b. alls et al .
b 336 ( 1994 ) 248 .
r. gupta , j.e .
mandula , phys .
d 50 ( 1994 ) 6931 .
r. altmeyer et al . , phys . rev .
d 49 ( 1994 ) r3087 b. alls , g. boyd , m. delia , a. di giacomo , e. vicari , phys .
b 389 ( 1996 ) 107 . s. gottlieb et al . , phys .
d 35 ( 1987 ) 2531 .
b. alls et al . , phys , lett .
b 350 ( 1995 ) 70 . c. christou , a. di giacomo , h. panagopoulos , e. vicari , phys . rev .
d53 ( 1996 ) 2619 .
a. di giacomo , e. vicari , phys .
b 275 ( 1992 ) 429 .
b. alls , m. delia , a. di giacomo , nucl .
b494 ( 1997 ) 281 .
r. altmeyer et al . , nucl .
b 389 ( 1993 ) 445 . | we present preliminary results on the proton spin structure function in full qcd .
the measurement has been done using 4 flavours of staggered fermions and an improved definition of the lattice topological charge density . |
in the microcirculation , it is often observed that the red blood cells ( rbcs ) flow in single or multiple files forming small trains of cells , called clusters@xcite .
the arrangement and organization of the rbcs depend on the diameter of the vessel and their concentration ( hematocrit ) .
each rbc interacts hydrodynamically with the other cells@xcite .
rbcs can also interact via another mechanism , namely an interaction mediated by plasma proteins .
the latter interaction is materialized either by bridging between rbcs or by a depletion force . in the bridging mechanisms proteins
make a real bridge between two neighboring rbcs , while in the depletion mechanism osmosis is responsible for the cluster formation .
we have recently discussed the implication of plasma proteins in the formation of rbc clusters in microcirculation @xcite .
the main objective is to gain further insight into the role of each mechanism .
therefore this paper will be directed towards numerical simulation in the presence of pure hydrodynamic interaction .
several studies have been devoted to understanding the hydrodynamical interaction between suspended particles in the stokes regime .
analytical models @xcite considered the motion of a linear array of rigid spheres at low reynolds number in a cylindrical tube under a pressure - driven flow ( i.e. , an imposed poiseuille flow ) .
wang and skalak@xcite estimated the hydrodynamic interaction s range between the spherical particles to be of the order of the tube diameter .
_ et al.__@xcite showed that the interparticle interactions were relatively small at low confinements , reached a maximum at intermediate confinement , and were quickly damped out at high confinement .
more recently , colloidal particles confined between two parallel plates in a quasi 2d geometry have been studied experimentally and theoretically @xcite .
the complexity in these systems arises from the difficulty to decouple the effect of brownian diffusion from hydrodynamic interactions .
an antidrag between the moving particles attributed to a negative hydrodynamic coupling has been reported . a change of sign of the hydrodynamic coupling ( from attraction to repulsion ) in a cylindrical channel
was also observed @xcite .
the effect of boundaries on the hydrodynamic interactions has been studied in the case of water - in - oil droplets in quasi-1d microfluidic devices having a square section of the order of the size of the droplet , so that the droplets are constrained to move along the channel axis @xcite .
this study reported on a non - monotonous behavior of the hydrodynamic interaction resulting from an interplay between the plug flow and the screening of the long - range hydrodynamic interaction induced by the confinement .
+ janssen _ _
et al.__@xcite have studied numerically pairs of rigid spheres and deformable drops driven by a poiseuille flow through a three - dimensional rectangular channel in the stokes regime . due to the reversibility of stokes equations ,
the interdistance between a pair of rigid spheres does not evolve in time .
however , for a pair of deformable drops , due to the up - stream / down - stream shape asymmetry , hydrodynamics interaction leads to an attraction at large separation distances and repulsion at small interdistances .
the long - range attraction was attributed to the source - quadrupole flows induced by droplets in the hele - shaw geometry .
the pair of drops tends to the same stationary separation distance independently of the capillary number a measure of the flow strength which only affects the time needed to reach the stationary state .
vesicles , a closed bilayer membrane of phospholipid molecules , have been used as a simplified model for rbcs to simulate blood flow in microcirculation @xcite with a focus on the role of the imposed flow velocity and the hematocrit ( concentration of cells in the tube ) in the organization of cells in a capillary .
the precise role of the interplay between confinement and the cell deformability is still not fully elucidated .
it has been shown in a recent study that stable clusters of vesicles can form in the absence of bounding walls under an imposed parabolic flow profile @xcite .
our objective in this paper is to take this unbounded case as a reference and see how confinement affects the hydrodynamic interactions .
we generically find repulsion at short distance , and attraction at long distance .
we also find that at a large enough vesicle interdistance the interaction can change sign , become repulsive , and then become attractive at higher interdistance .
this points to the existence of an interaction which changes sign with distance , pointing to the nontrivial effect of hydrodynamic interactions .
we further show that several branches of solutions can coexist , highlighting a complex imbricated interaction structure .
a systematic 2d numerical study ( based on a boundary integral formulation ) is undertaken here in order to analyze the evolution of a pair of vesicles , by exploring several parameters , such as channel width , initial interparticle spacing , and the flow strength .
we then consider the case of a larger clusters of vesicles and show that the size of a stable cluster can be drastically modified by the presence of walls .
implication of the results to real situations is discussed .
this paper is organized as follows . in section [ model ]
we introduce the model in details .
section [ results ] presents the main results , for both weak and strong confinement .
section [ conclusion ] is devoted to a discussion of the results and the implication for real situations .
vesicles , drops , capsules ( drops coated with polymers ) are endowed with different mechanical properties leading to different responses to external stresses . both vesicles@xcite and capsules@xcite are widely used to mimic rbcs under flow .
rbc s complex dynamics , such as tank - treading , tumbling and vacillating breathing ( aka swinging , trembling ) can be reproduced by both vesicles and capsules .
other shapes exhibited by rbcs under a poiseuille flow , such as parachute and slipper shapes , are also captured by the vesicle and capsule model ( see review @xcite ) .
vesicles resist bending , and possess an inextensible membrane ( constant area in 3d and perimeter in 2d ) but do not present a resistance to shearing , in contrast to rbcs and capsules , which are endowed with surface shear elasticity . despite the absence of a cytoskeleton in the vesicle model several features exhibited by rbcs
are also captured by vesicles , such as the above - mentioned dynamics and shapes under flow .
the vesicle bending energy@xcite is expressed in its two - dimensional form as @xmath0 where @xmath1 is the membrane bending modulus , @xmath2 is the curvature , @xmath3 is the arclength element , and @xmath4 is a lagrange multiplier which enforces membrane inextensibility.the two - dimensional membrane force is obtained by calculating the functional derivative of the total energy @xmath5\mathbf{n } - c \zeta \mathbf{n } + \frac{\partial{\zeta}}{\partial{s}}\mathbf{t } , \label{eq::membrane_force}\ ] ] where @xmath6 and @xmath7 are the outward normal and the tangent unit vectors , respectively .
the details of the derivation can be found in ref .
the vesicle is characterized by its reduced area @xmath8
^ 2},\ ] ] a dimensionless parameter that expresses the ratio between the actual fluid area enclosed by the vesicle contour @xmath9 and the area of a disk having the same perimeter as the vesicle . for a rbc
, the reduced volume ( 3d equivalent of the reduced area ) lies in the range of @xmath10 .
this will serve as a reference for our choice of parameter @xmath11 .
the vesicle s equilibrium shape obtained by minimizing ( [ eq::helfrich ] ) for a reduced volume @xmath12 is a biconcave shape similar to that exhibited by a rbc at rest @xcite .
it is convinient to decompose the total velocity @xmath13 into an undisturbed velocity in abscence of vesicles @xmath14 , and an induced velocity owing to the presence of vesicles @xmath15 .
this will help us to understand the interactions between vesicles ( see section [ results ] ) .
here we consider a pressure - driven flow in a confined geometry between two laterally infinite plates ( see fig.[fig::fig0 ] ) .
the corresponding undisturbed velocity @xmath16 at a position @xmath17 is given by @xmath18 \\
u_2^\infty ( \boldsymbol { x } ) & = & 0 , \\
\end{array } \right.\]]where @xmath19 is the midplane velocity and @xmath20 is the channel width .
the vesicles encapsulate an inner fluid with a viscosity @xmath21 and are suspended in an outer fluid with a viscosity @xmath22 . for simplicity
, we shall take @xmath23 .
the reynolds number associated with a vesicle ( with typical size @xmath24 , water viscosity and speed in the range of mm / s , as found in microcirculation ) is of order @xmath25 .
thus fluid s motion can safely be described by the stokes equations @xmath26 where @xmath27 is stress tensor associated to the total velocity and @xmath28 is the pressure .
the jump in the interfacial traction is balanced by the vesicle s membrane force @xmath29 \mathbf{n}$ ] . besides the latter force balance condition , the solution of equations ( [ eq::motion_stokes ] ) and ( [ eq::continuity_stokes ] ) must respect the following boundary conditions ( b.c . )
@xmath30 = \mathbf{0 } \\ \label{eq::bc3 } & \mathbf{u}_{out}(\mathbf{x})=\mathbf{u}_{in}(\mathbf{x } ) , \quad \quad \mathbf{x } \in \partial \omega_i \quad ( i=1, ... ,n_{v})\end{aligned}\ ] ] the vesicles can be simply regarded as a collection of lagrangian nodes advected by the flow .
the motion of the vesicles is obtained by solving the advection equation for each material node @xmath31 lying on the membrane @xmath32 the velocity of each membrane s discretization point can be deduced by solving an integral equation ( discussed in what follows ) , and the position of each node is updated in time using an explicit euler scheme .
@xmath33 in what follows we use dimensionless quantities defined as @xmath34where the relaxation time ( @xmath35 ) and radius @xmath36 of the vesicle are used as characteristic time and length scales .
the dimensionless integral equation of the velocity along the membrane is given by @xmath37 where @xmath17 and @xmath38 are two vector positions on the membrane .
the integral in the rhs expresses the induced velocity owing to the vesicles .
@xmath39 is the green s function associated with the stokes equations and vanishes at the two bounding walls .
the way this function is computed and its detailed expression are given in refs .
@xcite .
the dimensionless expression of the membrane s force reads @xmath40\mathbf{n } - c^ * \zeta^ * \mathbf{n } + \frac{\partial{\zeta}^*}{\partial{s^*}}\mathbf{t}\ } \label{eq::membrane_force_dimensionless}\end{aligned}\ ] ] in addition to the vesicle reduced area defined above ( @xmath41 ) , and the confinement given by @xmath42 , we have a new dimensionless number associated to the flow in equation ( [ eq::interfacialflowfinal ] ) , which is given by @xmath43 we will refer to this number as the capillary number , where @xmath44 is the shear rate defined at @xmath45 .
we recall that the shear rate in a parabolic flow is not a constant unlike in a shear flow .
thus , the definition of the capillary number for a parabolic flow is not unique in the literature .
the capillary number is the ratio between the flow stress and bending force density .
it may be viewed also as the ratio between the characteristic shape relaxation time @xmath46 and the time scale of the flow @xmath47 . in order to have an idea on the conversion of dimensionless units into physical ones ,
the following dimensional numbers can be used : @xmath48 , @xmath49 pa@xmath50s and @xmath51
. this leads to a characteristic time of relaxation of the cell of about @xmath52 s. this is quite consistent with measured values for rbcs@xcite .
in this section , simulations of a pair of vesicles in channels of different widths are performed .
the vesicles are initialized as ellipses elongated in the direction of the flow , if not stated otherwise .
the pair interdistance , a quantity that will be used to quantify the interaction between vesicles , is measured between the mass centers of adjacent vesicles .
we will also investigate the role of the capillary number , and the impact of initial conditions ( shapes and interdistances ) on the final state .
we shall see that there are several coexisting solutions .
a previous study has been devoted to the cluster formation in the absence of walls @xcite .
we first study the evolution of a pair of vesicles in a weak confinement in order to check that we can capture almost the same result .
we have analyzed the evolution of a pair of vesicles in a channel having a width @xmath53 , which corresponds to a weak enough confinement .
we find that the final equilibrium distance is equal to about @xmath54 for @xmath53 .
this result compares well with that obtained in unbounded flow @xcite where the final distance is of about @xmath55 .
we have analyzed systematically the behavior of a pair of vesicles for different confinements ( but still weak enough confinement ) and different initial conditions .
figure [ 13p1 ] shows a typical evolution of the pair interdistance as a function of time for different initial conditions , denoted as @xmath56 in that figure .
we see there that different initial conditions lead to the same final state .
figure [ floww20 ] shows the final configuration of the pair of vesicles as well as the induced flow field , that is the total flow field from which we subtract the imposed poiseuille flow ( as defined in section [ sec : model_subsublevel2 ] ) .
this figure also shows the total flow field .
it can be seen that the induced flow ahead of the leading vesicle ( the one on the right ) is converging towards the vesicle .
this means that if a tracer is placed ahead of the leading vesicle it will be attracted by that vesicle .
this qualitatively hints to the fact that the vesicle would attract small objects placed ahead of it .
the qualitative understanding of this flow field is quite simple , as shown schematically in figure [ flow_schematic ] .
we show there the imposed flow ( black arrows ) , the total flow ( blue ) and the induced flow ( red ) .
consider the flow field ahead of the vesicle at the center line . far away from the vesicle
the total flow field practically coincides with the imposed one ( the perturbation due to the vesicle decays away from the vesicle ) .
consider now a point close enough ahead of the vesicle ( point @xmath57 in the figure ) .
the vesicle moves with a velocity which is smaller than the maximum imposed velocity , so that close enough to the vesicle ( due to velocity continuity ) , the full velocity of the fluid ahead of the vesicle is smaller than the maximum imposed one ( the blue arrow is shorter than the black one ) .
if we consider the induced flow only ( which is what matters for interaction ) , we have to subtract the black arrows ( imposed flow ) from the blue ones ( the full field ) , so that at the center line ahead of the vesicle the induced flow ( red arrows ) points towards the vesicle .
the same reasoning applies to the rear part of the vesicle ( point @xmath58 ) , meaning that the induced flow is directed away from the vesicle .
consider now a point on the side part of the vesicle ( point @xmath59 in the figure ) .
since the vesicle is at the center its velocity is higher than that of the adjacent fluid near on top of point @xmath59 . at point @xmath59 , because of velocity continuity , the vesicle and the fluid move at the same velocity , which is higher than the imposed flow at that point . subtracting the imposed flow from the total flow field around point
@xmath59 shows that the flow field there point to the right .
joining now point @xmath57 , @xmath59 and @xmath58 ( see figure [ flow_schematic ] ) , we obtain the qualitative pattern for the induced flow field where the flow goes from the rear to the front of the vesicle . in order to analyze further this question we have systematically analyzed in details the behavior of the relative velocity of the two centers of mass of the pair of vesicles as a function of their actual interdistance by considering different initial interdistances .
let us denote the leading vesicle by @xmath60 and the following vesicle by @xmath61 , and let @xmath62 and @xmath63 denote their instantaneous positions , and by @xmath64 and @xmath65 the corresponding velocities ( velocity of center of mass has only a component along the wall direction ) .
their relative velocity is given by @xmath66 .
a positive value of @xmath67 means that the two vesicles repel each other , while a negative value means they attract each other .
an equilibrium position corresponds to @xmath68 .
in what follows we will refer to the dimensionless velocity , defined as @xmath69 .
figure [ vander1 ] shows @xmath70 as a function of the dimensionless interdistance between centers of mass , @xmath71 .
we observe a repulsion at short distance and an attraction at long distance .
there is a final equilibrium position characterized by @xmath72 ( denoted as @xmath60 within a dark circle in figure [ vander1 ] ) .
this equilibrium position is unique , independent from the choice of the initial shape ( figure [ vander12 ] ) .
this defines in a non ambiguous way the equilibrium position . clearly , that equilibrium point is stable , since for higher interdistance the relative velocity is negative , so that the vesicles will attract each other , while the opposite happens for smaller interdistance . generally ,
if the relative velocity crosses zero ( when @xmath73 increases ) from positive to negative values the equilibrium position is stable , and is unstable otherwise .
we have performed this study for several confinements , but still weak enough , and have determined the corresponding equilibrium position .
figure [ diag1 ] shows the branch of equilibrium interdistance @xmath74 as a function of @xmath75 .
the equilibrium interdistance weakly depends on confinement , and remains close to about @xmath76 .
let us now examine the generic behavior for a strong confinement .
consider the case @xmath77 . by studying the evolution of the interdistance as a function of time
, we find that for several initial conditions the pair of vesicles finally settles into a bound and stable steady state .
figure [ w31 ] shows the evolution of the interdistance with time for different initial conditions .
the first noticeable fact is that the equilibrium distance for @xmath77 is significantly larger than that obtained at weak confinement .
for example for @xmath77 the equilibrium interdistance is about @xmath78 , which is twice as large as that obtained in the weak confinement case ( about @xmath79 ) .
a first interpretation would be that the presence of the walls screens the hydrodynamics interaction .
this reasoning is , _ a priori _ , not justified , since the screening concerns both the attraction and repulsion and it is not clear how would the confinement shift the attraction and repulsion zones .
let us first analyze the velocity field in a strong enough confinement regime .
figure [ vw63 ] shows the induced flow field for two confinements @xmath80 and @xmath77 .
we also show there the flow field in the co - moving frame ( the frame moving with the pair ) .
the main difference with the weak confinement is the fact that the flow lines , which extend far away in the weak confinement ( see figure [ floww20 ] ) , are cut - off by confinement in the strong confinement case .
this partially hints to a weaker interaction in magnitude , as we shall see below .
as done above for the weak confinement case , we have analyzed the relative velocity as a function of the relative distance .
a typical result is shown in figure [ vander2 ] .
we obtain a stable equilibrium position , with short - range repulsion and long - range attraction .
we find that the relative velocity amplitude is significantly smaller than that obtained for a weak confinement ( compare with figure [ vander1 ] ) .
this is to be traced back to the screening effect due to the walls that weakens the interaction .
we have performed this study for several confinements , and have determined the equilibrium position .
figure [ diag2 ] shows the branch of equilibrium position @xmath81 as a function of @xmath75 .
in contrast to the weakly confined case ( also shown in figure [ diag2 ] ) the equilibrium distance strongly depends on confinement , and varies from about @xmath82 up to about @xmath83 .
some remarks are in order .
we have a new branch of solution for small values of @xmath20 ( the left branch in figure [ diag2 ] ) , different from that discussed above for large @xmath20 ( the right branch in figure [ diag2 ] ) .
the second observation is that we have two distinct branches , one presents an equilibrium distance which increases with @xmath20 , while the other has an equilibrium position which decreases with @xmath20 .. we have made further studies in order to clarify the overall structure of the topology of the phase diagram .
first , we have tried to analyze the bassin of attraction of the branches by exploring a wider range of initial conditions .
we begin our discussion with the weak confinement case ( large @xmath20 ) .
we have already seen in the previous section that different initial conditions lead to the same final solution . by exploring wider and wider regions of initial conditions ,
we have found a different scenario . by keeping the same confinement @xmath84 ,
as done above , we found that beyond a certain initial pair interdistance , the solution does not evolve anymore to the previous solution , but rather to another solution .
figure [ 13p2](left ) shows the temporal evolution where after @xmath85 the pair interdistance is still evolving , but converges ultimately to a distance of about @xmath86 . for different initial conditions ( smaller initial interdistance ) we found before a final distance of about @xmath87 ( figure [ 13p2](right ) ) .
this clearly shows the coexistence of two different stable solutions .
in general , these two stable solutions should be separated by an unstable solution .
therefore , we need to determine the location of the unstable branch .
we can have access to this by analyzing the relative velocity as a function of interdistance ( as performed before ) .
figure [ vandercom ] shows the result , where the points denoted as 1 , 2 and 3 in the figure , refer to the location where the pair relative velocity vanishes .
since the relative velocity around point 3 is very small , we have made a zoom in order to reveal the structure of the evolution of the relative velocity .
points 1 and 3 are stable , whereas point 2 is unstable . we have made a systematic analysis of this kind by varying the confinement ( i.e. @xmath20 ) in order to build the full diagram representing the equilibrium solutions ( like point 1 , 2 and 3 ) as a function of @xmath20 .
the results are summarized in figure [ diag_full ] .
we see that the branch for high @xmath20 undergoes a fold singularity in the form of a saddle - node bifurcation , where a stable solution ( represented by solid line ) merges with an unstable solution ( dashed line ) .
the branch arising at low @xmath20 continues to exist ( as a stable solution ) beyond the saddle - node point , and does not show , for the values of @xmath20 explored so far , any tendency of disappearance .
theoretically , this branch either continues to exist for any @xmath20 , or may undergo a fold singularity , which is one of the scenario expected from catastrophe theory .
since for large @xmath20 the relative velocity is too small , the question of the behavior of the branch at large enough @xmath20 has only an academic interest .
we shall thus not dwell further on this issue here .
finally , let us say few words on the strength of interaction as a function of confinement .
as can be expected , the confinement reduces the strength of interaction , due to screening . to understand the difference of the pairing mechanism at large and small channel widths we first consider a pair of vesicles in two different channels of width @xmath88 and @xmath89 .
the time evolution of the interdistances in each channel is shown in figure [ fig::fig6 ] .
the time needed to reach the steady state for different channel widths has been investigated .
we found that their interdistance reduces by approximatively @xmath90 the pair of vesicles needs around @xmath91 in a channel of width @xmath80 whereas only @xmath92 are needed to reduce their interdistance by @xmath93 in a channel of width @xmath94 .
since a typical time @xmath95 for rbc is about @xmath96 s , this gives about 10 s in the first case .
note also that in reality there are always fluctuations and imperfection so that the cells are not all at the same lateral position . for example
, if the leading cell is slightly off - centered its velocity may be sensibly different from the following cell , so that the time needed to reach the bound state may occur at a significantly shorter time scale than @xmath97 s. in this section we describe the effect of the capillary number on the main results .
we consider a pair of cells with four different capillary numbers @xmath98 and flowing in channels of width ranging from @xmath99 to @xmath100 ( @xmath101 @xmath102 to @xmath103 @xmath104 for rbcs ) .
figure [ fig::fig16_ca_w_di ] depicts the stationary interparticle spacing as a function of the channel width for different capillary numbers ( we do not show the full branch as before due to computational cost , and especially because we do not see any significant changes ) .
the stationary shape of the vesicles at high confinement ( @xmath105 ) is asymmetric for @xmath106 .
the shape asymmetry decreases with the increase of the capillary number and a transition from asymmetric to symmetric shapes is observed at @xmath107 ( figure [ fig::fig16_ca_w_di ] ( b ) ) .
the shapes become fully symmetric for a channel width @xmath108 and the final pair interdistance tend to the same value , independent from the capillary number . for
@xmath109 the choice of the initial shapes does not affect the stationary solution whereas for @xmath110 it leads to two different solutions in the region of transition resulting in a decrease of the stationary interdistance if starting with parachutes and the opposite if starting with ellipses .
( figure [ fig::fig16_ca_w_di ] ( c ) and ( d ) ) . the time needed to reach the stationary shape is proportional to the capillary number ( higher @xmath111 is shorter the time needed to transform an ellipse to a parachute ) which explains the dependency on the initial shapes at low @xmath111 .
globally , the capillary number seems to not significantly affect the behavior of the pair
. for vesicles , higher values of @xmath112 can be reasonably reached , such as @xmath113 . for a vesicle of radius @xmath114 ,
a channel radius of @xmath115 , and a velocity of @xmath116 , one finds approximately @xmath117 . by assuming that the relative velocity scale with @xmath112 ( as shown in figure [ vrelat_ca100 ] )
, one finds that the relative velocity is of about @xmath118 , which is no devoid of experimental testability .
the above study has focused on the case of a pair of vesicles .
the question thus arises of whether this would allow us to draw some general conclusion about the size of clusters when more cells are present .
the situation is quite complex , and here we only quote some basic examples leaving a systematic study to the future . in a recent work @xcite it has been reported that , in an unconfined poiseuille flow , the cluster size depends on the flow strength : a higher flow strength leads to a larger cluster size . while this tendency is confirmed here , we find that even for a large enough @xmath20 ( about @xmath100 ) where the confinement might be considered as weak , we obtain some important differences with the unconfined case @xcite .
we have considered the largest @xmath112 explored in @xcite ( @xmath119 ) , and for which it has been reported that the largest stable cluster contains 10 cells ( for a reduced area @xmath120 ) .
we have investigated here the largest stable cluster in the presence of a weak confinement .
figure [ cluster ] shows different initial configurations of parachute - like cells initialized as clusters of different sizes and their corresponding final state .
it is seen that for @xmath53 , the cluster is stable as long as the number of cells is smaller than , or equal to , 4 . beyond @xmath121
the leading cells detach progressively one after the other from the initial cluster with a variable time gap until decreasing the number of cells within the cluster down to 4 ( see the movies in the supplementary material ) .
the maximum size of a stable cluster obtained in unconfined flow was @xmath122 , whereas here we obtain @xmath123 in a weakly confined flow corresponding to @xmath53 ( @xmath124 ) .
this clearly means that the confinement , despite being weak , alters significantly the results . in a future study
we plan to analyze in detail the dependence of the largest stable cluster size on the confinement .
this study has reported on a complex phase diagram regarding hydrodynamic interaction between two vesicles in a confined pressure - driven flow .
it is found that several branches of stationary solutions coexist .
it would be interesting in the future to extend this study to the three - dimensional case both for vesicles and rbcs .
another important ingredient is to analyze the interplay between hydrodynamic interaction and that due to plasma proteins and study how the structure of the phase diagram reported here evolves in this case .
another interesting task for future investigations is to study collective behaviors when many cells are present and determine the link between the spatio - temporal organization and the rheology .
we acknowledge financial support from the german science foundation research initiative sfb1027 , the german french university ( dfh / ufa ) , cnes ( centre detudes spatiales ) and esa ( european space agency ) .
32ifxundefined [ 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.164.3880.717 [ * * , ( ) ] link:\doibase 10.1007/bf00396444 [ * * , ( ) ] @noop * * , ( ) @noop * * ( ) link:\doibase 10.1017/s002211206900005x [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.89.188302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.258301 [ * * , ( ) ] http://stacks.iop.org/0953-8984/17/i=31/a=003 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.4757394 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.028104 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.90.033011 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.crhy.2013.05.001 [ * * , ( ) ] , @noop * * , ( ) link:\doibase 10.1103/physreve.77.021903 [ * * , ( ) ] @noop * * , ( ) http://books.google.de/books?id=qxt5boqdegqc[__ ] , cambridge texts in applied mathematics ( , ) link:\doibase 10.1103/physreve.88.062707 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) | the hydrodynamics - induced pairing mechanism of deformable closed bilayer membranes ( vesicles , a model for red blood cells ) is studied numerically with a special attention paid to the role of the confinement ( the vesicles are within two rigid walls ) .
this study unveils the complexity of the pairing mechanism due to hydrodynamic interactions . at weak confinement
we find that two vesicles attract each other and form a stable pair if their initial distance is below a certain value .
if the initial distance is beyond that distance , the vesicles repel each other and adopt a larger stable interdistance .
this means that for the same confinement we have ( at least ) two stable branches and whether one branch prevails over the other depends only on initial conditions .
an unstable branch is found between these two stable branches . at a critical confinement the stable branch corresponding to
the smallest interdistance merges together with the unstable branch in the form of a saddle - node bifurcation . at this critical confinement
we have a finite jump from a solution corresponding to an unbounded case to a solution which is induced by the presence of walls .
the results are summarized in a complex phase diagram . |
a packing can be described as an ensemble of particles with given position and orientation , having interparticle contacts such that ( _ i _ ) there is no overlap between particles , ( _ ii _ ) force and torque balances are satisfied on every particle , and ( _ iii _ ) the packing is rigid , _ ie _ there is no displacement of particles , individual or collective , that does not create a violation of conditions ( _ i _ ) or ( _ ii _ ) ( except for global translation or rotation of the packing ) . in this work ,
we consider the geometric configuration as given ( _ ie _ condition ( _ i _ ) is fulfilled and taken for granted ) , hence we do not discuss about the spatial degrees of freedom but rather focus on the force degrees of freedom . now , let us argue that , in the case of random packings of spheres under finite pressure , fulfilling condition ( _ ii _ ) ( force and torque balances ) implies that condition ( _ iii _ ) ( stability ) is also fulfilled . force and torque balances under finite pressure ensure ( by definition )
that the system is stable against isotropic compression .
the average force in the packing is then proportional to the pressure .
but , as such , nothing is guaranteed concerning the stability of the packing against more general perturbations , as defined by the condition ( _ iii _ ) . in the case of sphere packings
, however , it is very likely that packings verifying force and torque balances while still being mechanically unstable have some sort of order , at least locally ( one can think of the case of a row of frictionless spheres perfectly aligned for example , which can maintain force balance , but is unstable ) .
this is what is observed at the jamming transition , when packings prepared under isotropic pressure are naturally stable @xcite .
hence , for random packings of spheres , the requirement of force and torque balances is argued to be enough for having stability too , _
ie _ conditions ( _ ii _ ) and ( _ iii _ ) are simultaneously fulfilled .
similarly to the global rigidity condition , force and torque balances are entirely constrained by the contact network of the packing . if we define @xmath32 as the vector joining the center of particle @xmath33 and the contact @xmath34 on it , the contact network is uniquely defined by the complete set @xmath35 .
calling @xmath36 the force acting on particle @xmath33 from contact @xmath34 , force and torque balances read : @xmath37{$\qquad \forall \ , a$}\\ \displaystyle \sum_{i \in \partial a } \vec{r}^{\,a}_i \times \vec{f}^{\,a}_i = 0 , \qquad & f_{i}^{\,t } \leq \mu f_{i}^{\,n } , & \end{array}\ ] ] where the notation @xmath38 denotes the set of contacts of particle @xmath33 .
we explicitly take into account friction by decomposing the force into normal and tangential parts @xmath39 , where @xmath40 and @xmath41 are normal and tangential unit vectors to the contact , respectively .
the inequality @xmath42 ensures the repulsive nature of the normal force .
the last inequality ensures coulomb s law with friction coefficient @xmath5 . ) .
( b ) part of factor graph used to compute @xmath43 and @xmath44 by using eq .
( [ eq : p2pto ] ) and eq .
( [ eq : cavity_eq ] ) . on the edges around particle @xmath33 ,
the arrows indicate that we consider the force probabilities @xmath45 as ` messages ' going through contacts in the contact network . in our algorithm
, @xmath44 is iteratively updated by the uncorrelated force probability @xmath46 on neighboring edges with the force / torque balance constraint @xmath47 on that particle @xmath33 , verifying the cavity equation eq .
( [ eq : cavity_eq ] ) .
the marginal probability @xmath48 on site @xmath34 is calculated as a product of @xmath44 and @xmath49 as in eq .
( [ eq : p2pto ] ) .
, scaledwidth=61.8% ] the constraints induced by force and torque balances on the forces @xmath50 are not always satisfiable . in the case of frictionless spheres
, we can recover the known @xmath51 directly from force balance alone .
the naive maxwell counting argument @xcite applied to frictionless spheres reduces eq .
( [ eq : force_balance ] ) to a set of linear equations by taking into account only force balance and neglecting the repulsive nature of the forces .
maxwell argument considers the minimal number of forces needed to satisfy eq .
( [ eq : force_balance ] ) which gives , per sphere , @xmath7 equations and @xmath52 variables ( forces ) , implying @xmath51 to have a solution . below this threshold , there is generically no solution to eq .
( [ eq : force_balance ] ) . to accurately extend this counting argument to more general conditions ( frictional , repulsive , and/or non - spherical particles )
, one must take into account all the constraints in eq .
( [ eq : force_balance ] ) , including the repulsive nature of the forces and the coulomb condition for frictional packings .
indeed , the naive maxwell argument , neglecting those constraints , concludes @xmath53 for any frictional packing of spheres , ignoring the dependance on the friction coefficient @xcite . on the other hand , below
we show that including the above mentioned constraints we obtain an accurate lower bound @xmath4 , explicitely depending on @xmath5 . we tackle the problem of satisfiability of force and torque balances eq .
( [ eq : force_balance ] ) by looking at the contact network in an amorphous packing as an instance of _ random graph_. as depicted in fig .
[ fig : from_network_to_graph]a , starting from a packing of @xmath54 particles , we explicitly construct a so - called factor graph @xcite , considering the @xmath55 contacts as ` sites ' , and the @xmath54 particles as ` interaction nodes ' .
each site @xmath34 bears two vectors @xmath56 and @xmath57 and two opposite forces @xmath58 ( one per particle involved in the contact ) .
note that @xmath59 are uniquely determined by the contact network @xmath60 and represents the ` quenched ' disorder in the system , whereas @xmath61 are free to rotate in the plane tangent to contact . on each interaction node ( particle )
@xmath33 , we enforce force balance , torque balance , repulsive interactions and coulomb friction conditions on its @xmath62 neighboring sites by an interaction function , @xmath63 we define the partition function @xmath64 and entropy @xmath65 for the problem of satisfiability of force and torque balances for a fixed realization of the quenched disorder @xmath59 ( as shown in appendix ) : @xmath66 without loss of generality , we work in the constant pressure @xmath67 ensemble : if we find a solution to the force and torque balances problem having a pressure @xmath68 , we can always find one solution with pressure @xmath67 by multiplying all forces by @xmath69 .
note that the only effect of the applied pressure in a hard sphere packing at zero temperature is to set the average normal force .
the contact network is unmodified by a change in pressure as the particles are hard .
if the entropy is finite , there exists a solution to force and torque balances .
the satisfiability / unsatisfiability threshold of force and torque balances is the coordination number @xmath4 that separates the region of finite @xmath65 from the region for which @xmath70 , corresponding to an underdetermined / overdetermined set of eq .
( [ eq : force_balance ] ) , respectively . within this framework , the satisfiability problem eq .
( [ eq : force_balance ] ) is one of the well studied class of constraint satisfiability problems ( csp ) defined on random networks @xcite .
these problems are ubiquitous in statistical physics and computer science and have attracted a lot of attention in recent years .
thus powerful methods from statistical physics have been developed to study them @xcite .
we will work here with random graphs , retaining from actual packings the distribution of coordination number @xmath71 . in this work ,
we use as @xmath71 a truncated gaussian distribution , more precisely , @xmath72 in 2-d and @xmath73 in 3-d , both providing very good fit to experimental and numerical data @xcite .
( note that due to the truncation @xmath74 , we choose @xmath75 to achieve a desired @xmath76 . )
we clarify that although this choice of @xmath71 relies on quantities extracted from previous numerical studies , most of the results we obtain on the force distribution are insensitive to the exact form of @xmath71 .
the main change observed for frictionless sphere packings is the behavior near the peak of the force distribution , as is discussed in sec .
[ sec : poff ] . for the distribution of contacts around one particle @xmath77
, we use a flat measure over all contacts positions that do not create overlaps between the @xmath62 corresponding particles .
for spherical particles this distribution is @xmath78 } \theta(1/2 - \vec{r}^{\,a}_i.\vec{r}^{\,a}_j)$ ] with @xmath79 the heaviside function . a real contact network of a two or three - dimensional packing shows some finite dimensional structure , of course , and treating it as random graphs can only be an approximation .
this amounts to a mean - field approximation , neglecting correlations between the different contacting forces acting on one particle .
this approximation is routinely used in the context of spin - glasses for example @xcite . from this point of view
, we stand on the same ground as the q - model approach @xcite , and the edwards approach of bruji _
et al_. @xcite .
even though we keep a part of the finite dimensional geometric constraints through the distributions @xmath80 and @xmath71 , the mean - field approach neglecting correlations between neighboring contact forces should only be exact in the limit of high dimension , where we expect those correlations to vanish ( some recent progress however suggest that correlations are not completely trivialized by the high dimensionality in packings , due to a fullrsb phase transition close to jamming @xcite ) . in finite dimension , it is clear that short range correlations matter , and having a mean - field approximation can only be a first step .
one finite dimensional feature that we miss is the detailed link between local structure and force distribution .
a stability analysis set constraints relating the detailed features of the near contact structure and the force distribution at low forces @xcite .
the mean - field computation we are providing here does not include that constraint , and there may be some finite dimensional effects at low forces that modify the mean - field picture we are giving here , as shown in the result section [ sec : poff ] .
as the simplest case , we restrict the description of this section to packings of spheres , with obvious generalization to non - spherical objects . for a given disordered packing ,
each particle @xmath33 has unique surroundings , different from its neighbors or other particles in the packing .
these surroundings are defined by the contact number @xmath62 and contact vectors @xmath81 .
if the system is underdetermined , several sets of forces in the system satisfy force and torque balances , and each contact force has a certain probability distribution @xmath48 .
the local disorder makes each contact unique , and the probability distributions of forces @xmath48 are different from contact to contact .
we define the overall force distribution in a packing @xmath82 as an average over the probability distributions of forces over the contacts : @xmath83 next , we show that on a random graph , we can access the distributions @xmath48 with a self - consistent set of local equations using the cavity method @xcite . in this description ,
we work at fixed pressure @xmath67 , _ ie _ we consider any two solutions differing only by an overall rescaling of the pressure to be only one genuine solution . each contact is linked to two particles , @xmath33 and @xmath84 .
we denote @xmath85 the probability distribution of the force @xmath86 of a site ( contact ) @xmath34 , if @xmath34 is connected only to the interaction node ( particle ) @xmath33 , that is , if we remove particle @xmath84 ( dig a cavity ) from the packing .
the main assumption of the cavity method is to consider that @xmath85 and @xmath87 are uncorrelated .
therefore , we can write the probability of forces at contact @xmath34 as : @xmath88 with @xmath89 the normalization . under the mean - field assumption
a set of local equations ( called cavity equations ) relates the @xmath45 s , as depicted in fig .
[ fig : from_network_to_graph]b : @xmath90 where the notation @xmath91 stands for the set of neighbors of @xmath92 on the graph except @xmath93 , and @xmath94 is the normalization .
crucially , we do not average over the contact directions @xmath95 at this stage ( whereas the q - model @xcite and edwards model @xcite do ) .
this implies that every link @xmath96 has a different distribution , due to the local ` quenched ' disorder provided by the contact network @xmath95 and contact number @xmath62 .
hence , finding a set of @xmath45 that are solutions of eq .
( [ eq : cavity_eq ] ) allows to get the distribution of forces on every contact _ individually _ , through the use of eq .
( [ eq : p2pto ] ) . looking for a solution of eq .
( [ eq : cavity_eq ] ) for a given instance of the contact directions ( meaning for one given packing ) is possible .
these equations are commonly encountered as ` cavity equations ' in the context of spin glasses or optimization problems defined on random graphs @xcite , and they can be solved by message passing algorithms like belief propagation . here
we follow another route , since we are interested in @xmath82 for not only one packing but over the ensemble of all random packings .
thus , we study the solutions of the cavity equations in the thermodynamic limit to provide typical solutions for large packings . as in statistical mechanics ,
the partition function will be dominated by the relevant typical configurations which we expect will be realized in experiments .
the set of cavity equations eq .
( [ eq : cavity_eq ] ) might _ a priori _ admit several solutions .
however , at the satisfiability / unsatisfiability threshold , there is only one solution to force balance eq .
( [ eq : force_balance ] ) , which means only one solution to eq .
( [ eq : cavity_eq ] ) ( a @xmath97-function on every site , centered on the solution of force and torque balances ) . to get this threshold
, we may thus take for granted that there is a unique solution to the cavity equations , a case known as replica symmetric ( rs ) in the spin glass terminology .
this assumption will be fully justified _ a posteriori _ by the fact that we recover the correct threshold in the known case of frictionless spheres . in the thermodynamic limit , the set of @xmath45 s that are solutions of eq .
( [ eq : cavity_eq ] ) are distributed according to the probability @xmath98 : @xmath99\ ] ] in this case , we can replace the sum over @xmath100 by a continuum description of the @xmath45 s based on their distribution @xmath98 .
the probability that a given @xmath45 is set by a cavity equation eq .
( [ eq : cavity_eq ] ) involving @xmath101 contact is proportional to @xmath102 .
thus , averaging over the ensemble of random graphs , eq .
( [ eq : def_qofq ] ) becomes a self - consistent equation @xcite : @xmath103 .
\end{array}\ ] ] where @xmath104 is the normalization .
note that the value of the integral does not depend on the choice of @xmath105 .
once a solution to eq .
( [ eq : cavity_averaged ] ) is known , we deduce the force distribution @xmath82 in the overall packing as the average of all these probability distributions and contacts : @xmath106 ^ 2 \end{array } \label{fdis}\ ] ] where @xmath107 is the normalization to ensure @xmath108 .
equation ( [ eq : cavity_averaged ] ) stands out as the main and crucial difference with previous approaches , in particular the q - model @xcite and edwards description @xcite .
although these approaches also neglect correlations , our work does not reduce to those models due to a fundamentally different way of treating the disorder in the packing . here , we consider a site - dependent @xmath48 , where the edwards model and q - model create all sites equal .
thus in our method the average over the packing configurations is not done at the same level as the average over forces .
that is , we perform a quenched average over the disorder of the graph . as random packings in two or three dimensions
have a rather small connectivity , the fluctuations in the environment of one particle are large : no particle stands in a ` typical ' surrounding .
hence , the average over the ` quenched ' disorder ( the packing configurations ) must be done with care .
averaging directly eq .
( [ eq : cavity_eq ] ) ( a so - called ` annealed ' average in spin - glass terminology ) , as the previously cited approaches do @xcite , amounts to neglect the site to site fluctuations . performing a ` quenched ' average as in eq .
( [ eq : cavity_averaged ] ) , however , allows to take into account these fluctuations correctly @xcite , and leads to a force distribution which is the average force distribution over the ensemble of possible packings , as opposed to the force distribution of an averaged packing .
this issue becomes also crucial to the study of the satisfiability transition of force and torque balances .
for example , the q - model describes a force distribution in a packing of frictionless spheres ( _ ie _ it finds solution to force balance ) , even in cases where we know there is no solution , such as when @xmath109 in 3-d frictionless systems .
this can be understood by looking at the entropy defined in eq .
( [ eq : partition_function ] ) . the annealed average over disorder done in the q - model amounts to compute the averaged partition function @xmath110 , and get the entropy through @xmath111 , with @xmath64 defined in eq .
( [ eq : partition_function ] ) .
but one can show that @xmath110 is always finite for @xmath112 .
indeed , for @xmath113 , an infinite row of perfectly aligned spheres will satisfy force and torque balances and will contribute to the partition function .
a straightforward generalization of this example shows that for @xmath112 , the annealed entropy is finite . on the contrary ,
the quenched average amounts to compute the averaged entropy @xmath114 .
now , for our frictionless sphere example , typical configurations with @xmath109 can not satisfy force balance , and their diverging negative entropy will dominate the average @xmath115 .
hence , the quenched average correctly captures the satisfiability / unsatisfiability transition at @xmath116 while the annealed average does not .
equation ( [ eq : cavity_averaged ] ) is typically hard to solve , since it is a self - consistent equation for a distribution of distributions @xmath98 . for this purpose ,
we use a population dynamics algorithm ( shown in the next section ) , familiar to optimization problems @xcite .
this method consists to describe the distribution @xmath117 via a discrete sampling ( a ` population ' ) made of a large number of distributions @xmath45 . applying iteratively eq .
( [ eq : cavity_averaged ] ) , we find , if it exists , a fixed - point of the distribution @xmath98 .
it is interesting to discuss the different types of solutions expected from eq .
( [ eq : cavity_averaged ] ) . for a given contact network ,
if the system is satisfiable and underdetermined , hence admits an infinite set of solutions for force and torque balances , the distributions @xmath118 should be broad , allowing each force to take values in a non - vanishing range .
this means that on each contact , @xmath119 and @xmath120 should be broad and overlapping . if the system is neither under- nor overdetermined ( _ ie _ isostatic ) , there is only one solution to force and torque balances for every site @xmath34 , and each @xmath118 is a dirac @xmath97-function centered on the solution .
if the system is overdetermined or unsatisfiable , there is typically no solution to eq .
( [ eq : cavity_averaged ] ) , meaning that an algorithm designed to solve it would not converge . in practice , since we perform a population dynamics algorithm to average over all possible packings , if one starts with a set @xmath121 of broad distributions as a guess for the solution , both isostatic and overdetermined ensembles show that all @xmath121 shrink to @xmath97-distributions after a few iterations , while underdetermined ensemble always gives broad ( not vanishing ) probabilities . therefore the threshold @xmath4 of the satisfiability / unsatisfiability transition for force transmission can be located by measuring the width of the force distributions @xmath121 .
the location of this transition , in turn , constitutes a lower bound for the possible coordination number @xmath4 , which extends maxwell s counting argument for @xmath30 to any friction .
an additional quantity available is the force distribution itself , as a function of @xmath4 .
therefore , our approach explicitly relates two essential properties of the jamming transition : the average coordination number and the force distribution . as discussed in the previous section , satisfiability of force
/ torque balance is studied via the behavior of the probability distribution @xmath98 of the distributions @xmath122 , defined by eq .
( [ eq : def_qofq ] ) .
this quantity is obtained for a given graph ensemble , defined by the distribution of connectivity ( contact number ) @xmath123 and the joint probability distribution of the contact directions @xmath124 around every interaction node ( particle ) , via the cavity equations eq .
( [ eq : cavity_eq ] ) and eq .
( [ eq : cavity_averaged ] ) . equation ( 8) is a self - consistent equation for @xmath98 .
self - consistent equations can generally be solved by an iterative algorithm , and this is the method we use here .
however , there is one difficulty arising from the fact that @xmath98 , which is a distribution of distributions , is a complex object in itself and is hard to simply describe in a computer program .
this is a recurring problem of solving cavity equations , and the solution developed to overcome this difficulty is called population dynamics algorithm .
population dynamics algorithm solves the problem of representing @xmath98 by describing it by a large sample drawn from @xmath117 .
this sample @xmath125 is called a ` population ' , and it matches @xmath98 in the sense that it contains more elements close ( in a function norm sense ) to @xmath126 than @xmath127 if @xmath128 .
more precisely , the approximation of @xmath98 it gives is @xmath129 $ ] .
this population of course needs to be as large as possible in order to give an accurate description of @xmath98 .
now , with this change of description , the algorithm needs to specify an iterative method ( a ` dynamics ' ) to make the population converge to the solution of eq .
( [ eq : cavity_averaged ] ) .
this is the heart of the population dynamics algorithm .
one starts from an initial population guess @xmath130 at iterative time @xmath131 .
the only requirements on those initial distributions are ( a ) repulsive normal force condition @xmath132 , ( b ) coulomb friction condition @xmath133 and ( c ) fixed pressure @xmath134 . in practice , we tested several initial population guesses ( gaussian , gaussian plus random noise , uniform on the region @xmath135\times [ 0 , \mu f^{\,n}]$ ] of the @xmath136 plane ) , with identical results at the end of the iteration procedure .
one then iterates in the following way at time @xmath137 , starting from @xmath131 : \(1 ) pick randomly a distribution @xmath138 in the current population .
\(2 ) pick a contact number @xmath139 within the distribution @xmath140 , and generate a set of @xmath139 random contact directions @xmath141 with the distribution @xmath124 .
\(3 ) pick @xmath101 distributions @xmath142 in the current population ( one can exclude the already chosen @xmath138 , but this is not necessary ) , and assign to each of them one of the directions @xmath143 chosen at step ( 2 ) .
\(4 ) generate a new field @xmath144 according to eq .
( [ eq : cavity_eq ] ) : @xmath145 and assign it to @xmath138 \(5 ) repeat operations ( 1 ) to ( 4 ) @xmath146 times , and increment the iteration time by @xmath147 .
now , rename the population as @xmath148 , \(6 ) test the convergence of the obtained @xmath98 .
again , due to the complexity of the object @xmath98 , testing the convergence is hard to do numerically , as it requires a very large population of @xmath149 to describe the @xmath45 space precisely enough .
however , in this work we will use the population dynamics only for determining the satisfiability threshold @xmath4 .
hence for our purpose in this work , we do not need to describe @xmath98 in detail , since we just need to know if it exists .
we thus adopt a simpler criterion to test the convergence of our population dynamics .
it is based on the convergence of the average width of the distribution @xmath149 .
if this width converges to a finite value , a solution to the cavity equations exists ( satisfiability ) , whereas if it vanishes , no solution exists ( unsatisfiability ) .
if the width have converged , we stop here , otherwise , we restart from step ( 1 ) for a new iteration .
one last technical difficulty is to perform the integration involved in eq .
( [ eq : cavity_eq ] ) during step ( 4 ) .
the integrand contains a delta - function via the @xmath150 constraint , coming from force and torque balance constraints .
this means the integrand is non - zero only on a set that has a dimension smaller than that of the integration domain .
a naive integration scheme would therefore never probe the integrand in this region .
the solution to this problem is a simple change of variable .
let us first rewrite the @xmath97-functions in the @xmath150 appearing in eq .
( [ eq : cavity_eq ] ) in a linear algebra form : @xmath151 where @xmath152 is a @xmath153 vector concatenating the @xmath101 normal and @xmath101 tangential forces , @xmath154 and @xmath155 , for @xmath156 , @xmath157 is a @xmath158 vector , and @xmath159 is a @xmath158-by-@xmath153 matrix encoding the geometrical configuration determined by the force directions @xmath160 and @xmath161 . the constraint appearing in the @xmath97-function of eq .
( [ eq : interaction_node ] ) is now expressed as linear algebra problem .
now , if @xmath162 , this linear problem is readily invertible , but if @xmath163 ( which is the most common situation ) , it is an underdetermined problem .
the vector @xmath164 is given , and we wish to integrate over the space of solutions @xmath152 of this underdetemined linear problem .
finding one of the infinitely many solutions to @xmath165 is done with a qr decomposition @xcite .
the qr decomposition is the fact that the @xmath153-by-@xmath158 rectangular matrix @xmath166 ( transpose of @xmath167 ) can be factorized as : @xmath168 where @xmath169 is an @xmath153-by-@xmath153 square orthogonal matrix , and @xmath170 is @xmath153-by-@xmath158 matrix whose top @xmath158-by-@xmath158 part @xmath171 is an upper triangular matrix and bottom part is identically zero .
it is easy to verify that a solution to the linear problem @xmath165 is given by : @xmath172 \\\ ] ] where @xmath173 is _ any _ @xmath174 vector .
this freedom of choice for @xmath173 corresponds exactly to the @xmath174 dimensional space of solution .
so , eventually , the cavity equation eq .
( [ eq : cavity_eq ] ) can be rewritten as : @xmath175 with a jacobian of @xmath147 .
obtaining a solution to eq .
( [ eq : cavity_averaged ] ) via the population dynamics algorithm as described above is _ a priori _ enough to get the force distribution through eq .
( [ fdis ] ) .
unfortunately , in practice the sampling provided by the population used in the population dynamics is too small to get a detailed description of the force distribution .
quite fortunately however , drastic simplifications of the self - consistent cavity equations eq .
( [ eq : cavity_eq ] ) and eq .
( [ eq : cavity_averaged ] ) occur at the isostatic point which allow us to access the force distribution in an easier way .
indeed , in an isostatic configuration , there exists a unique set of solutions of the contact forces for a given graph ( at a given pressure ) , meaning that the force probabilities @xmath118 on every contact are @xmath97-functions .
this , in turn , constraints the distributions @xmath176 to be also @xmath97-functions .
this trivializes the cavity equations eq .
( [ eq : cavity_eq ] ) to : @xmath177 using eq .
( [ eq : cavity_averaged ] ) and eq .
( [ fdis ] ) , we then obtain that the overall force distribution in an isostatic configuration satisfies : @xmath178 \chi_a(\ { f^{\,n},f^{\,t},\hat{n},\hat{t } \ } ) \end{array } \label{eq : p_of_f_at_transition}\ ] ] where the proportionality constant is just the normalization .
this is a self - consistent equation of the form @xmath179 $ ] that can be solved iteratively .
starting with an initial distribution ( ` guess ' ) @xmath180 , we iterate through @xmath181 $ ] , and obtain the force distribution via @xmath182 . in practice ,
a few tens of iterations are sufficient to obtain convergence .
in frictionless spheres packing in ( a ) 2-dimensional and ( b ) 3-dimensional systems .
results obtained from the cavity method ( open triangles ) show a flat regime ( in a log - log plot ) with exponent @xmath183 at small forces and @xmath184 in the intermediate region in both cases .
we use here as a contact number distribution @xmath185 . in inset , log - linear plot of the same distribution exhibits approximate exponential tail at large forces .
the red solid lines correspond to the fitting functions @xmath186 , as defined in the text .
( c ) comparison of force distributions obtained with narrower distributions @xmath71 ( at fixed average @xmath76 ) .
the exponent @xmath183 is obtained independently of the width of @xmath71 , but the peak in the force distribution becomes smaller when using a narrower @xmath71.,title="fig:",scaledwidth=48.0% ] in frictionless spheres packing in ( a ) 2-dimensional and ( b ) 3-dimensional systems .
results obtained from the cavity method ( open triangles ) show a flat regime ( in a log - log plot ) with exponent @xmath183 at small forces and @xmath184 in the intermediate region in both cases .
we use here as a contact number distribution @xmath185 . in inset ,
log - linear plot of the same distribution exhibits approximate exponential tail at large forces .
the red solid lines correspond to the fitting functions @xmath186 , as defined in the text .
( c ) comparison of force distributions obtained with narrower distributions @xmath71 ( at fixed average @xmath76 ) .
the exponent @xmath183 is obtained independently of the width of @xmath71 , but the peak in the force distribution becomes smaller when using a narrower @xmath71.,title="fig:",scaledwidth=48.0% ] in frictionless spheres packing in ( a ) 2-dimensional and ( b ) 3-dimensional systems .
results obtained from the cavity method ( open triangles ) show a flat regime ( in a log - log plot ) with exponent @xmath183 at small forces and @xmath184 in the intermediate region in both cases .
we use here as a contact number distribution @xmath185 . in inset , log - linear plot of the same distribution exhibits approximate exponential tail at large forces .
the red solid lines correspond to the fitting functions @xmath186 , as defined in the text .
( c ) comparison of force distributions obtained with narrower distributions @xmath71 ( at fixed average @xmath76 ) .
the exponent @xmath183 is obtained independently of the width of @xmath71 , but the peak in the force distribution becomes smaller when using a narrower @xmath71.,scaledwidth=96.0% ] we start by computing the force distribution @xmath187 for two and three - dimensional frictionless spheres packings , when we fix the average contact number @xmath188 and @xmath189 respectively , by solving eq .
( [ eq : p_of_f_at_transition ] ) .
results in fig .
[ fig : pf ] reproduce the force distributions similar as seen in numerical simulations @xcite and experiments @xcite .
the force distribution we obtain can be well fitted with @xmath190e^{-2.67x}$ ] for 2-d ( fig .
[ fig : pf]a ) and @xmath191e^{-2.65x}$ ] for 3-d ( fig .
[ fig : pf]b ) , with @xmath192 .
both fitting functions are close to the empirical fit @xmath193e^{-2.25x}$ ] to the force distribution of dense amorphous packings generated by lubachevsky - stillinger algorithm in 3-d by donev _
_ @xcite .
note that although the tails of the force distributions can be well fitted to exponential , claims of the precise form of the tails are difficult to conclude , as the presented data only varies over @xmath194 decades .
our method allows to access the small force region with unprecedented definition .
we gather data down to @xmath195 times the peak force ( which is of the order of the pressure ) .
this range is way below what is accessible with state of the art simulations of packings @xcite .
the reason for this is that we avoid two problems : ( i ) we work with eq .
( [ eq : p_of_f_at_transition ] ) directly in the thermodynamic limit , whereas simulated packings typically are limited to few tens of thousands of forces @xcite which limits the definition of the obtained force distribution , and ( ii ) we can work exactly at the jamming transition point , as we set @xmath196 , which contrasts with actual numerical or experimental studies where the limit of vanishing pressure with @xmath196 is very challenging .
the behavior of @xmath0 at small forces has recently attracted attention @xcite , due to its central role for the mechanical stability of packings @xcite .
lerner et al .
@xcite pointed out a relation between the small force scaling @xmath1 and the distribution function of the gaps @xmath197 between particles close to contact @xmath198 via the inequality @xmath199 .
few empirical data exist so far for the @xmath12 exponent .
some recent efforts greatly improved the data statistics @xcite , but however still widely disagree over the value of the exponent , which is found anywhere between @xmath200 and @xmath201 .
this calls for more insight from theory . here
we find for frictionless spheres a distribution of contact forces having a finite value for @xmath27 in the mean - field approximation .
this translates as an exponent @xmath28 over four decades of data in both 2-d and 3-d packings ( fig .
[ fig : pf ] ) .
we stress here that this observation is not dependent on the input distribution of coordination number @xmath71 .
the comparison of the force distributions we obtain with various shapes of @xmath71 is shown in fig .
[ fig : pf]c .
in particular , the exponent @xmath3 as well as the large - force tail stays the same when we change @xmath71 from the empirical fit @xmath73 ( at fixed average @xmath76 ) to _
e.g. _ a regular graph @xmath202 in 3-d packings .
however , the value of the exponent found in recent investigations @xcite seems to be incompatible with the value of @xmath12 obtained from our theory .
the exponent might be dependent on the protocol by which jammed packings are generated @xcite , although this point is still debated @xcite .
the value @xmath28 we obtain is an ensemble averaged , mean - field exponent .
the value of the mean - field exponent might be lost in a finite - dimensional packing , but the fact that the exponent seems to be dimension independent @xcite is an encouraging sign for a mean - field approach , as the mean - field value should be valid in high dimension .
we note that a very recent work by charbonneau et al .
finds an exponent @xmath17 in infinite dimensions from replica theory @xcite .
our exponent is incompatible with the relations @xmath203 derived by wyart @xcite , or @xmath199 derived by lerner et al @xcite , where @xmath204 is the exponent describing the distribution function of the gaps @xmath197 between particles close to contact @xmath198 .
those two different relations are obtained from considerations of mechanical stability against respectively extended or local ( buckling ) excitations . in 3-d ,
the local excitations seem to be the dominant ones , and thus only the relation @xmath199 should hold @xcite in that case . in our mean - field framework , we miss the relation between excitation modes and force distribution because we neglect the spatial structure of finite dimensional packings .
in fact , the pair correlation function in our framework has no structure beyond contact .
particle positions are omitted , apart from a local constraint of non overlaping , and there is therefore no excitation mode in our approach .
on the theoretical side , several values of @xmath12 have been predicted : the replica theory at the 1-rsb level also predicts a finite value for @xmath2 @xcite ( it predicts a gaussian for @xmath0 ) , but the fullrsb solution gives @xmath17 in infinite dimensions @xcite ; the q - model can give several values for @xmath12 , depending on the underlying assumptions , but in any case predicts @xmath205 @xcite ; and edwards model predicts @xmath206 @xcite .
the differences between edwards and q - model approaches , which mostly stem from the treatment of local disorder in the contact normal directions , indicate that the way to deal with this disorder is crucial for a correct description of the small force behavior .
interestingly , we find an intermediate regime for @xmath10 slightly smaller than its average value @xmath207 , for which @xmath208 ( fig . [ fig : pf]a , fig .
[ fig : pf]b ) .
we observe that this pseudo - exponent gets smaller when using a narrower @xmath71 as input ( fig .
[ fig : pf]c ) .
the smallest value it can reach for regular graph @xmath202 are @xmath209 in 2-d and @xmath210 in 3-d packings respectively . the results of this regime , extending from @xmath211 to @xmath212 is the one probed by most experiments and simulations .
this suggests that careful measurements at very small @xmath10 are needed to avoid pre - asymptotic behavior in the estimation of @xmath12 from experimental or numerical data , which might explain a lot of discrepancies observed in the current literature .
only few very recent numerical data managed to reach this regime @xcite . when ( a ) @xmath213 and ( b ) when @xmath214 for 2-d frictionless spheres packing . in ( a ) ,
the ` width ' of @xmath215 on @xmath216 is defined by @xmath217 as the difference of two extreme values of @xmath216 at which @xmath215 is equal to @xmath218 .
, title="fig:",scaledwidth=48.0% ] when ( a ) @xmath213 and ( b ) when @xmath214 for 2-d frictionless spheres packing . in ( a ) ,
the ` width ' of @xmath215 on @xmath216 is defined by @xmath217 as the difference of two extreme values of @xmath216 at which @xmath215 is equal to @xmath218 .
, title="fig:",scaledwidth=48.0% ] we turn to the determination of the force distribution for arbitrary friction coefficient @xmath5 and a lower bound on @xmath20 for the existence of random packings of spheres at a given @xmath5 .
this threshold corresponds to the point where solutions of the cavity equations eq .
( [ eq : cavity_eq ] ) and eq .
( [ eq : cavity_averaged ] ) no longer exist .
we search for the existence of a solution by applying the population dynamics algorithm described in section [ sec : pd ] .
a solution exists if this process leads to a converged averaged width of the distributions of the population @xmath149 used to describe @xmath98 .
even though this algorithm does not require a very large population ( as we do not seek to obtain a detailed description of @xmath98 and only want to know if it exists ) , its computational cost is still high , and we here apply the method to frictionless packings and 2-d frictional packings only . + vs. @xmath5 for various friction coefficients in 2-d spheres packing .
@xmath4 shows a monotonic decrease with increasing @xmath5 from @xmath219 to @xmath220 .
error bar indicates the range from the largest @xmath4 having no solution to the smallest @xmath4 having solution .
data point represents the mean of the range .
( b ) two power law scaling relations @xmath221 , @xmath222 are found with @xmath223 , @xmath224 , respectively .
, title="fig:",scaledwidth=48.0% ] vs. @xmath5 for various friction coefficients in 2-d spheres packing .
@xmath4 shows a monotonic decrease with increasing @xmath5 from @xmath219 to @xmath220 .
error bar indicates the range from the largest @xmath4 having no solution to the smallest @xmath4 having solution .
data point represents the mean of the range .
( b ) two power law scaling relations @xmath221 , @xmath222 are found with @xmath223 , @xmath224 , respectively . , title="fig:",scaledwidth=48.0% ] for frictionless packings , the ` width ' of a force distribution @xmath225 , denoted by @xmath217 , is defined as the difference of two extreme values of @xmath226 at which @xmath225 is equal to @xmath218 ( see fig .
[ fig : width]a ) . to determine the existence of solution
, we calculate the average width over the sites as @xmath227 .
[ fig : wt]a shows the evolution of the average width of distributions at different @xmath76 versus time step in population dynamics for the particular case of 2-d frictionless spheres packings .
results indicate that the final population of distributions @xmath228
after @xmath229 iterations have dramatically different shapes at various @xmath76 .
we find that the population @xmath149 rapidly tends to a set of non - overlapping dirac peaks when @xmath76 is small ( shown in fig .
[ fig : width]b for @xmath230 ) . in this case , the average width of fields decreases as a function of time step in the population dynamics iteration and finally goes below data discretization ( the interval of force to integrate the force distribution ) , leading to no solution of @xmath231 according to eq .
( [ eq : p2pto ] ) .
thus the system is overdetermined in the sense that @xmath76 contacts per particle are not enough to stabilize the whole packing .
in contrast , when @xmath76 is increased ( shown in fig . [
fig : width]a for @xmath232 ) , the distributions in @xmath149 become broad .
the average width converges to a finite value as seen in fig .
[ fig : wt]a , allowing a non - vanishing force distribution @xmath231 at each contact . in this case a set of solutions to force and torque balances exists .
the range between the largest @xmath76 having no solution and the smallest @xmath76 having solution is the range where @xmath4 belongs to .
in addition to 2-d frictionless packings , result of 3-d frictionless packings is shown in fig .
[ fig : wt]b .
the transition points @xmath233 are found in the expected range for both 2-d and 3-d frictionless packings . for frictional packings ,
the force distribution @xmath234 has the shape of a ` blob ' that spreads on the @xmath136-plane . in this case
the width of the force distribution @xmath217 and @xmath235 are defined as the width of the spread from side to side on axis @xmath216 and @xmath236 respectively with @xmath234 higher than @xmath218 : _ e.g. _
@xmath237 , where @xmath238 is the smallest value satisfies the condition @xmath239 and @xmath240 is the largest value satisfies the condition @xmath241 . here
we obtain data of average width @xmath227 and @xmath242 of distributions @xmath149 in 2-d frictional spheres packings with arbitrary friction coefficient .
similarly , the force distributions have different shapes after @xmath229 iterations , from which we determine , for example , the satisfiability transition point of force and torque balances in 2-d spheres packing at @xmath243 $ ] ( fig .
[ fig : wt]c ) and @xmath244 $ ] ( fig .
[ fig : wt]d ) .
the full curve @xmath4 is shown in fig .
[ fig : zmu]a for 2-d frictional spheres packing .
we observe a monotonic decrease with increasing @xmath5 from @xmath245 at @xmath246 , a well - known behavior of frictional packings , previously found in numerous studies , both experimentally and numerically @xcite .
notice that the critical contact number we obtain at infinite friction is slightly above the maxwell argument @xmath247 ; also a typical feature @xcite .
this is interesting , as it means that the naive counting argument , ignoring the repulsive nature of the forces , fails to reproduce the correct bound for such a simple case as sphere packings with @xmath248 , where neither coulomb condition nor non - trivial geometrical features complexify the picture .
the fact that the naive maxwell counting argument still gives the correct bound for frictionless sphere packing can therefore be seen as a quite fortunate isolated prediction . in fig .
[ fig : zmu]b , two power law scaling relations @xmath249 are found with @xmath223 , @xmath224 respectively .
our result @xmath250 agrees well with the prediction of 2-d monodisperse packing @xmath251 by wang _
@xcite , and is not far away from previous simulation of 2-d polydisperse packings @xmath252 @xcite , while @xmath253 is much smaller than their predicted value @xmath254 and the result of @xmath255 obtained from simulation @xcite .
+ similar to the frictionless case , the cavity method can generate the joint force distribution @xmath256 for frictional spheres packings with friction coefficient @xmath5 .
the simplest case of infinite friction , @xmath257 for 2-d and 3-d sphere packings are shown in fig .
[ fig : pff]a and fig .
[ fig : pff]d respectively , and results follow similar behavior as the empirical fitting formula @xmath258 , ( where @xmath259 , @xmath260 , @xmath261 and @xmath262 ) measured in previous numerical studies @xcite ( fig .
[ fig : pff]b and fig .
[ fig : pff]e ) .
in particular , we recover the non - trivial qualitative change between 2-d and 3-d @xcite : while in 3-d , @xmath263 , in 2-d , this symmetry is clearly broken .
this is a consequence of the more symmetrical role tangential and normal forces play in 3-d with as many torque balance as force balance equations , whereas in 2-d , there are twice less torque balance than force balance equations . when the friction coefficient is finite ( see fig . [
fig : pff02]a ) , the pattern inside the coulomb cone looks similar to the one obtained at infinite friction .
quite interestingly , we do not observe any excess of forces at the coulomb threshold @xmath264 , implying that there are no sliding contacts .
this offers a theoretical explanation to a singular fact already observed in simulations : control parameters ( essentially volume fraction and friction coefficient ) being equal , the percentage of plastic contacts in a packing depends on the preparation protocol @xcite .
our formalism takes into account those different packings ( hence protocols ) by performing a statistical average over possible packings , and the outcome shows that packings without plastic contacts are dominant . in this regards , the fragility associated with the large number of plastic contacts in many experimentally or numerically generated packings could be mostly attributed to the preparation protocol , rather than to an inherent property of random packings of frictional spheres .
furthermore , we obtain the distributions of the normal and tangential components for frictionless and frictional packings as plotted in fig . [
fig : pff]c , fig .
[ fig : pff]f and fig .
[ fig : pff02]b .
the normal force distributions all have slight peaks around the mean and approximate exponential tails at large forces . below the mean ,
the normal force distribution for infinite friction has a nonzero probability at zero force whereas it shows a dip towards zero for @xmath265 .
the tangential force distribution also has an approximate exponential tail , however , it decreases monotonically without an obvious rise at small forces .
our results of the probability distribution of normal forces and tangential forces agree with previous experimental measurements in 2-d frictional spheres packing @xcite . with friction coefficient @xmath266 in 2-d spheres packing .
( b ) plots of the probability distribution of normalized normal forces and frictional forces .
, title="fig:",scaledwidth=40.0% ] with friction coefficient @xmath266 in 2-d spheres packing .
( b ) plots of the probability distribution of normalized normal forces and frictional forces .
, title="fig:",scaledwidth=48.0% ]
in conclusion , we develop a theoretical framework by using the cavity method , introduced initially for the study of spin - glasses and optimization problems , to obtain a statistical physics mean - field description of the force and torque balances constraints in a random packing .
this allows us to get the force distribution and the lower bound on the average coordination number in frictional and frictionless spheres packings .
we find a mean - field signature of jamming in the finite value @xmath267 of the force distribution at small force .
we also notice that there is a power law rise @xmath268 in the intermediate region of @xmath0 .
thus it is likely that one obtains an exponent @xmath269 if the simulations or experiments can not achieve data down to low enough forces .
however , we must stress that the mean - field approach we develop does not include the detailed structure of finite dimensional systems , which may have an important effect on the @xmath12 exponent .
indeed some recent data seem to extend beyond the intermediate regime @xmath270 and find a finite exponent at small force @xcite . for frictional packings
, we can access the complete joint distribution @xmath271 . concerning the average coordination number , we describe its lower bound @xmath4 , which interpolates smoothly between the isostatic frictionless case @xmath272 , and a large @xmath5 limit @xmath273 slightly above @xmath247 .
this confirms that there is no discontinuity at @xmath246 .
we predict two scalings for small and large friction coefficients as @xmath274 and @xmath275 .
the statistical mechanics point of view on force and torque balances for random packings thus proves fruitful .
many features of these systems can be inferred from those simple considerations .
the use of the cavity technique enables us to tackle this problem with a correct treatment of the disorder , leading to several new results .
this formalism will be extended to packings of more general shapes and could be used to predict other properties of disordered packings , like the yield stress for instance .
granular materials are not the only systems subject to force and torque balances , and this constraint is seen in all overdamped systems , among which suspensions at low reynolds number constitute an important example , both conceptually and practically .
we hope that our results will motivate investigations in this direction .
here we give the statistical physics description of the packing problem motivating the definition of the partition function @xmath64 in eq .
( [ eq : partition_function ] ) of the main text .
we define for a given contact network @xmath276 , of @xmath54 particles and @xmath277 forces @xmath39 exerted on particle @xmath33 at position @xmath278 ( with respect to the center of the particle ) , an energy function which is the sum of the square of total forces and torques on each particle , @xmath279 we then define the partition function of contact network @xmath276 with quenched disorder on the contact normals @xmath59 , @xmath280 \mathrm{e}^{-\beta h } , \end{array}\ ] ] where @xmath253 is an inverse temperature , which is only here as a parameter ( it is not the actual temperature , which is irrelevant for granular packings ) , and will be set to @xmath281 soon .
the heaviside @xmath79 functions ensure repulsive normal forces , and coulomb friction condition , respectively .
the ground state of the hamiltonian @xmath282 at @xmath281 provides the balance condition in the packing .
consider the repulsive nature of contact forces and the coulomb condition when friction exists , by using the factor graph representation of the boltzmann probability , the weight of interaction node @xmath33 ( a particle on which force / torque balance is satisfied ) is @xmath283 \displaystyle \prod_{i \in \partial a } \theta(f_{i}^{\,n } ) \theta(\mu f_{i}^{\,n } - f_{i}^{\,n } ) , \end{array}\ ] ] with @xmath284 ensuring @xmath285 in the zero temperature limit @xmath281 , the above expression becomes the force / torque balance constraint on each node : @xmath286 notice that in this limit the exact shape of hamiltonian eq .
( [ h ] ) is irrelevant , as its only condition is that it provides the force and torque balances in the limiting case @xmath281 .
the associated partition function @xmath287 is then the one defined in eq .
( [ eq : partition_function ] ) of the main text .
this partition function is defined at the level of a single graph , but we study the problem for an ensemble of random graphs , _
ie _ we average the entropy @xmath288 over the selected ensemble of random graphs , defined by the distribution of connectivity ( contact number ) @xmath123 and the joint probability distribution of the contact directions @xmath124 around every interaction node ( particle ) , as explained in the main text .
this ensemble is .
we gratefully acknowledge funding by nsf - cmmt and doe office of basic energy sciences , chemical sciences , geosciences , and biosciences division . we thank f. krzakala and y. jin for interesting discussions . | the force distribution of jammed disordered packings has always been considered a central object in the physics of granular materials . however , many of its features are poorly understood .
in particular , analytic relations to other key macroscopic properties of jammed matter , such as the contact network and its coordination number , are still lacking . here we develop a mean - field theory for this problem , based on the consideration of the contact network as a random graph where the force transmission becomes a constraint optimization problem .
we can thus use the cavity method developed in the last decades within the statistical physics of spin glasses and hard computer science problems .
this method allows us to compute the force distribution @xmath0 for random packings of hard particles of any shape , with or without friction .
we find a new signature of jamming in the small force behavior @xmath1 , whose exponent has attracted recent active interest : we find a finite value for @xmath2 , along with @xmath3 .
furthermore , we relate the force distribution to a lower bound of the average coordination number @xmath4 of jammed packings of frictional spheres with coefficient @xmath5 .
this bridges the gap between the two known isostatic limits @xmath6 ( in dimension @xmath7 ) and @xmath8 by extending the naive maxwell s counting argument to frictional spheres .
the theoretical framework describes different types of systems , such as non - spherical objects in arbitrary dimensions , providing a common mean - field scenario to investigate force transmission , contact networks and coordination numbers of jammed disordered packings .
mechanically stable packings of granular media are important to a wide variety of technical processes @xcite .
one approach to characterize jammed granular packings is via the interparticle contact force network . in turn
, this network determines the probability density of inter - particle contact forces @xmath0 and the average coordination number @xmath9 .
while the force network has been studied for years , there is yet no unified theoretical framework to explain the common observations in granular packings , ranging from frictional to frictionless systems , from spherical to non - spherical particles and in any dimensions .
experimental force measurements @xcite and simulations @xcite have shown that the interparticle forces are inhomogeneously distributed with common features : @xmath0 near the jamming transition has a peak at small forces and an approximate exponential tail in the limit of large @xmath10 .
it is argued that the development of a peak is a signature of the jamming transition @xcite .
the scaling of @xmath0 in the limit of small forces is directly related to the mechanical stability of the packing @xcite : the contacts bearing small loads are the ones for which a local buckling is the most easily achieved @xcite , and they must thus exist in small enough density for the packing to be stable .
some parts of the qualitative behavior of @xmath0 are correctly captured by simplified mean - field models , such as the q - model @xcite and edwards model @xcite .
both of them describe the exponential decay at large force and a power law behavior @xmath11 for small forces @xcite . however , the exponent @xmath12 obtained by the q - model ( an integer @xmath13 @xcite ) is larger than the one obtained by recent numerical simulations accessing the low force limit with increasing accuracy @xmath14 @xcite , while the exponent obtained by edwards model @xmath15 depends strongly on the dimension ( through @xmath16 ) , when simulations show it does not @xcite .
a recent replica calculation finds an exponent @xmath17 in infinite dimensions of space @xcite .
other theoretical approaches based on entropy maximization similar to edwards statistical mechanics @xcite also recover the large force exponential decay @xcite .
some of those works which simulated the force network ensemble @xcite on strongly hyperstatic @xmath18 triangular lattice @xcite and on 2-d ( 3-d ) disordered hyperstatic frictionless bead packings generated using molecular dynamics @xcite however predict a force distribution that decays faster than exponential , and so does a mean - field theory based on replica theory of spin glasses @xcite .
recent measurements in the bulk of packings @xcite find downward curvature on semi - log plot of @xmath0 .
the average coordination number per particle @xmath16 is another key signature of jamming .
close to jamming , many observables ( pressure , volume fraction , shear modulus or viscosity @xcite to name a few ) scale with the distance @xmath19 to the average coordination number @xmath20 at the transition . understanding the value of @xmath20 is thus of primary importance .
the case of frictionless spheres , for which , @xmath21 where @xmath7 is the dimension , has been rationalized based on counting arguments leading to the isostatic conjecture @xcite : a lower bound @xmath22 is provided by maxwell s stability argument @xcite , and an upper bound @xmath23 is given by the geometric constraint of having the particles exactly at contact , without overlap .
the problem , however , turns out to be more complicated when friction is considered , and no method is known so far to predict @xmath20 in such a case .
the ( naive ) generalization of the counting arguments gives the bounds @xmath24 in frictional packings , independently of the value of the interparticle friction coefficient @xmath5 .
indeed , there is a range of @xmath20 obtained numerically and experimentally @xcite .
however , for small @xmath5 , this range never extends to the predicted lower bound , as packings with low friction coefficient lie close to @xmath25 . here
we present a theoretical framework at a mean - field level to consider force transmission as a constraint optimization problem on random graphs , and study this problem with standard tools , namely the cavity method @xcite .
we first obtain the force distribution for spheres , in frictionless and frictional cases , both in two and three dimensions .
besides showing the experimentally known approximate exponential fall - off at large forces , these distributions bring a new insight in the much less known small force regime . in particular
, we find for frictionless spheres in both 2 and 3 dimensions a finite value for @xmath26 in @xmath27 , leading to a mean - field exponent @xmath28 for small forces .
our framework is also the first one to show theoretically how the frictional coefficient @xmath5 affects the average number of contacting neighbors @xmath20 at the jamming transition , and we find a lower bound @xmath4 for this number ( which is also a lower bound on @xmath16 , since @xmath29 in a packing ) .
we achieve this by generalizing in a careful way the naive maxwell counting arguments , considering the satisfiability of force and torque balances equations .
linking @xmath20 to the behavior of force and torque balances is not a new idea , as it was already suggested by silbert _
et al_. @xcite .
furthermore the generalized isostaticity picture @xcite gives a bound on the number of fully mobilized forces ( _ ie _ the number of tangential forces which are taking the maximally value allowed by the coulomb law ) based on the value of @xmath20 .
however , none of these works derive a bound for @xmath20 itself , at a given @xmath5 .
the bound @xmath4 that we obtain interpolates smoothly between the two isostatic limits at @xmath30 and @xmath31 . |
over the past couple of decades , both observational and analytical work have suggested a physical connection between the formation and growth of galaxies and the growth of their central black holes .
for example , the comoving emissivity of active galactic nuclei ( agns ) and the cosmic star formation rate have both similarly declined since @xmath4 ( @xcite ; @xcite ; @xcite ; @xcite ) , which might imply a link between star formation and agn activity .
in addition , luminous agns are more often found in massive early - type galaxies with young stellar populations ( e.g. , @xcite ; @xcite ; @xcite ) . further support of a black hole
galaxy connection comes in the form of a number of correlations between properties of the host galaxies and the masses of their central black holes ( bhs ) .
a key relationship is between black hole mass ( @xmath5 ) and bulge stellar velocity dispersion ( @xmath6 ) , observed in both quiescent ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) and active galaxies ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
this relation was first predicted by @xcite and @xcite and has been explained by various analytic models
( e.g. , @xcite ; @xcite ; @xcite ) as well as recovered in numerical simulations of evolving and interacting galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
the @xmath0@xmath1 relationship can be used to infer @xmath0 in large samples of galaxies .
this allows for the exploration of the bh mass function on much larger scales ( e.g. , @xcite ) and thus helps investigate the role of bhs in galaxy formation and evolution processes .
direct @xmath0 measurements are made with stellar kinematics and gas dynamics , although these methods require good spatial resolution and are presently only feasible for nearby galaxies .
agns , however , offer the most robust tracer of the evolution of the bh population over much of the history of the universe . under the assumption that the motion of the gas in the broad line region ( blr ) of agns is dominated by the gravitational influence of the black hole
, one can use the virial relation @xmath7 to obtain @xmath5 , where @xmath8 is the average radius of the emitting gas in the blr , usually either determined with reverberation mapping ( e.g. , @xcite ) or estimated with the radius - luminosity relation ( e.g. , @xcite ) , @xmath9 is the velocity dispersion of the gas , deduced from the width of the emission line , and @xmath10 is a dimensionless factor that accounts for the unknown geometry and orientation of the blr and may be different for each agn . with current technology , we are unable to directly observe the structure of the blr , as it is unresolvable even with the largest telescopes , so the true value of @xmath10 for each object is unknown .
this has contributed significantly to the uncertainties in @xmath0 measurements using blr emission lines .
recent reverberation mapping ( rm ) efforts have begun to reveal more information about the actual structure of the blr and the value of @xmath10 in some objects ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
however , limited data for most agns requires the use of an average virial factor @xmath2 to estimate @xmath0 .
currently , @xmath2 is calculated with the assumption that agns follow the same @xmath0@xmath1 relation as quiescent galaxies ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
most estimates of @xmath11 are somewhat larger than @xmath12 ; @xcite find @xmath2 = 5.5 @xmath3 1.8 , @xcite find @xmath11 = 5.2 @xmath3 1.2 , and more recently , analysis by @xcite and @xcite both yield @xmath2 = 5.1 .
@xcite obtain a slightly lower value , @xmath2 = 3.8@xmath13 .
the difference between slopes and virial factors among studies using similar regression methods ( whether @xmath0 is considered the independent or dependent variable ) arise when different galaxy samples are used to determine these two quantities , which may suggest a morphological dependence or selection bias in the relation .
in fact , recent studies do report a morphological dependence in the quiescent @xmath0@xmath1 relation , such that there are systematic differences in the relation for early - type ( higher - mass ) and late - type ( lower - mass ) galaxies ( e.g. , @xcite ; @xcite ) .
others have found that barred galaxies lie systematically below the @xmath0@xmath1 relation of normal unbarred galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and still others have found deviations in both the slope and intercept for galaxies hosting pseudobulges ( e.g. , @xcite ; @xcite ; @xcite ) . the idea of a non - universal @xmath0@xmath1 relation has been supported by theoretical work as well ( e.g. , @xcite ; @xcite ) , which has also suggested that the relation may depend on environment .
morphological deviations from a single @xmath0@xmath1 relation have also been claimed in agns ( e.g. , @xcite ; @xcite ) , and there has been some question as to whether or not objects at the high - mass / high-@xmath1 end of the relation follow a different slope ( e.g. , @xcite ; @xcite ) .
for example , four out of the six objects with @xmath0 above 10@xmath14 _
@xmath15 included in the study of @xcite lie significantly above the relation .
the appearance of outliers could be due to systematic errors in @xmath1 or @xmath0 measurements , or simply a fluke due to small number statistics .
alternatively , @xcite suggest that offsets at the high - mass end may be due to a selection bias .
specifically , when a sample is selected based on agn properties , one is more likely to find a high - mass bh in a lower - mass galaxy ( based on a bh host galaxy correlation ) because high mass galaxies are rare and there is intrinsic scatter in bh host galaxy correlations .
an important step in evaluating the @xmath0@xmath1 relation and any possible deviations from it is to obtain secure @xmath1 and @xmath0 measurements in agns that sample the entire mass range of the relation .
while the high - mass end of the quiescent @xmath0@xmath1 relation is relatively well - populated to beyond 10@xmath16 _
@xmath15 ( @xcite ) , the current sample of agns used to calculate @xmath2 still contains just three or four objects with @xmath0 above 10@xmath14 _ @xmath15(@xcite ; @xcite ; @xcite ) .
more measurements for luminous agns are needed to measure the high - mass end of the agn @xmath0@xmath1 relation . however , accurate @xmath1 measurements for high - luminosity agns are difficult to obtain because the agn light overpowers the light from the host .
moreover , more luminous agns are relatively scarce and thus typically found at large distances , so the host galaxy has a small angular size and is easily lost in the glare of the agn .
it is only in the past few years that high - precision measurements in very luminous objects have been obtained on account of the availability of adaptive optics ( ao ) and integral field spectrographs ( ifus ) such as gemini north s near - infrared integral field spectrometer ( nifs ) combined with the gemini north laser guide star ao system , altair .
@xcite used nifs+altair to measure @xmath1 for pg1426 + 015 with much higher precision than previous measurements for high - luminosity quasars .
this success prompted us to undertake additional observations of quasars at the high - mass end of the @xmath0@xmath1 relation .
in this paper we present the results of our nifs observations of eight high - luminosity quasars .
we successfully measured @xmath1 in four objects and use these results to improve the population of the @xmath0@xmath1 relation at the high - mass end .
we also recalculate virial products for the entire agn sample with updated time lag measurements to re - derive @xmath2 , calibrate black hole masses in agns , and reexamine the agn @xmath0@xmath1 relation . in this work
we adopt a cosmological model of @xmath17 , @xmath18 , and @xmath19 km sec@xmath20 mpc@xmath20 .
observations of eight quasars were carried out at the gemini north telescope in 2008 and 2010 under the programs gn-2008b - q-28 , gn-2010a - q-11 , and gn-2010b - q-24 .
we chose our sample from the database of objects with rm - based black hole mass measurements from @xcite with @xmath0 @xmath21 10@xmath14 _ @xmath15 .
basic information on our targets is given in table [ table : tbl1 ] .
we used nifs in conjunction with the altair laser guide star ao system to carry out our observations .
nifs has a 3@xmath22 3 field of view that is divided into 29 individual spectroscopic slices , with a spectral resolution @xmath23 = @xmath24 @xmath25 5290 in both the @xmath26 and @xmath27 bands . with the ao correction ,
nifs yields a spatial resolution on the order of 0.1 .
there are several strong stellar absorption lines that fall within the wavelength range of the @xmath26 band , which has a central wavelength of 1.65 @xmath28 m and covers from about 1.49 @xmath28 m to 1.80 @xmath28 m , so we observed seven of our targets with the @xmath26 band filter .
we list the most prominent stellar absorption features in this wavelength region in table [ table : tbl3 ] .
we observed our eighth object , pg1700 + 518 , in the @xmath27 band due to its higher redshift .
the @xmath27 filter on nifs covers from about 1.99 @xmath28 m to 2.40 @xmath28 m .
we estimated the integration time for each object with @xmath29 acs or wfpc2 images of the sources from @xcite . to simulate the data we would obtain from nifs , we measured the flux within a 3x 3 aperture , except for a central circle of diameter 0.2 .
we estimated the exposure time for each object based on its brightness relative to pg 1426 + 015 , for which @xcite obtained a host - galaxy signal - to - noise ratio ( @xmath30 ) @xmath31 200 in about two hours of on - source integration . with both poisson and background - limited trials
, we estimated the integration time required for each object to obtain a @xmath30 @xmath31 200 .
table [ table : tbl2 ] gives details of the observations , most notably the on - source integration time for each object .
reconstructed images from the ifu spectra of all eight targets are shown in figure 1 .
we observed telluric standard stars ( usually a0v ) once every 1.5 hours for the purpose of telluric corrections .
data were processed through the standard nifs pipeline from the gemini iraf package .
our reductions deviated from the standard pipeline tasks in only two ways .
first , we found that the original sky frames did not adequately remove sky lines from our spectra . to remedy this , we manually scaled the individual sky spectra to obtain better sky subtraction in each individual object frame .
second , to remove stellar absorption lines in our telluric spectra , we used methods described by @xcite and applied in the idl - based code _ xtellcor_. this code uses a theoretical model of vega to remove the hydrogen features in our telluric standard spectra , and is specifically written for use with a0v stars . to separate the host - galaxy spectra from the agns
, we extract the spectrum from an annulus that excludes the quasar - dominated nucleus .
the use of ao in these observations allowed us in most cases to confine at least the core of the quasar flux to the very central pixels of the image .
generally , the ao - assisted seeing was on the order of @xmath32 , so we used either a @xmath33 , 0.3 , or @xmath34 inner radius ( @xmath35 ) to isolate the quasar component . the outer radius for each extraction annulus ( @xmath36 )
was chosen to include as much host galaxy light as possible while minimizing the amount of noise contributed by the sky . for most of our objects
, we evaluated this by eye and chose windows that minimized noise .
however , in the targets where we could see identifiable galactic absorption lines , we chose @xmath35 and @xmath36 to obtain the highest equivalent width measurements in the visible absorption lines . @xmath35 and @xmath36 for each object
are listed in table [ table : tbl4 ] and are shown on the reconstructed images in figure 1 .
the total galaxy+quasar spectra for all eight objects , extracted from within a radius of @xmath36 , are shown in the top panels of figure 2 .
although the use of the altair ao system helps confine the nuclear light to the central few pixels of the image , the ao - corrected psf still has nuclear light in its wings .
typical strehl ratios for the altair ao system are on the order of 0.1 to 0.3 in the @xmath26 band ( @xcite ) . for a typical strehl ratio of 0.2 , @xcite report a 50% encircled psf energy radius of about 0.4 ; i.e. , half of the quasar light falls in a radius outside 0.4 .
thus , there is still significant quasar contamination in the host - galaxy spectra . to further remove the quasar emission , we scaled and subtracted the nuclear spectrum from our annulus in each target .
the scaling was done empirically we chose a scale factor that best eliminated the most prominent quasar emission lines seen in the spectra . our final nucleus - subtracted , observed - frame spectra of all eight objects are shown in the bottom panels of figure 2 .
note that while the spectrograph coverage in the @xmath26 band extends from about 1.48 @xmath28 m to 1.8 @xmath28 m , in some cases the telluric contamination was sufficiently significant that we cropped the spectra before using them .
the wavelength coverage of the @xmath26-band spectra shown in figure 2 is not uniform for this reason .
we used the penalized pixel fitting method ( ppxf ) of @xcite to measure @xmath6 .
this method convolves a stellar template spectrum and a line - of - sight velocity distribution to model the host galaxy .
the best line - of - sight velocity distribution is calculated in the ppxf code with a @xmath37 minimization technique .
the velocity templates used to make our measurements were obtained by @xcite , and include stars of four different spectral classes : k0iii , k5iii , m1iii , and m5ia .
the k5iii , m1iii , and m5ia templates all resulted in somewhat similar fits in each spectrum the reduced @xmath38 values of the fit in these three cases were always very close to one another , with the k5iii template usually a slightly better fit than the other two .
the k0iii template provided a poor fit to the co(3 - 0 ) absorption line in all objects and was therefore not used in our analysis
. there are several different factors that affect the uncertainty in @xmath6 .
first , no single stellar template is expected to be a perfect match to the host - galaxy stellar absorption features .
we therefore adopt the average @xmath1 value from the three templates as our estimate , and fold the differences into our uncertainties .
one exception to this is mrk 509 , for which the k5iii template provided a significantly better fit .
we take the standard deviation in @xmath1 reported by the ppxf software among the three stellar templates as representative of the template mismatch uncertainty .
we also consider the location of @xmath39 = 1 in the fitting a component of our uncertainties . to do this , we allowed @xmath1 to vary but held the rest of the parameters fixed at their best - fit values .
we identified the value of @xmath1 at which the @xmath38 had changed by @xmath3 1 from the best - fit value , and took the average difference between these two values and our best - fit value as our uncertainty .
because our measurements for pg1411 + 442 and pg1617 + 175 were made with spectra that are significantly noisier than the other two spectra , we also include a signal - to - noise ( @xmath30 ) component in our uncertainties for those two objects .
to estimate this component , we degraded our two best spectra ( mrk 509 and pg2130 + 099 ) to match the @xmath30 in pg1617 + 175 and pg1411 + 442 . we then re - measured @xmath1 in the degraded spectra and took the deviation from our original measurements to represent the @xmath30 component of the uncertainties . for most of the templates
, the @xmath1 measurements tended to be overestimated in the degraded spectra by about 20 ^-1 km s@xmath20 .
we combine the 1@xmath40 uncertainties , the template mismatch uncertainties , and any @xmath30 component in quadrature and adopt these as our formal uncertainties .
we measured @xmath1 from the spectra of four of our objects : mrk 509 , pg1411 + 442 , pg1617 + 175 , and pg2130 + 099 , and these values are listed in table [ table : tbl4 ] .
the normalized galaxy spectra and the best - fit broadened stellar templates for these four objects are shown in figure 3 .
our measurement for pg2130 + 099 of 147 @xmath3 17 km s@xmath20 constitutes a significant improvement in precision over its previous measurement of 172 @xmath3 46 km s@xmath20 ( @xcite ) , while mrk 509 and pg1411 + 442 have no previous measurements reported in the literature .
our measurement for pg1617 + 175 of 201 @xmath3 37 km@xmath20 is slightly more precise but consistent with measurements made by @xcite , who report @xmath1 = 183 @xmath3 47 km s@xmath20 . in the other four objects
, we were unable to identify any absorption lines in our nucleus - subtracted spectra and were unable to fit stellar templates and recover @xmath6 .
we believe this was caused mainly by the overwhelming strength of the quasar emission in these objects and a lack of strong absorption lines within the observed wavelength range .
these four objects are also the most distant of the eight we observed . in an attempt to minimize the quasar contamination , we experimented with different radii for both the inner and outer regions in these objects , but in all cases were unable to remove the quasar contamination enough to see absorption in the host galaxy spectrum . even when subtracting off a scaled nuclear spectrum to eliminate the emission lines , we see only noise ( with residual sky and telluric contamination
; see figure 2 ) in the spectra of objects in which we were unable to measure @xmath6 . because we optimized our extraction radii to obtain the best host - galaxy - to - quasar ratio , our @xmath1 measurements were made within different effective physical apertures for each target .
the measurements for mrk 509 , pg1411 + 442 , and pg1617 + 175 were made within an effective physical aperture of 1.25 kpc , 3.2 kpc , and 3.3 kpc respectively .
the effective radii ( @xmath41 ) for the spheroid components of these three objects from the surface decompositions of @xcite are 1.85 kpc , 5.05 kpc , and 3.3 kpc , so our apertures for these three objects lie between , or very close to , the commonly quoted aperture sizes of @xmath41 and @xmath42 . however , the measurement for pg2130 + 099 was made within 1.64 kpc , which is 4.3 times @xmath41 for this object .
we apply the relation derived by @xcite to determine the velocity dispersion within the effective radius ( @xmath43 ) for all four objects .
these values are listed in table [ table : tbl4 ] for comparison with our measurements . for mrk 509 , pg1411 + 442 , and pg1617 + 175
, there is very little difference between our measured @xmath1 and @xmath43 .
the value of @xmath43 for pg2130 + 099 is notably higher than our quoted value , but the change does not qualitatively affect our analysis .
we use the corrected @xmath1 measurements ( @xmath43 ) for the subsequent discussion and analysis .
there are now 30 reverberation - mapped agns with @xmath1 measurements .
the virial products for most of these objects were calculated from time lags that were determined with traditional cross correlation methods ( e.g. , @xcite ; @xcite ) .
however , @xcite recently introduced a different method to determine time lags that they called stochastic process estimation for agn reverberation ( spear ) .
this method works as follows : we assume all emission - line light curves are scaled and shifted versions of the continuum light curve .
the continuum is modeled as an autoregressive process using a damped random walk ( drw ) model , which has shown to be a good statistical model of quasar variability ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the transfer function is modeled as a simple top - hat function .
we fit the continuum and emission - line light curves simultaneously , maximizing the likelihood of the model using bayesian markov chain monte carlo iterations .
the main advantage of the spear method is that it treats gaps in the temporal coverage of light curves in a statistically self - consistent way , so the gaps in the data are filled in a well - defined manner with well - defined uncertainties .
@xcite re - measured the time lags in reverberation - mapped agns with the spear method and demonstrated its ability to recover accurate time lags .
the method has since been successfully used to improve rm measurements ( @xcite ; @xcite ) and even recover velocity - delay maps ( @xcite ) .
because many of the light curves from the agn @xmath0@xmath1 sample , particularly high - luminosity objects , have large gaps , we went through and re - determined the virial products for the entire sample with updated h@xmath44 time lags recovered with the spear method .
eighteen of the objects already have h@xmath44 time lags from @xcite . for eight of the remaining objects
, we applied the same spear method with the latest version of the software called javelinyingzu / codes.html#javelin ] to calculate new time lags and virial products .
three of the remaining agns have recently - published virial products calculated with the spear method , so we use the published virial products for 3c390.3 from @xcite , 3c120 from @xcite , and pg2130 + 099 from @xcite .
we do not have the light curve for mrk 50 , so for this object we use the virial product from @xcite .
all recalculated time lags , original line widths , and updated virial products for each rm data set are given in table [ table : tbl5 ] .
for objects with just one measurement , we use that virial product . for objects with multiple measurements ,
we adopt the mean of the logarithm of the virial products . in table
[ table : tbl6 ] we show the adopted virial products and @xmath1 measurements for the entire sample . in most cases ,
the updated virial products are very similar to the previously - quoted values .
the median fractional change in the virial products due to the updated time lags is about 18% , and the majority of these changes are consistent to within the measurement errors . for eight of the objects ,
the virial products changed somewhat significantly .
the largest increase was for 3c390.3 , for which the new virial product increased by factor of almost five over the previous measurement .
most of these eight objects had light curves with significant gaps at key locations in the light curves that likely interfered with the time lag determinations .
we do not see any systematic increase or decrease in the fractional change as a function of virial product ( i.e. , the virial products did not increase more in objects with larger virial products , or vice versa , in general ) , but we do note that the virial products of both of the highest - mass objects , 3c390.3 and pg1426 + 015 , both increased substantially from previous measurements .
the objects for which we successfully measure @xmath1 ( pg1411 + 442 , pg1617 + 175 , mrk 509 , and pg2130 + 099 ) are the four lowest - redshift galaxies in our sample .
mrk 509 and pg2130 + 099 have the highest @xmath30 host - galaxy spectra , with an average @xmath30 per pixel of @xmath31 250 for mrk 509 and @xmath31 190 for pg2130 + 099 .
these have the most easily identifiable galaxy absorption features in our sample ( see figure 3 ) .
the higher - luminosity quasars , pg1411 + 442 and pg1617 + 175 , have lower @xmath30 , specifically a @xmath30 of 100 per pixel in pg1617 + 175 and 130 per pixel in pg1411 + 442 , which is lower than our anticipated @xmath30 .
although we made several attempts to fit and remove sky features , residual sky contamination and telluric absorption lines remain in the subtracted spectra , which makes the velocity dispersion measurements more uncertain and contributes to the lower @xmath30 of the spectra .
we also see stronger quasar emission features which we were unable to eliminate entirely from the host - galaxy spectra .
there seems to be three main factors that compromise the quality of our host galaxy spectra .
first , we were unable to satisfactorily remove the sky emission from the spectra in all eight targets .
this caused a significant decrease in @xmath30 in all of our spectra .
secondly , in three of our objects , pg0026 + 129 , pg1126 + 023 , and pg0052 + 251 , many of the strong stellar absorption features that allow us to measure @xmath1 were redshifted out of the @xmath26-band .
this limited us to very few absorption lines , and these few remaining lines fell in regions with severe telluric contamination .
as such , we did not detect any absorption and were unable to measure @xmath1 in these objects .
third , and possibly most importantly , in these four objects the quasar contamination becomes strong enough to overwhelm the host galaxy flux despite our long integrations and attempts to optimize the extraction radius . to quantify the amount of quasar contamination remaining in the original extraction annuli of these spectra , we estimated the ratio of the quasar flux to the host flux , both within our extraction radius , for the case of pg0026 + 129 .
we based this calculation on measurements of the psf magnitude , host galaxy magnitude , and host galaxy sersic index reported by @xcite from their analysis of @xmath29 nicmos @xmath26-band images .
we integrate the sersic profile over the extraction annulus used in our study ( for pg0026 + 129 , we used an inner radius of 0.2and an outer radius of 0.6 , corresponding to 0.62 and 1.86 kpc , respectively ) to estimate the amount of host flux inside our aperture .
given the previously discussed findings of @xcite , we assume that half of the psf light falls inside the annulus and find that the psf flux inside the extraction annulus is a factor of 4.5 times the amount of host flux inside the extraction annulus .
we expect similar , possibly even more , contamination in the other targets for which we were unsuccessful , and thus this contamination limits our ability to explore the hosts of quasars at the high end of the luminosity distribution .
while we appear to have reached the limit of the nifs+altair system for these measurements , ao systems continue to move towards diffraction - limited resolution with high strehl ratios .
these advances may lead to successful measurements with similar exposure times .
the future availability of the @xmath45 ( @xmath46 ) may also lead to successful attempts at @xmath1 measurements in high - luminosity quasars .
@xmath46 is currently expected to launch in 2018 and will be equipped with an ifu spectrograph of sufficient resolution , and the major problems of sky and telluric contamination , which were prohibitive for our higher - redshift targets , will be completely eliminated in space . observing from space will allow us to see the whole spectrum continuously , so we will not be limited to specific redshift windows , and will also include the caii triplet region .
@xmath46 will offer a compact , stable psf , and will have more sensitivity than our current equipment and thus offers much promise for future efforts to measure @xmath1 in high - luminosity , high - redshift agns .
it is also possible that the host galaxies of the four quasars in which we were unable to measure @xmath1 were simply fainter galaxies with lower @xmath6 .
if this is the case , excluding them in our subsequent examination of the @xmath0@xmath1 relation could result in a bias in the distribution at the high-@xmath0 end . as a separate check on the expected velocity dispersions of these hosts ( as well as the rest of the sample ) , we place the entire rm sample on the @xcite relation ( hereafter the faber - jackson relation ) , which is the correlation between @xmath1 and the absolute magnitude or luminosity of the host bulge .
some of the galaxies had already - published absolute @xmath47-band magnitudes from @xcite , and we used host galaxy decompositions by @xcite and @xcite to determine the bulge magnitudes in the rest ( except mrk 50 , which was not included in these studies ) . to compare our sample with a previous determination of the faber - jackson relation by @xcite , we convert our magnitudes to the @xmath48 band .
the @xmath29 observations @xcite were taken using the acs f550 m filter and the wfpc2 f547 m filter , while the images used by @xcite were taken with the wfpc2 f814w filter .
we use the iraf package @xmath49 with bulge templates from @xcite to determine the @xmath50 and @xmath51 colors for each galaxy type and use this to transform our bulge magnitudes to the magnitude in the f814w filter , which is extremely close to the @xmath48 band .
figure 4 shows our rm sample on the faber - jackson relation from @xcite , which was calculated with the sample from @xcite .
the rm sample as a whole appears to follow this relation , with some scatter .
we use the same relation measured by @xcite with the sample from @xcite to calculate the expected @xmath1 values for the four objects in which we were unable to measure @xmath6 .
as figure 4 shows , the hosts of the four targets in wich we were unable to measure @xmath1 are some of the most luminous in the sample , and thus the predicted @xmath1 values are quite high .
as such , we conclude that our inability to measure @xmath1 in these objects is most likely due to the high agn luminosities and high redshifts of the systems and not due to systematically fainter host galaxies in these targets .
it is important again to note that measuring bulge properties in agn , particularly at the high - luminosity end of the distribution , is very difficult , and thus the bulge luminosities themselves are subject to large uncertainties .
because we have updated the virial products in the agn @xmath0@xmath1 sample and added a few objects at the high - luminosity end of the distribution , we also present an updated measurement of the average virial factor @xmath2 used to calibrate the agn @xmath0 scale . in order to measure @xmath2
, we assume that the agn @xmath0@xmath1 relation follows the same slope as quiescent galaxies .
however , previous studies have found that agns appear to follow a slightly shallower relation than quiescent galaxies ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
we measure the slope of the relation between the virial product , @xmath52 , and @xmath6 : @xmath53 with our updated agn sample .
we use the traditional forward regression for our calculation : we consider @xmath1 as the independent variable and @xmath0 the dependent variable . using the fitexy algorithm ( @xcite ) to fit the agn relation , we obtain @xmath54 .
this slope is steeper than found in previous studies ( e.g. , @xcite , who report @xmath44 = 3.46 @xmath3 0.61 ) , and is slightly flatter than but consistent with the most recent measurement of the slope in quiescent galaxies of 5.31 @xmath3 0.33 ( @xcite ) .
we then use fitexy to determine the @xmath11 necessary to place the agn sample on the same @xmath0@xmath1 relation as quiescent galaxies .
using our updated agn sample , we obtain log @xmath2 = 0.63 @xmath3 0.11 , corresponding to @xmath2 = 4.31 @xmath3 1.05 . this number is slightly lower than , but consistent with , the recent values found by @xcite and @xcite and is in closer agreement with that found by @xcite when using a forward regression analysis .
however , @xcite also raised the issue of a potential sample selection bias in the @xmath0@xmath1 relation specifically , there are no @xmath1 measurements in objects with @xmath0 @xmath55 _ @xmath15 , causing a bias against low-@xmath0 systems .
they use an inverse regression approach to avoid this selection bias .
we prefer a forward regression because the @xmath0@xmath1 relation is typically used to determine @xmath0 in galaxies with measured @xmath6 .
we use our new value of @xmath2 = 4.31 to transform the virial products of the agn sample to @xmath0 .
we place our four objects with successful @xmath1 measurements on the @xmath0@xmath1 relation in figure 5 , with the rest of the agn sample shown for comparison .
we also show the most recent quiescent @xmath0@xmath1 relation from @xcite .
we see that all four of our objects lie within the expected scatter of the relation , and see no evidence for an offset in the objects at the high end of the relation . because we were unable to obtain @xmath1 measurements for our highest-@xmath0 objects , we fall short of our original goal of densely populating the very high end of the @xmath0@xmath1 relation , though we do increase the sample of objects with 10@xmath14 _
@xmath15@xmath56 @xmath0 @xmath57 _ @xmath15 .
as noted in previous studies , the agn sample is still biased towards objects with relatively lower masses than the majority of the quiescent galaxies with dynamical mass measurements .
a morphological or environmental dependence of the @xmath0@xmath1 relation could have important implications on the use of a single mean @xmath10 factor to transform the virial product , @xmath52 , into @xmath0 in agns . to test
whether or not the agn sample shows any morphology - dependent effects , we divided our agn sample into different groups to see if there was any visible offset or change in slope in the @xmath0@xmath1 relation . in the left panel of figure 6 , we divide the agn sample into two groups : those suspected of hosting a pseudobulge , and those thought to host a classical spheroid at their centers . in the right panel
, we divide the sample based on whether or not a central bar is observed .
we classified the galaxies using the host - galaxy decompositions from @xcite ( see their table 4 ) .
following @xcite , the galaxies were inspected for the presence of the following pseudobulge indicators : flattened bulge morphology , the presence of a nuclear bar , a low ( @xmath58 ) sersic index , and the presence of copious dust and star formation in the nucleus without any signature of an ongoing merger .
none of these indicators by themselves constitute sufficient evidence that a bulge is a pseudobulge .
therefore , we assumed that galaxies with most of these indicators host a pseudobulge , and galaxies with few or none of these indicators host a classical bulge .
our classifications are listed in table [ table : tbl6 ] .
it should be noted that host galaxy classifications are particularly difficult to make in high - luminosity quasar hosts , and some galaxies in our sample show signs of disturbed morphologies , so our classifications are somewhat uncertain .
we do not see any signs of an offset in either the galaxies hosting pseudobulges or the galaxies with central bars , unlike several studies that were previously noted in our introduction .
if we determine @xmath2 with our samples of barred and unbarred galaxies separately , we obtain a result similar to that of @xcite .
namely , the galaxies with bars in them yield a lower value ( @xmath2 = 3.07 @xmath3 1.00 ) than the galaxies without central bars ( @xmath59 = 5.92 @xmath3 1.80 ) .
however , the barred sample lacks any objects with @xmath0 @xmath60 _ @xmath15and similarly , the unbarred sample has only three objects with @xmath0 below 10@xmath61 _ @xmath15 .
observations of the megamaser sample by @xcite show that about half of their host galaxies ( with @xmath0 @xmath62 _ @xmath15 ) fall below the @xmath0@xmath1 relation , which again demonstrates that different samples yield different results .
in addition , @xcite and this work have demonstrated that extending the sample to cover the entire @xmath0 range can significantly impact the measured slopes and mean virial factor , and as such we conclude the difference seen in our samples is not statistically significant due to the systematic uncertainties associated with the lack of a substantial dynamic range in @xmath5 .
we measured @xmath1 in four quasars at the high - mass end of the agn @xmath0@xmath1 distribution .
the measurements of mrk 509 and pg1411 + 442 are the first ever @xmath1 measurements for these targets , and the measurements for pg1617 + 175 and pg2130 + 099 are updates with improved precision .
we were unable to measure @xmath1 in the spectra of our highest - redshift objects due to substantial contamination of the host - galaxy spectra by the quasar nucleus and poor sky line subtraction .
future measurements of @xmath1 for high - luminosity qsos will require improved ao technology , substantially longer integration times , or @xmath46 .
we also updated the virial products for the agn @xmath0@xmath1 sample with recalculated time lags from the spear method of @xcite and recalculated the mean virial factor @xmath11 used to calibrate the agn @xmath0 scale .
we obtained log @xmath2 = 0.63 @xmath3 0.11 , corresponding to @xmath11 = 4.31 @xmath3 1.05 .
this is consistent with previous results based on forward regression methods . with our new @xmath1 measurements ,
all four of our objects fall within the expected scatter of the quiescent @xmath0@xmath1 relation .
we find no evidence in agns for a morphology - based deviation from the standard quiescent @xmath0@xmath1 relation .
the authors would like to thank jenny greene and alister graham for their helpful suggestions that improved this work .
c.j.g . would like to thank gisella de rosa for help with some of the code used in this analysis .
b.m.p . , r.w.p . , and c.j.g .
gratefully acknowledge the support of the national science foundation through grant ast-1008882 to the ohio state university .
is also supported by a presidential fellowship at the ohio state university .
p.m. is grateful for support from the sabbatical visitor program at the north american alma science center ( naasc ) at nrao and the hospitality of both the naasc and the university of virginia while this work was completed .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this work is based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) . , d. 1986 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
627 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
d. l. crawford , 733 lccc pg0026 + 129 & 00 29 13.6 & + 13 16 03 & 0.142 + pg0052 + 251 & 00 54 25.1 & + 25 25 38 & 0.154 + pg1226 + 023 & 12 29 06.7 & + 02 03 09 & 0.158 + pg1411 + 442 & 14 13 48.3 & + 44 00 14 & 0.089 + pg1617 + 175 & 16 20 11.3 & + 17 24 28 & 0.112 + pg1700 + 518 & 17 01 24.8 & + 51 49 20 & 0.292 + mrk 509 & 20 44 09.7 & -10 43 25 & 0.034 + pg2130 + 099 & 21 32 27.8 & + 10 08 19 & 0.063 [ table : tbl1 ] lccccr pg0026 + 129 & 2008b & 1.33 & h + pg0052 + 251 & 2010b & 1.33 & h + pg1226 + 023 & 2010a & 0.83 & h + pg1411 + 442 & 2010a & 1.33 & h + pg1617 + 175 & 2010a & 2.50 & h + pg1700 + 518 & 2010a & 3.00 & k + mrk 509 & 2008b & 2.00 & h + pg2130 + 099 & 2010b & 1.00 & h lcccccccc pg0026 + 129 & 0.2 & 0.6 & & & & & & + pg0052 + 251 & 0.2 & 0.7 & & & & & & + pg1226 + 023 & 0.4 & 1.4 & & & & & & + pg1411 + 442 & 0.2 & 1.6 & k5iii & 216 @xmath3 31 & 858.80 & 914 & 0.940 & 209 @xmath3 30 + pg1617 + 175 & 0.3 & 1.3 & m5ia & 201 @xmath3 37 & 491.33 & 649 & 0.757 & 201 @xmath3 37 + pg1700 + 518 & 0.4 & 1.1 & & & & & & + mrk 509 & 0.3 & 1.7 & k5iii & 189 @xmath3 12 & 1046.17 & 1134 & 0.923 & 184 @xmath3 12 + pg2130 + 099 & 0.2 & 1.2 & k5iii & 147 @xmath3 17 & 548.70 & 914 & 0.600 & 163 @xmath3 19 [ table : tbl4 ] lccccc 3c120 & 27.2@xmath63 & 1 & 1514 @xmath3 65 & 1 & 12.2@xmath64 + 3c390.3 & 44.3@xmath65 & 2 & 5455 @xmath3 278 & 2 & 260@xmath66 + ark 120 & 35.7@xmath67 & 3 & 1959 @xmath3 109 & 7 & 26.7@xmath68 + ark 120 & 29.7@xmath69 & 3 & 1884 @xmath3 48 & 7 & 20.6@xmath70 + arp 151 & 3.6@xmath71 & 4 & 1252 @xmath3 46 & 8 , 9 & 1.1@xmath72 + mrk 50 & 10.4@xmath73 & 5 & 1740 @xmath3 101 & 5 & 6.2@xmath74 + mrk 79 & 25.5@xmath75 & 3 & 2137 @xmath3 375 & 7 & 22.7@xmath76 + mrk 79 & 30.9@xmath77 & 3 & 1683 @xmath3 72 & 7 & 17.1@xmath78 + mrk 79 & 17.2@xmath79 & 3 & 1854 @xmath3 72 & 7 & 11.5@xmath80 + mrk 79 & 43.6@xmath81 & 3 & 1883 @xmath3 246 & 7 & 30.1@xmath82 + mrk 110 & 25.3@xmath83 & 3 & 1196 @xmath3 141 & 7 & 7.1@xmath84 + mrk 110 & 33.9@xmath85 & 3 & 1115 @xmath3 103 & 7 & 8.2@xmath86 + mrk 110 & 21.5@xmath87 & 3 & 755 @xmath3 29 & 7 & 2.4@xmath88 + mrk 202 & 3.5@xmath89 & 4 & 659 @xmath3 65 & 8 , 9 & 0.30@xmath90 + mrk 279 & 18.3@xmath91 & 3 & 1420 @xmath3 96 & 7 & 7.2@xmath63 + mrk 509 & 69.9@xmath92 & 3 & 1276 @xmath3 28 & 7 & 22.2@xmath93 + mrk 590 & 19.0@xmath94 & 3 & 789 @xmath3 74 & 7 & 2.3@xmath95 + mrk 590 & 19.5@xmath96 & 3 & 1935 @xmath3 52 & 7 & 14.2@xmath97 + mrk 590 & 32.6@xmath98 & 3 & 1251 @xmath3 72 & 7 & 9.9@xmath99 + mrk 590 & 30.9@xmath100 & 3 & 1201 @xmath3 130 & 7 & 8.7@xmath101 + mrk 817 & 20.9@xmath102 & 3 & 1392 @xmath3 78 & 10 & 7.9@xmath64 + mrk 817 & 17.2@xmath103 & 3 & 1971 @xmath3 96 & 10 & 13.0@xmath104 + mrk 817 & 35.9@xmath105 & 3 & 1729 @xmath3 158 & 10 & 20.9@xmath106 + mrk 817 & 10.8@xmath107 & 3 & 3150 @xmath3 295 & 10 & 20.9@xmath108 + mrk 1310 & 4.2@xmath109 & 4 & 755 @xmath3 138 & 8 , 9 & 0.5@xmath110 + ngc3227 & 10.6@xmath111 & 3 & 1925 @xmath3 124 & 10 & 7.7@xmath112 + ngc3227 & 4.4@xmath113 & 3 & 2018 @xmath3 174 & 10 & 3.5@xmath114 + ngc3516 & 14.6@xmath115 & 3 & 1591 @xmath3 10 & 10 & 7.2@xmath13 + ngc3783 & 7.3@xmath116 & 3 & 1753 @xmath3 141 & 7 & 4.4@xmath117 + ngc4051 & 2.5@xmath89 & 3 & 1034 @xmath3 41 & 10 & 0.5@xmath89 + lccccc 3c120 & classical & 12.2@xmath64 & 52.6@xmath118 & 162 @xmath3 20 & 1 + 3c390.3 & classical & 260@xmath66 & 1120.6@xmath119 & 273
@xmath3 16 & 2 + ark 120 & classical & 23.4@xmath120 & 100.9@xmath121 & 192 @xmath3 8 & 3 + arp 151 & classical & 1.1@xmath72 & 4.7@xmath122 & 118 @xmath3 4 & 4 + mrk 50 & classical & 6.2@xmath74 & 26.3@xmath123 & 109 @xmath3 14 & 5 + mrk 79 & barred pseudobulge & 19.2@xmath124 & 82.8@xmath125 & 130 @xmath3 12 & 2 + mrk 110 & pseudobulge & 5.2@xmath126 & 22.4@xmath127 & 91 @xmath3 7 & 6 + mrk 202 & classical & 0.3@xmath89 & 1.3@xmath128 & 78 @xmath3 3 & 4 + mrk 279 & pseudobulge & 7.2@xmath129 & 31.0@xmath130 & 197 @xmath3 12 & 2 + mrk 509 & classical & 22.2@xmath93 & 95.7@xmath131 & 184 @xmath3 12 & 7 + mrk 590 & pseudobulge & 7.3@xmath132 & 31.5@xmath133 & 189 @xmath3 6 & 2 + mrk 817 & barred pseudobulge & 14.6@xmath134 & 62.9@xmath135 & 120 @xmath3 15 & 2 + mrk 1310 & pseudobulge & 0.47@xmath136 & 2.2@xmath74 & 84 @xmath3 5 & 4 + ngc3227 & barred pseudobulge & 5.2@xmath137 & 22.4@xmath138 & 92 @xmath3 6 & 3 + ngc3516 & barred pseudobulge & 7.2@xmath13 & 31.0@xmath139 & 181 @xmath3 5 & 2 + ngc3783 & barred pseudobulge & 4.4@xmath117 & 19.0@xmath140 & 95 @xmath3 10 & 8 + ngc4051 & barred pseudobulge & 0.5@xmath89 & 2.2@xmath141 & 89 @xmath3 3 & 2 + ngc4151 & barred pseudobulge & 8.4@xmath142 & 36.2@xmath143 & 97 @xmath3 3 & 2 + ngc4253 & barred pseudobulge & 0.3@xmath110 & 1.3@xmath74 & 93 @xmath3 32 & 4 + ngc4593 & barred pseudobulge & 2.1@xmath144 & 9.1@xmath145 & 135 @xmath3 6 & 2 + ngc4748 & barred pseudobulge & 0.7@xmath110 & 3.0@xmath74 & 105 @xmath3 13 & 4 + ngc5548 & pseudobulge & 13.8@xmath146 & 59.5@xmath147 & 195 @xmath3 13 & 4 + ngc6814 & barred pseudobulge & 3.7@xmath95 & 15.9@xmath148 & 95 @xmath3 3 & 4 + ngc7469 & barred pseudobulge & 4.8@xmath149 & 20.7@xmath150 & 131 @xmath3 5 & 2 + pg1229 + 204 & barred pseudobulge & 16.0@xmath151 & 69.0@xmath152 & 162 @xmath3 32 & 9 + pg1411 + 442 & classical & 26.9@xmath153 & 115.9@xmath154 & 209 @xmath3 30 & 7 + pg1426 + 015 & classical & 373.3@xmath155 & 1609@xmath156 & 217 @xmath3 15 & 10 + pg1617 + 175 & classical & 118.6@xmath157 & 511.2@xmath158 & 201 @xmath3 37 & 7 + pg2130 + 099 & pseudobulge & 20.1@xmath159 & 86.6@xmath160 & 163 @xmath3 19 & 7 + sbs1116 + 583a & barred pseudobulge & 1.1@xmath95 & 4.7@xmath148 & 92 @xmath3 4 & 4 | we present new stellar velocity dispersion measurements for four luminous quasars with the nifs instrument and the altair laser guide star adaptive optics system on the gemini north 8-m telescope .
stellar velocity dispersion measurements and measurements of the supermassive black hole masses in luminous quasars are necessary to investigate the coevolution of black holes and galaxies , trace the details of accretion , and probe the nature of feedback . we find that higher - luminosity quasars with higher - mass black holes are not offset with respect to the @xmath0@xmath1 relation exhibited by lower - luminosity agns with lower - mass black holes , nor do we see correlations with galaxy morphology . as part of this analysis ,
we have recalculated the virial products for the entire sample of reverberation - mapped agns and used these data to redetermine the mean virial factor @xmath2 that places the reverberation data on the quiescent @xmath0@xmath1 relation . with our updated measurements and new additions to the agn sample ,
we obtain @xmath2 = 4.31 @xmath3 1.05 , which is slightly lower than , but consistent with , most previous determinations . |
we study the three - user multi - way relay channel ( mwrc ) with correlated sources , where each user transmits its data to the other two users via a single relay , and where the users messages can be correlated .
the mwrc is a canonical extension of the extensively studied two - way relay channel ( twrc ) , where two users exchange data via a relay @xcite . adding users to
the twrc can change the problem significantly @xcite .
the mwrc has been studied from the point of view of channel coding and source coding . in channel coding problems ,
the sources are assumed to be independent , and the channel noisy .
the problem is to find the capacity , defined as the region of all _ achievable _ channel rate triplets ( bits per channel use at which the users can encode / send on average ) . for the gaussian mwrc with independent sources , gndz et al .
@xcite obtained asymptotic capacity results for the high snr and the low snr regimes . for the finite - field mwrc with independent sources , ong et al .
@xcite constructed the _ functional - decode - forward _ coding scheme , and obtained the capacity region . for the general mwrc with independent sources ,
however , the problem remains open to date . in source coding problems ,
the sources are assumed to be correlated , but the channel noiseless .
the problem is to find the region of all _ achievable _ source rate triplets ( bits per message symbol at which the users can encode / send on average ) .
the source coding problem for the three - user mwrc was solved by wyner et al .
@xcite , using _ cascaded slepian - wolf _ source coding @xcite . in this paper , we study both source and channel coding in the same network , i.e. , transmitting correlated sources through noisy channels ( cf .
our recent work @xcite on the mwrc with correlated sources and orthogonal uplinks ) . for most communication scenarios ,
the source correlation is fixed by the natural occurrence of the phenomena , and the channel is the part that engineers are `` unwilling or unable to change '' @xcite .
given the source and channel models , we are interested in finding the limit of how fast we can feed the sources through the channel . to this end , define _ source - channel rate _ @xcite ( also known as bandwidth ratio @xcite ) as the average channel transmissions used per source tuple . our aim is then to derive the minimum source - channel rate required such that each user can _ reliably _ and _ losslessly _ reconstruct the other two users messages . in the multi - terminal network
, it is well known that separating source and channel coding , i.e. , designing them independently , is not always optimal ( see , e.g. , the multiple - access channel @xcite ) .
designing good joint source - channel coding schemes is difficult , let alone finding an optimal one .
gndz et al .
@xcite considered a few networks with two senders and two receivers , and showed that source - channel separation is optimal for certain classes of source structure . in this paper , we approach the mwrc in a similar direction .
we show that source - channel separation is optimal for three classes of source / channel combinations , by constructing coding schemes that achieve the minimum source - channel rate .
recently , mohajer et al .
@xcite solved the problem of linear deterministic relay networks with correlated sources .
they constructed an optimal coding scheme , where each relay injectively maps its received channel output to its transmitted channel input .
while this scheme is optimal for deterministic networks , such a scheme ( e.g. , the amplify - forward scheme in the additive white gaussian noise channel ) suffers from noise propagation in noisy channels and has been shown to be suboptimal for the mwrc with independent sources @xcite .
we consider the mwrc depicted in figure [ fig : mwrc ] , where three users ( denoted by 1 , 2 , and 3 ) exchange messages through a noisy channel with the help of a relay ( denoted by 0 ) .
for each node @xmath0 , we denote its source by @xmath1 , its input to the channel by @xmath2 , and its received channel output by @xmath3 .
we let @xmath4 , as the relay has no source .
we consider correlated and discrete - memoryless sources for the users , where @xmath5 , @xmath6 , and @xmath7 are generated according to some joint probability mass function @xmath8 the channel consists of a finite - field _ uplink _ from the users to the relay , which takes the form @xmath9 and a finite - field _ downlink _ from the relay to each user @xmath10 , which takes the form @xmath11 where @xmath12 , for all @xmath0 , for some finite field @xmath13 of cardinality @xmath14 with the associated addition @xmath15 . here
, @xmath16 can be any prime power .
we assume that the noise @xmath17 is not uniformly distributed , i.e. , its entropy @xmath18 ; otherwise , it will randomize the channel , and no information can be sent through .
each user sends @xmath19 source symbols to the other two users ( simultaneously ) in @xmath20 channel uses .
we refer to the @xmath19 source symbols of user @xmath21 as its message , denoted by @xmath22 , w_i[2 ] , \dotsc,$ ] @xmath23)$ ] , where each symbol triplet @xmath24 , w_2[u ] , w_3[u])$ ] for @xmath25 is generated independently according to .
the channel is memoryless in the sense that the channel noise @xmath17 for all nodes and all channel uses are independent , and the distribution @xmath26 is fixed for all channel uses .
the source - channel rate , i.e. , the number of channel uses per source triplet , is denoted by @xmath27 .
we assume that each user has all its @xmath19 source symbols prior to the @xmath20 channel uses source triplets and @xmath20 channel uses per unit time ) , we can always transmit in blocks .
we first wait for @xmath19 source triplets to be generated , and then use the channel @xmath20 times to transmit these source symbols ( while waiting for the next @xmath19 triplets generation ) , and so on . taking the number of blocks to be sufficiently large , the source - channel rate can be made as close to @xmath28 as desired . ] , and consider the following block code of source - channel rate @xmath28 : 1 .
the @xmath29-th transmitted channel symbol of each node @xmath30 depends on its message and its previously received channel symbols , i.e. , @xmath31 = f_{d , t } ( { \ensuremath{\boldsymbol{w}}}_d , y_d[1 ] , y_d[2 ] , \dotsc , y_d[t-1])$ ] , for all @xmath0 and for all @xmath32 .
each user @xmath21 estimates the messages of the other users from its own message and all its received channel symbols , i.e. , user @xmath21 decodes the messages from users @xmath33 and @xmath34 as @xmath35 , for all distinct @xmath36 .
we denote @xmath37,y_i[2],$ ] @xmath38)$ ] .
for source symbols or @xmath20 for channel symbols is clear from context . ]
note that utilizing _
feedback _ is permitted in our system model .
this is commonly referred to as the _ unrestricted _ mwrc ( cf .
the _ restricted _ mwrc @xcite ) .
we will see later that for the classes of source / channel combinations for which we find the minimum source - channel rate , feedback is not used .
this means that feedback provides no improvement to source - channel rate for these cases .
user @xmath21 makes a decoding error if @xmath39 .
we define @xmath40 as the probability that one or more users make a decoding error , and say that source - channel rate @xmath27 is _ achievable _ if the following is true : for any @xmath41 , there exists at least one block code of source - channel rate @xmath42 with @xmath43 .
the aim of this paper is to find the infimum of achievable source - channel rates , denoted by @xmath44 . for the rest of the paper
, we refer to @xmath44 as the minimum source - channel rate .
theoretical interest aside , the finite - field channel considered in this paper shares two important properties with the awgn channel ( commonly used to model wireless environments ) .
firstly , the channel is linear , i.e. , the channel output is a function of the sum of all inputs .
secondly , the noise is additive .
sharing these two properties , optimal coding schemes derived for the finite - field channel shed light on how one would code in awgn channels .
for example , the optimal coding scheme derived for the finite - field mwrc with independent sources @xcite is used to prove capacity results for the awgn mwrc with independent sources @xcite .
we will now state the main result of this paper .
the technical terms ( in italics ) in the theorem will be defined in section [ sec : definition ] following the theorem .
[ theorem : main ] the minimum source - channel rate is given by @xmath45 if the sources have any one of the following : 1 .
_ almost - balanced conditional mutual information _ , or 2 .
_ skewed conditional entropies _ ( on any _ symmetrical _ finite - field channel ) , or 3 . _
their common information equals their mutual information_. for cases 1 and 2 , we derive the achievability ( upper bound ) of @xmath44 using existing ( i ) slepian - wolf source coding and ( ii ) functional - decode - forward channel coding for independent sources .
we abbreviate this pair of source and channel coding scheme by sw / fdf - is .
we derive a lower bound using cut - set arguments . while the achievability for these two cases is rather straightforward
, what we find interesting is that using the scheme for independent messages is actually optimal for two classes of source / channel combinations .
furthermore , although the source - channel rates achievable using sw / fdf - is can not be expressed in a closed form , we are able to derive closed - form conditions for two classes of sources where the achievability of sw / fdf - is matches the lower bound . in sw
/ fdf - is , the source coding while compressing destroys the correlation among the sources , and hence channel coding for independent sources is used . for case 3 ,
the sources have their common information equal their mutual information , meaning that each source is able to identify the parts of the messages it has in common with other source(s ) . for this case
, we again use slepian - wolf source coding , but we conserve the parts that the sources have in common .
we then design a _
channel coding scheme that takes the common parts into account . here
, the challenge is to optimize the functions of different parts that the relay should decode .
we show that the new coding scheme is able to achieve @xmath44 . for all three cases ,
the coding schemes are derived based on the separate source - channel coding architecture .
also , for cases 1 and 3 , @xmath44 is found when only the sources satisfy certain conditions , and this is true independent of the underlying finite - field channel , i.e. , any @xmath16 and any noise distribution . in this section ,
we define the technical terms in theorem [ theorem : main ] .
[ def : symmetrical ] a finite - field mwrc is _ symmetrical _ if @xmath46 otherwise , we say that the channel is _
we can think of @xmath47 as the noise level on the downlink from the relay to user @xmath21 .
so , a symmetrical channel requires that the downlinks from the relay to all the users are equally noisy .
we do not impose any condition on the uplink noise level , @xmath48 .
[ def : abcmi ] the sources are said to have _ almost - balanced conditional mutual information _
( abcmi ) if @xmath49 for all distinct @xmath36 .
otherwise , the sources are said to have _ unbalanced conditional mutual information_. putting it another way , for unbalanced sources , we can always find a user @xmath50 , such that @xmath51 for some @xmath52 and distinct @xmath53 .
[ def : sce ] sources with unbalanced conditional mutual information are said to have _ skewed conditional entropies _ ( sce ) if , in addition to , @xmath54 for the same @xmath55 as in .
[ def : sce ] sources with unbalanced conditional mutual information are said to have _ skewed conditional entropies _
( sce ) if , in addition to , @xmath56 for the same @xmath55 as in .
lastly , we define _ common information _ in the same spirit as gcs and krner @xcite . for two users ,
gcs and krner defined common information as a value on which two users can _ agree _ ( using the terminology of witsenhausen @xcite ) .
the common information between two random variables can be as large as mutual information ( in the shannon sense ) , but no larger .
the concept of common information was extended to multiple users by tyagi et al .
@xcite , where they considered a value on which all users can agree . in this paper
, we further extend common information to values on which different subsets of users can agree .
we now formally define a class of sources , where their common information equals their mutual information .
[ def : cc ] three correlated random variables @xmath57 are said to have their common information equal their mutual information if there exists four random variables @xmath58 , @xmath59 , @xmath60 , and @xmath61 such that @xmath62 for some deterministic functions @xmath63 , and @xmath64 @xmath65 we give graphical interpretations using information diagrams for sources that have abcmi and sce in appendix [ appendix : abcmi ] , and examples of sources that have abcmi and their common information equal their mutual information in section [ section : conclusion ] .
definitions [ def : abcmi ] and [ def : sce ] are mutually exclusive , but definitions [ def : cc ] and [ def : abcmi ] ( or [ def : cc ] and [ def : sce ] ) are not .
this means correlated sources that have their common information equal their mutual information must also have either abcmi , sce , or unbalanced mutual information without sce .
this leads to the graphical summary of the results of theorem [ theorem : main ] in figure [ fig : results ] .
the rest of this paper is organized as follows : we show a lower bound and an upper bound ( achievability ) to @xmath44 in section [ section : lower ] . in section [ section : upper-1 - 2 ] , we show that for cases 1 and 2 in theorem [ theorem : main ] , the lower bound is achievable . in section [
section : upper-3 ] , we propose a coding scheme that takes common information into account , and show the source - channel rate achievable using this new scheme matches the lower bound .
we conclude the paper with some discussion in section [ section : conclusion ] .
denote the rhs of as @xmath66 we first show that @xmath67 is a lower bound to @xmath44 . using cut - set arguments @xcite
, we can show that if source - channel rate @xmath42 is achievable , then @xmath68 , \label{eq : upper-2}\end{aligned}\ ] ] @xmath69 , \label{eq : upper-2}\end{aligned}\ ] ] for all distinct @xmath36 . here follows from mohajer et al .
( 11)(12 ) ) and follows from ong et al .
* section iii ) .
re - arranging the equation gives the following lower bound to all achievable source - channel rates @xmath42and hence also to @xmath44 : [ lemma : lower - bound ] for any three - user finite - field mwrc with correlated sources , the minimum source - channel rate is lower bounded as @xmath70 we now present the result of sw / fdf - is coding scheme that first uses slepian - wolf source coding for the noiseless mwrc with correlated sources @xcite , followed by functional - decode - forward for independent sources ( fdf - is ) channel coding for the mwrc @xcite .
this scheme achieves the following source - channel rates : [ lemma : achievability - separate ] for any three - user finite - field mwrc with correlated sources , sw / fdf - is achieves all source - channel rates in @xmath71 , where @xmath72 where @xmath73 is the set of real numbers .
so , the minimum source - channel rate is upper bounded as @xmath74 @xmath75 where @xmath73 is the set of real numbers .
so , the minimum source - channel rate is upper bounded as @xmath74 the proof is based on random coding arguments and can be found in appendix [ appendix : slepian - wolf - fdf ] .
the variables @xmath76 are actually the channel code rates , i.e. , the number of message bits transmitted by user @xmath21 per channel use . from lemmas [ lemma : lower - bound ] and [ lemma : achievability - separate ] , we have the following result : [ corollary : separation ] for a three - user finite - field mwrc , if @xmath77 , then @xmath78 , meaning that the minimum source - channel rate is known and is achievable using sw / fdf - is .
[ remark : separation ] the collection of source / channel combinations that satisfy corollary [ corollary : separation ] forms a class where the minimum source - channel rate is found , in addition to theorem [ theorem : main ]
. the challenge , however , is to characterize in closed form classes of source / channel combinations for which @xmath77 . for this , we need to guarantee the existence of three positive numbers @xmath79 and @xmath80 satisfying the inequalities in for every @xmath81 .
next , we will show that for cases 1 and 2 in theorem [ theorem : main ] , sw / fdf - is achieves all source - channel rates @xmath81 .
in this subsection , we will show that if the sources have abcmi , then @xmath77 .
since any @xmath82 relies on the existence of channel code rates @xmath83 , we first show the following proposition : [ proposition : existence ] consider sources with abcmi . given any source - channel rate @xmath84 , and any positive number @xmath85
, we can always find positive @xmath79 and @xmath80 such that @xmath86 it can be shown that choosing @xmath87 + \frac{\delta}{4 } , \label{eq : case-1-r1-r3}\ ] ] @xmath88 + \frac{\delta}{4 } , \label{eq : case-1-r1-r3}\end{aligned}\ ] ] for all distinct @xmath36 satisfies . the expression in the square brackets is non - negative due to the abcmi condition . with this result ,
we now prove case 1 of theorem [ theorem : main ] .
we need to show that any source - channel rate @xmath81 is achievable , i.e. , the source - channel rate @xmath89 for any @xmath85 , lies in @xmath71 . here , @xmath90 is independent of @xmath91 and @xmath92 . for a source - channel rate in ,
we choose @xmath93 as in . substituting into , the second inequality in is satisfied . also ,
imply the first inequality in .
hence , @xmath82 .
this proves case 1 in theorem [ theorem : main ] .
@xmath94 we need to show that if the sources have sce and the channel is symmetrical , then the source - channel rate in is achievable for any @xmath85 . recall that sources that have sce must have unbalanced conditional mutual information , for which we can always re - index the users as @xmath95 , @xmath96 , and @xmath97 satisfying for some fixed @xmath98 . for achievability in lemma [ lemma : achievability - separate ]
, we first show the existence of @xmath99 satisfying the following conditions : [ proposition : existence-2 ] consider sources with unbalanced mutual information . given any source - channel rate @xmath84 , and any positive number @xmath85 , we can always find positive @xmath100 and @xmath101 such that @xmath102 for @xmath98 defined in .
constraint implies the following : @xmath103 @xmath104 first , we can always choose a positive number @xmath105 as in .
in addition , we choose @xmath106 + \frac{\delta}{4 } , \label{eq : case-2-r2}\\ \kappa r_c & = h(w_c|w_a , w_b ) + \frac{1}{2 } \big [ i(w_b;w_c|w_a ) + i(w_a;w_c|w_b ) - i(w_a;w_b|w_c)\big ] + \frac{\delta}{4}. \label{eq : case-2-r3}\end{aligned}\ ] ] @xmath107 + \frac{\delta}{4 } , \label{eq : case-2-r2}\\ \kappa r_c & = h(w_c|w_a , w_b ) + \frac{1}{2 } \big [ i(w_b;w_c|w_a)\nonumber \\ & \quad + i(w_a;w_c|w_b ) - i(w_a;w_b|w_c)\big ] + \frac{\delta}{4}. \label{eq : case-2-r3}\end{aligned}\ ] ] substituting into , we get ; substituting into , we get . summing different pairs from , , and ,
we get .
furthermore , for a symmetrical channel , we can define @xmath108 so , for sce and for symmetrical channels imply that the source - channel rate in equals @xmath109 hence , we only need to show that the source - channel rate is achievable for any @xmath110 .
we first choose @xmath100 and @xmath101 as in , , and , respectively . from , we get @xmath111 - \frac{\delta}{2 } \label{eq : b } \\ & < \kappa \big[\log f - \max\ { h(n_0 ) , h_\text{downlink } \ } \big ] , \label{eq:2a-1}\\ \kappa ( r_a + r_b ) & = h(w_a , w_b|w_c ) + \frac{\eta}{2 } + \frac{\delta}{2 } \label{eq : c } \\ & \leq h(w_b , w_c|w_a ) - \frac{\eta}{2 } + \frac{\delta}{2 } \label{eq : d}\\ & < \kappa \big[\log f - \max\ { h(n_0 ) , h_\text{downlink } \ } \big ] , \label{eq:2a-2}\\ \kappa ( r_a + r_c ) & = h(w_a , w_c|w_b ) + \frac{\eta}{2 } + \frac{\delta}{2 } \label{eq : e}\\ & \leq h(w_b , w_c|w_a ) - \frac{\eta}{2 } + \frac{\delta}{2 } \label{eq : f } \\ &
< \kappa \big[\log - \max\ { h(n_0 ) , h_\text{downlink } \ } \big ] , \label{eq:2a-3}\end{aligned}\ ] ] where and follow from ; , , and follow from . , which is determined by the sources correlation , is strictly greater than zero ; see its definition in .
] this means the second inequality in is satisfied . from , we know that the first inequality in is also satisfied . hence , the source - channel rate is indeed achievable for any @xmath110 .
@xmath94 in this section , we give an example showing that sw / fdf - is can be suboptimal .
consider the following sources : @xmath112 , @xmath113 , and @xmath114 , where @xmath115 is uniformly distributed in @xmath116 , @xmath117 is uniformly distributed in @xmath118 , and @xmath119 are each uniformly distributed in @xmath120 . in addition , all @xmath121 and @xmath122 are mutually independent . here
, each @xmath122 represents common information between @xmath123 and @xmath124 . for the channel ,
let the finite field be @xmath125 and @xmath15 be modulo-@xmath126 addition , i.e. , @xmath127 .
furthermore , let @xmath128 for @xmath129 , and @xmath130 for @xmath131 ; let @xmath132 for @xmath133 , and @xmath134 for @xmath135 , for all @xmath10 . for this source and channel combinations , we have @xmath136 , @xmath137 , @xmath138 @xmath139 , @xmath140 , @xmath141 , @xmath142 , for all @xmath10 .
one can verify that these sources have unbalanced conditional mutual information and do not have sce . in this example ,
suppose that @xmath144 is achievable using sw / fdf - is .
from lemma [ lemma : achievability - separate ] , there must exists three positive real numbers @xmath145 , @xmath146 , and @xmath80 such that @xmath147 @xmath148 from , we must have that @xmath149 and @xmath150 .
these imply @xmath151 . hence , and can not be simultaneously true .
this means the source - channel rate 1.05 is not achievable using sw / fdf - is .
the sources described here have their common information equal their mutual information .
we will next propose an alternative scheme that is optimal for this class of sources .
the following proposed scheme achieves all source - channel rates @xmath152 for this source / channel combination , meaning that the minimum source - channel rate for this example is @xmath153 .
so sw / fdf - is is strictly suboptimal for this source / channel combination .
while the achievability for cases 1 and 2 uses existing source and channel coding schemes , for case 3 ( i.e. , sources that have their common information equal their mutual information ) , we will use an existing source coding scheme and design a new channel coding scheme to achieve all source - channel rates @xmath81 .
while case 3 in general requires a new achievability scheme , these sources may have abcmi ( as shown in figure [ fig : results ] ) .
for such cases , optimal codes can also be obtained using the coding scheme for case 1 . in this section , without loss of generality , we let @xmath154 this means we can re - write @xmath67 as follows : @xmath155 as mentioned earlier , we will use a separate - source - channel - coding architecture , where we first perform source coding and then channel coding .
we will again use random coding arguments . more specifically
, we will use random linear block codes for channel coding .
we encode each @xmath156 , v_\mathcal{s}[2 ] , \dotsc , v_\mathcal{s}[m])$ ] to @xmath157 , which is a length-@xmath158 finite - field ( of size @xmath16 ) vector , for all @xmath159 ( see definition [ def : cc ] for the definition of @xmath160 ) .
we also encode each @xmath161 to @xmath162 , which is a length-@xmath163 finite - field vector .
so , each message @xmath161 is encoded into four subcodes , e.g. , @xmath164 is encoded into @xmath165 . some subcodes the common parts are shared among multiple sources . using the results of distributed source coding @xcite , if @xmath19 is sufficiently large and if @xmath166 then we can decode @xmath165 to @xmath164 , @xmath167 to @xmath168 , and @xmath169 to @xmath170 with an arbitrarily small error probability .
we show the proof in appendix [ app : gk - source - coding ] .
after source coding , user 1 has @xmath171 .
in order for it to decode @xmath172 , it must receive @xmath173 from the other users through the channel .
similarly , users 2 and 3 must each obtain subcodes that they do not already have through the channel .
in contrast to the source coding used for cases 1 and 2 , here , we have generated source codes where the users share some subcodes .
so , instead of using existing fdf - is channel codes ( designed for independent sources ) , we will design channel codes that take the common subcodes into account . [
cols="^,^,^,^,^,^,^,^",options="header " , ] the messages in each column are transmitted simultaneously .
note that in the third row of messages in the table , @xmath174 is split into two parts @xmath175 if and only if @xmath176 .
else , the entire message @xmath174 will be transmitted in the first column , i.e. , together with @xmath177 . since @xmath178
, the message @xmath175 can always fit into the first and third columns , with the remaining `` space '' padded with zero , denoted by @xmath179 .
the message @xmath180 is transmitted in a similar way .
the relay decodes the modulo addition of the messages in each column . if @xmath20 is sufficiently large and if is satisfied , then the relay can reliably decode its intended messages , i.e. , @xmath181 .
the relay broadcasts @xmath182 on the downlink . using lemma [ lemma : downlnk ]
, we can show that if , , and are satisfied , then each user can reliably decode @xmath182 , from which it can recover the messages of the other two users .
this completes the proof for the achievability of case 3 in theorem [ theorem : main ] .
note that the lower bound in lemma [ lemma : lower - bound ] and the achievable source - channel rates in lemma [ lemma : achievability - separate ] are applicable to the general finite - field mwrc with correlated sources , in addition to cases 1 and 2 in theorem [ theorem : main ] .
however , the coding technique in section [ section : upper-3 ] is useful only for sources that have their common information equal their mutual information . besides the coding schemes considered in this paper
, one could treat the uplink ( as a multiple - access channel with correlated sources ) and the downlink ( as a broadcast channel with receiver side information ) separately to get potentially different achievable source - channel rates . in this case , on the uplink , we let the relay decode all three users messages @xmath183 .
an achievable channel rate region for the two - sender multiple - access channel with correlated sources was found by cover et al .
then , on the downlink , we can use the result by tuncel @xcite for the relay to transmit @xmath184 to the users taking into account that each user @xmath21 has side information @xmath185 . extending the results of cover et al .
* theorem 2 ) to three senders , for the relay to be able to reliably decode @xmath183 on the uplink , @xmath42 must satisfy the following : @xmath186 , \\
\rightarrow & & \kappa & > \frac{h(w_1,w_2,w_3)}{\log f - h(n_0)}. \label{eq : cover - tuncel}\end{aligned}\ ] ] @xmath187 , \\
\rightarrow & & \kappa & > \frac{h(w_1,w_2,w_3)}{\log f - h(n_0)}. \label{eq : cover - tuncel}\end{aligned}\ ] ] for the case where each user has non - zero message , @xmath188 for all @xmath21 .
comparing to , we see that this coding strategy is strictly suboptimal for all cases 13 in theorem [ theorem : main ] when @xmath189 .
however , this strategy _ may _ achieve better ( i.e. , lower ) source - channel rates than sw / fdf - is in general .
we leave the derivation of the rates to the reader it is straightforward given the results by cover et al . and tuncel . in this paper
, we have only investigated the separate - source - channel coding architecture without feedback .
though we have identified source / channel combinations where the minimum source - channel rate is found , the problem remains open in general .
two directions which one could explore are _ joint _ source - channel coding and _
feedback_. we now give a numerical example where none of the coding schemes in this paper achieves the lower bound @xmath67 .
let the sources be @xmath190 , @xmath191 , and @xmath192 where all @xmath193 are independent random variables , and @xmath194 denotes modulo - two addition .
we choose @xmath195 , @xmath196 , @xmath197 , and @xmath198 . for this choice
, we have @xmath199 , @xmath200 , @xmath201 . for the channel , let @xmath202 , which gives @xmath203 .
we choose @xmath204 and @xmath205 for @xmath206 , giving @xmath207 . for each @xmath10 , we set @xmath208 , @xmath209 , and @xmath210 for @xmath211 , giving @xmath212 .
this means , @xmath213 , and @xmath214 for all @xmath10 . in this example , @xmath215 .
if @xmath216 is achievable , then ( from lemma [ lemma : achievability - separate ] ) we must be able to find some @xmath146 and @xmath80 satisfying @xmath217 , and @xmath218 . since the conditions can not be simultaneously met , @xmath216 is not achievable using sw / fdf - is . as the sources common information does not equal their mutual information , we can not use the coding scheme derived in section [ section : upper-3 ] .
this example shows that the separation schemes derived in this paper can not achieve the lower bound in some cases .
however , to show that separation is suboptimal , one has to explore all possible separation schemes and show that some joint source - channel scheme achieves a better source - channel rate . in this paper ,
we have identified three classes of source / channel combinations where the minimum source - channel rate is found .
the first class is sources that have almost - balanced conditional mutual information ( abcmi ) .
an example of sources that have abcmi is interchangeable random variables in the sense of chernoff and teicher @xcite , where `` every subcollection of the random variables has a joint distribution which is a symmetric function of its arguments . ''
this can model sensor networks where the sensors are equally - spaced on a circle to detect a phenomenon occurring at the center of the circle .
however , the abcmi conditions are looser than that of interchangeable random variables as the former only requires that mutual information between any two sources has _ roughly _ the same value ( see appendix [ appendix : abcmi ] ) , and also , the marginal distribution of the variables can be vastly different .
another class for which we have derived the minimum source - channel rate is when the sources have their common information equal their mutual information .
an example is correlated sources in the sense of han @xcite , where the sources can be written as @xmath219 , @xmath220 , and @xmath221 , where @xmath222 @xmath223 and @xmath224 are _ mutually independent _ random variables . using sensor networks as an example again , each node here has multiple sensing capabilities , e.g. , temperature , light , sound . as these measurements display different behavior spatially , some remain constant across subsets of sensors , e.g. , nodes 1 and 2 always sense the same temperature but different light intensity . as for the class of sources with skewed conditional entropies , the conditions appear to be purely mathematical in nature .
the authors would like to thank roy timo for discussions on gcs and krner s common information and other helpful comments .
figure [ fig : case1 - 2 ] shows the relationship among the entropies and mutual information for the three source messages @xmath5 , @xmath6 , and @xmath7 for the cases described above .
referring to figure [ fig : case-1 ] , the shaded areas represent the mutual information between any two source messages given the third source message .
for abcmi , we have that any of the three shaded areas must not be bigger than the sum of the other two shaded areas .
suppose that the sources do not have abcmi , then they must have unbalanced conditional mutual information , i.e. , we can find a user @xmath95 where @xmath225 is larger than the sum of @xmath226 and @xmath227 by an amount @xmath98 ( see figure [ fig : case-2a ] ) .
in addition , for sources that have sce , we also have that for the two messages , @xmath228 and @xmath229 , whose mutual information conditioned on @xmath230 , i.e. , @xmath225 , is larger than the sum of the other two pairs by the amount @xmath55 , their entropy conditioned on @xmath230 , i.e. , @xmath231 , is also greater than that of any other pair ( conditioned on the message of the third user ) by at least @xmath55 . the information diagram for sce is depicted in figure [ fig : case-2a ] .
we first quote two existing results of ( i ) channel coding for the three - user mwrc with independent messages and ( ii ) source coding with side information .
the following channel - coding setting assumes that the source messages are independent : [ lemma : channel - coding ] consider the mwrc depicted in figure [ fig : mwrc ] , where each user @xmath21 s message @xmath123 is uniformly distributed in @xmath232 ( we consider a single copy , i.e. , @xmath233 ) , and where @xmath5 , @xmath6 , and @xmath7 are independent .
the channel is used @xmath20 times according to the block code ( with @xmath233 ) specified in section [ sec : model ] .
each user can then decode the messages of the other two users ( with its @xmath20 received channel symbols and its own message ) with an arbitrarily small error probability if @xmath20 is sufficiently large , and if @xmath234 for all distinct @xmath36 .
the following source - coding setting assumes that the channel is noiseless : [ lemma : source - coding ] consider only the three users with their respective length-@xmath19 messages @xmath235 generated according to ( [ eq : source ] ) .
each user @xmath21 encodes its messages @xmath185 to an index @xmath236 , and gives its index to the other two users .
each user can then decode the messages of the other two users ( with the received indices and its own message ) with an arbitrarily small error probability if @xmath19 is sufficiently large , and if @xmath237 for all distinct @xmath36 .
note that wyner et al .
@xcite derived a similar result with an additional constraint on the relay . in their setup , the users present their indices to a relay ; the relay in turn re - encodes and presents its index to the users .
we use slepian - wolf source coding .
each user @xmath21 encodes its length-@xmath19 message @xmath185 to an index @xmath238 , satisfying and , for @xmath10 .
each user @xmath21 randomly generates a dither @xmath239 uniformly distributed in @xmath240 , and forms its encoded message @xmath241 .
the dithers are made known to all nodes .
now , @xmath242 , @xmath243 , and @xmath244 are mutually independent , and each @xmath245 is uniformly distributed in @xmath240 .
we then use fdf - is channel coding for the users to exchange the encoded independent messages @xmath242 , @xmath243 , and @xmath244 via the relay in @xmath20 channel uses . from lemma
[ lemma : channel - coding ] , if is satisfied , then each user @xmath21 can reliably recover @xmath246 and @xmath247 . knowing the dithers , it can also recover @xmath248 and @xmath249 .
from lemma [ lemma : source - coding ] , if and are satisfied , then each user @xmath21 can reliably recover @xmath250 . noting that @xmath27 and defining @xmath251 , the conditions for achievability , i.e. , , , and , can be expressed as .
we perform source coding for correlated sources @xcite@xcite . consider the source message @xmath164 . clearly , @xmath252 , since @xmath253 are deterministic functions of @xmath5 . now since @xmath58 captures all information that @xmath5 and @xmath6 have in common ( because @xmath254 ) , we have @xmath255 .
similarly @xmath256 .
so , we can reliably reconstruct @xmath164 from @xmath171 if @xmath19 is sufficiently large , and if the following inequalities hold ( * ? ? ? * theorem 2 ) : @xmath257 \log f ) /m & > h(w_1 , v_{12 } | v_{13 } , v_{123 } ) = h(w_1|w_2,w_3 ) + i(w_1;w_2|w_3 ) , \\ ( [ \ell_{1 } + \ell_{13 } ] \log f ) /m & > h(w_1 , v_{13 } |v_{12 } , v_{123 } ) = h(w_1|w_2,w_3 ) + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{1 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{123 } | v_{12 } , v_{13 } ) = h(w_1|w_2,w_3 ) , \\ ( [ \ell_{12 } + \ell_{13 } ] \log f ) /m & > h(v_{12 } , v_{13 } | w_1 , v_{123 } ) = 0 , \\ & \vdots\\ ( [ \ell_{1 } + \ell_{12 } + \ell_{13 } ] \log f ) /m & > h(w_1 , v_{12 } , v_{13 } | v_{123 } ) \nonumber \\ & = h(w_1|w_2,w_3 ) + i(w_1;w_2|w_3 ) + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{1 } + \ell_{12 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{12 } , v_{123 } | v_{13 } ) = h(w_1|w_2,w_3 ) + i(w_1;w_2|w_3 ) , \\ ( [ \ell_{1 } + \ell_{13 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{13 } , v_{123 } | v_{12 } ) = h(w_1|w_2,w_3 ) + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{12 } + \ell_{13 } + \ell_{123}\ \log
f ) /m & > h(v_{12 } , v_{13 } , v_{123}|w_1 ) = 0 , \\ ( [ \ell_{1 } + \ell_{12 } + \ell_{13 } + \ell_{123 } ] \log f ) /m &
= h(w_1 , v_{12 } , v_{13 } , v_{123 } ) = h(w_1).\end{aligned}\ ] ] @xmath258 \log f ) /m & > h(w_1 , v_{12 } | v_{13 } , v_{123})\\ & = h(w_1|w_2,w_3)\\ & \quad + i(w_1;w_2|w_3 ) , \\ ( [ \ell_{1 } + \ell_{13 } ] \log f ) /m & > h(w_1 , v_{13 } |v_{12 } , v_{123})\\ & = h(w_1|w_2,w_3)\\ & \quad + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{1 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{123 } | v_{12 } , v_{13})\\ & = h(w_1|w_2,w_3 ) , \\ ( [ \ell_{12 } + \ell_{13 } ] \log f ) /m & > h(v_{12 } , v_{13 } | w_1 , v_{123 } ) = 0 , \\ & \vdots\\ ( [ \ell_{1 } + \ell_{12 } + \ell_{13 } ] \log f ) /m & > h(w_1 , v_{12 } , v_{13 } | v_{123 } ) \nonumber \\ & = h(w_1|w_2,w_3)\\ & \quad + i(w_1;w_2|w_3)\\ & \quad + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{1 } + \ell_{12 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{12 } , v_{123 } | v_{13})\\ & = h(w_1|w_2,w_3)\\ & \quad + i(w_1;w_2|w_3 ) , \\ ( [ \ell_{1 } + \ell_{13 } + \ell_{123 } ] \log f ) /m & > h(w_1 , v_{13 } , v_{123 } | v_{12})\\ & = h(w_1|w_2,w_3)\\ & \quad + i(w_1;w_3|w_2 ) , \\ ( [ \ell_{12 } + \ell_{13 } + \ell_{123}\ \log f ) /m & > h(v_{12 } , v_{13 } , v_{123}|w_1 ) = 0,\end{aligned}\ ] ] @xmath259 \log f ) /m & = h(w_1 , v_{12 } , v_{13 } , v_{123})\\ & = h(w_1).\end{aligned}\ ] ] here , we need to consider all possible non - empty subsets of @xmath260 on the lhs .
we have omitted some trivial inequalities where the rhs equals zero . repeating the above for sources 2 and 3 , and simplifying , we have .
d. gndz , e. tuncel , and j. nayak , `` rate regions for the separated two - way relay channel , '' in _ proc .
46th allerton conf .
control comput .
( allerton conf . ) _ , monticello , usa , sept . 2326 2008 , pp . 13331340 . c. schnurr , s. stanczak , and t. j. oechtering , `` achievable rates for the restricted half - duplex two - way relay channel under a partial - decode - and - forward protocol , '' in _ proc .
theory workshop ( itw ) _ , porto , portugal , may 59 2008 , pp . 134138 .
x. liang , s. jin , x. gao , and k .- k .
wong , `` outage performance for decode - and - forward two - way relay network with multiple interferers and noisy relay , '' _ ieee trans .
_ , vol .
61 , no . 2 ,
521531 , feb .
s. j. kim , b. smida , and n. devroye , `` capacity bounds on multi - pair two - way communication with a base - station aided by a relay , '' in _ proc .
inf . theory ( isit ) _ , austin , usa , june 1318 2010 , pp . 425429 .
r. timo , g. lechner , l. ong , and s. j. johnson , `` multi - way relay networks : orthogonal uplink , source - channel separation and code design , '' _ ieee trans .
_ , vol . 61 , no . 2 , pp .
753768 , feb .
j. l. massey , `` channel models for random - access systems . '' in _ performance limits in communication theory and practice , nato advances studies institutes series e142 _ , j. k. skwirzynski , ed.1em plus 0.5em minus 0.4emkluwer academic , 1988 , pp . 391402 .
a. jain , d. gndz , s. r. kulkarni , h. v. poor , and s. verd , `` energy - distortion tradeoffs in gaussian joint source - channel coding problems , '' _ ieee trans .
inf . theory _
58 , no . 5 , pp .
31533168 , may 2012 . | this paper studies the three - user finite - field multi - way relay channel , where the users exchange messages via a relay . the messages are arbitrarily correlated , and the finite - field channel is linear and is subject to additive noise of arbitrary distribution .
the problem is to determine the minimum achievable source - channel rate , defined as channel uses per source symbol needed for reliable communication .
we combine slepian - wolf source coding and functional - decode - forward channel coding to obtain the solution for two classes of source and channel combinations . furthermore ,
for correlated sources that have their common information equal their mutual information , we propose a new coding scheme to achieve the minimum source - channel rate .
bidirectional relaying , common information , correlated sources , linear block codes , finite - field channel , functional - decode - forward , multi - way relay channel |
in the superfluid state , the bogoliubov wave function @xmath96 obeys the bogoliubov de gennes equation : @xmath97 , where the bogoliubov de gennes hamiltonian is given by @xmath98 for the superfluid @xmath0he b - phase ( @xmath0he - b ) , the normal state hamiltonian is @xmath99 and the pair potential is @xmath100 , where @xmath101 is the fermi wave number , @xmath102 is the effective mass of @xmath0he , @xmath25 is the superfluid gap , @xmath103 is the pauli matrix , and @xmath41 is the unit matrix in the spin - space . in the bulk , the excitation energy of the bogoliubov quasiparticles is given by @xmath104 .
when @xmath0he - b fills @xmath105 and there is a specular surface at @xmath106 , low energy quasiparticles emerge below the superfluid gap @xcite . within the andreev approximation , @xmath107 with the fermi velocity @xmath108 , excitation energy of the low energy quasiparticles is given by @xmath24 , where @xmath109 .
the eigenfunction @xmath110 for the positive ( negative ) eigenvalue is @xmath111 where @xmath112 is the coherence length and @xmath113 the eigenfunction for @xmath114 implies the wave function of the quasiparticles reflected from the surface and that for @xmath115 implies the wave function of the quasiparticles injected into the surface .
the wave function of the surface bound state is produced by the wave functions of the reflected and injected quasiparticles in order to satisfy the boundary condition @xmath116 at @xmath106 as @xmath117 where @xmath72 is the parallel component to the surface and @xmath118 . ignoring the gapped modes ,
the fermion field operator @xmath119 is described by using the wave function as @xcite @xmath120,\end{aligned}\ ] ] which gives the majorana relation @xmath121 .
therefore , we can regard the quasiparticles in the surface bound state as the majorana fermions .
when we consider the scattering process on the majorana fermions by an impurity with a small potential radius @xmath122 , such as the electron bubble @xcite , at @xmath123 , the spatial modulation with the scale @xmath124 of the wave function in eq . can be neglected as @xmath125 therefore , the normalized spinor for the majorana fermions in each spin state are described by @xmath126 indicates the plane wave .
the quasiclassical green s function for the surface bound state @xmath127 in @xmath0he - b is obtained by wu and sauls @xcite as @xmath128 where @xmath129 .
each component is given by @xmath130,\\ & g_{\parallel } ( \theta , e , z)=\delta^2\sin\theta|\cos\theta|p(\theta , e)e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z|},\\ & f_{\parallel } ( \theta , e , z)=i\frac{\delta\sin\theta } { \sqrt{\delta^2-e^2}}\left[1-\delta^2\cos^2\theta p(\theta , e)e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z|}\right],\\ & f_z(\theta , e , z)=i\frac{\delta\cos\theta } { \sqrt{\delta^2-e^2}}\left[1-e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z|}\right]+e\delta|\cos\theta|p(\theta , e)e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z|},\\ & \underline{f}_z(\theta , e , z)=i\frac{\delta\cos\theta } { \sqrt{\delta^2-e^2}}\left[1-e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z|}\right]-e\delta|\cos\theta|p(\theta , e)e^{-2\sqrt{\delta^2-e^2}z/\hbar|v_z| } , \end{split}\ ] ] where @xmath131 and @xmath132^{-1}$ ] with @xmath133 .
41ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop | we have formulated the scattering theory on majorana fermions emerging in the surface bound state of the superfluid @xmath0he b - phase ( @xmath0he - b ) by an impurity . by applying the theory to the electron bubble , which is regarded as the impurity , trapped below a free surface of @xmath0he - b , the observed mobility of the electron bubble [ ikegami _ et al .
_ , j. phys .
. jpn . * 82 * , 124607 ( 2013 ) ] is quantitatively reproduced .
the mobility is suppressed in low temperatures from the expected value in the bulk @xmath0he - b by the contribution from the surface majorana fermions . by contrast , the mobility does not depend on trapped depth of the electron bubble in spite of the spatial variation of the wave function of the surface majorana fermions .
our formulated theory demonstrates the depth independent mobility by considering intermediate states in the scattering process .
therefore , we conclude that the experiment has succeeded in observing majorana fermions in the surface bound state . _
introduction. _ detection of majorana fermions in topological superfluids / superconductors @xcite is a modern objective in condensed matter physics . among several systems hosting majorana fermions @xcite ,
the superfluid @xmath0he b - phase ( @xmath0he - b ) is an ideal one because the bulk properties have been well established @xcite .
the established theory allows us quantitative comparison with experimental results . in @xmath0he - b ,
majorana fermions emerge in the surface bound state with linear dispersion below the superfluid gap @xcite .
there have been a few reports on the successful detection of the majorana fermions in @xmath0he - b @xcite . in an experiment ,
the impedance by the surface acoustic wave of @xmath0he - b was measured in a sample cell @xcite .
the observed impedance _
qualitatively _ corresponds with the theoretically evaluated impedance by considering the surface bound state .
another experiment demonstrates the enhancement of heat capacity of @xmath0he - b in low temperatures @xcite .
this enhancement is interpreted as caused by the contribution from the majorana fermions in the surface bound state ; however , the heat capacity is larger than an estimated value from the geometrical surface area of the sample owing to the surface roughness .
thus , it is hard to _ quantitatively _ compare theoretical predictions and experimental results for @xmath0he - b in a vessel owing to the surface roughness with the atomic scale . at the rough surface ,
the majorana fermions have a broad spectrum deviated from linear dispersion @xcite .
the specular reflection of @xmath0he quasiparticles ( qps ) is realized on a free surface .
the experiment to observe the majorana fermions in the surface bound state at the free surface of @xmath0he - b was performed by ikegami _
et al . _
@xcite . in the experiment ,
mobility of electron bubbles trapped below the free surface is measured .
note that the electron bubble is a self - trapped electron by the pauli exclusion with electrons surrounding @xmath0he atoms @xcite .
since the drag force on the slowly moving electron bubbles is caused by momentum transfer from the @xmath0he qps , the majorana fermions in the surface bound state are expected to be observed via the mobility . the formulation on the mobility of the electron bubbles in the bulk @xmath0he is well established @xcite , which reproduces experimental results @xcite .
furthermore , the transverse force on the moving electron bubbles in the superfluid @xmath0he a - phase , which is the direct evidence of chiral symmetry breaking @xcite , is quantitatively demonstrated by considering momentum transfer from weyl fermions in the bound state around the electron bubble @xcite . in the experiment by ikegami _
et al . _
@xcite , the observed mobility is suppressed in low temperatures from the theoretically evaluated value for @xmath0he - b in the bulk @xcite .
the suppression of the mobility can be interpreted as caused by the scattering with low energy majorana fermions whose local density of states ( dos ) , @xmath1 , depends on the depth from the free surface , @xmath2 @xcite , where @xmath3 nm is the coherence length . by contrast , the observed mobility is independent of the trapped depth of the electron bubbles in the whole measuring range @xmath4 .
the depth independent mobility seems to contradict the depth dependence of the local dos .
therefore , it has been unclear on the origin of the suppressed mobility and whether the experiment succeeded in observing the majorana fermions in the surface bound state . in this letter
, we formulate the scattering between an impurity and the majorana fermion in the surface bound state so as to understand the mobility observed by the above experiment @xcite . the mobility obtained by our formulation
quantitatively reproduces the observed depth independent mobility as a result of that the depth dependence of the scattering cross section and the local dos of the surface bound state cancel out . _
formulation. _ the mobility of impurities , such as the electron bubbles , is determined by the momentum transfer from the @xmath0he qps to the impurity .
the equation of motion for the momentum of the impurity is given by @xcite @xmath5 where @xmath6^{-1}$ ] is the fermi distribution at temperature @xmath7 for the qp with excitation energy @xmath8 and @xmath9 is the transition rate for the qps from wave number @xmath10 and spin @xmath11 to @xmath12 and @xmath13 by the scattering on the moving impurity with velocity @xmath14 . by the time reversal symmetry in @xmath0he - b ,
the equation of motion becomes up to the first order of @xmath14 @xcite , @xmath15({\bm k}'-{\bm k})(1-f_{\bm k})f_{\bm k}\\
\times\gamma({\bm k},\sigma\to{\bm k}',\sigma'),\end{gathered}\ ] ] where @xmath16 is the transition rate for the qps by the scattering on the static impurity : @xmath17 here , we only consider the elastic scattering because the recoil energy of the impurity is sufficiently low according to the experiments @xcite .
since the drag force on the moving impurity with @xmath18 parallel to the free surface is given by @xmath19 , the stokes drag coefficient is obtained by @xmath20 the squared @xmath7-matrix element is given by @xmath21 for the surface bound state in @xmath0he - b , the spinors for the majorana fermions are described by @xmath22 and @xmath23 for the qp energy @xmath24
@xcite , where @xmath25 is the bulk gap and @xmath26 and @xmath27 are the azimuthal angle and the polar angle from the normal axis to the free surface , respectively .
the wave number of the majorana fermions is fixed on the fermi wave number , @xmath28 .
the @xmath7-matrix element @xmath29 is given by the following equation based on the lippman - schwinger equation @xcite : @xmath30t_{\rm s}(\hat{\bm k}'',\hat{\bm k},e , z ) , \label{eq : ls}\end{gathered}\ ] ] where @xmath31 is the dos in the normal state per spin . since the size of the impurity is much less than the coherence length , the @xmath7-matrix depends on the quasiclassical green s function @xmath32 for the surface bound state only at the impurity position @xmath33 @xcite . the quasiclassical green s function in the normal state
is given by @xmath34 and the expression of @xmath35 is obtained by ref .
( see also supplemental material @xcite ) .
the @xmath7-matrix in the normal state is given by @xmath36 $ ] with @xmath37 where @xmath38 is the legendre polynomial and @xmath39 is the phase shift depending on the potential of the impurity . here , @xmath40
( @xmath41 ) is the unit matrix in the nambu- ( spin-)space . by solving eq . , we derive the squared @xmath7-matrix element in eq . .
the summation of wave number for initial and final states in eq . can be replaced by the integral : @xmath42 where the angle - resolved local dos of the surface bound state is given by @xcite @xmath43 .
\label{eq : dos}\end{gathered}\ ] ] thus , the depth dependence of the impurity mobility originates from the dos of the surface bound state and the @xmath7-matrix element through the quasiclassical green s function . by the replacement of the summation by the integral , the drag coefficient in eq .
is reduced to @xmath44 where @xmath45 is @xmath0he density , @xmath46 is the fermi momentum , and @xmath47 .
the polar angle averaged differential cross section @xmath48 is given by @xmath49 where the amplitude of the perpendicular wave number @xmath50 is fixed by setting the qp energy and the sign of that only remains as degree of freedom in the @xmath7-matrix element .
the fourth power of the qp energy , @xmath51 , originates from the linear dispersion of the surface bound state , @xmath52 , and the jacobian @xmath53 being replaced by @xmath54 owing to the delta function in the dos of the surface bound state in eq . . here
, we define the total cross section and transport cross section as @xmath55 respectively . then , the drag force on the impurity moving parallel to the free surface from the surface majorana fermions is obtained by the energy integral of the transport cross section .
so far , we generally formulate the scattering theory on the majorana fermions in the surface bound state by an impurity with the following assumptions .
( i ) the velocity of the moving impurity is sufficiently low , i.e. , the drag force is independent of the impurity velocity .
( ii ) the recoil energy of the impurity is sufficiently low , i.e. , the collision process is the elastic scattering .
( iii ) the size of the impurity is much less than the coherence length , i.e. , the @xmath7-matrix depends on the quasiclassical green s function only at the impurity position . from now on ,
we consider the electron bubble as the impurity whose potential is well modeled by the hard sphere potential .
this model potential quantitatively reproduces not only the observed mobility in the a- and b - phases @xcite but also the transverse force on the electron bubble in the a - phase caused by chiral symmetry breaking @xcite . for the hard sphere potential with radius @xmath56 ,
the phase shift @xmath39 is given by @xmath57 , where @xmath58 and @xmath59 are the spherical bessel and neumann functions , respectively .
the only parameter , the radius @xmath56 of the hard sphere potential , is fixed by the observed mobility in the normal state @xmath60 @xcite at @xmath61 @xcite . _ scattering cross section. _ first , we discuss the scattering cross section between the electron bubble and the majorana fermion in the surface bound state . in fig .
[ fig1](a ) , we show the total cross section @xmath62 and the transport cross section @xmath63 at @xmath64 .
the cross sections have the peak structure which mainly originates from the bound state around the electron bubble . the peak structure by the bound state is clearly seen deep in @xmath0he - b from the free surface , namely in the bulk , as the total cross section at @xmath65 shown in fig .
[ fig1](b ) .
the number of peaks is of the order of @xmath66 .
note that the bound state around the electron bubble is never detected in the bulk because low energy qps below the superfluid gap are absent .
the influence of the bound state emerges only after the presence of the low energy majorana fermions in the surface bound state .
the transport cross section is strongly suppressed from the total cross section . in order to clarify the reason of the suppression , the polar angle averaged differential cross section scaled by @xmath67 at @xmath64 for @xmath68 is depicted in fig .
[ fig1](d ) by a solid line .
the polar angle averaged differential cross section should be compared with that in the normal state depicted by a dashed line with the scale @xmath69 , where the total cross section and transport cross section in the normal state are @xmath70 and @xmath71 , respectively . by the comparison between them ,
it is clear that the strong suppression of the transport cross section is due to the reduction of the backscattering .
in particular , the perfect backscattering of the majorana fermion from momentum @xmath72 to @xmath73 is forbidden .
it is similar to the edge state in topological insulators which is not scattered to backward by nonmagnetic impurities owing to the time - reversal symmetry @xcite .
note that the electron bubble is regarded as a nonmagnetic impurity for the @xmath0he qps with small nuclear magnetic moment .
( color online ) ( a ) total cross section and transport cross section at @xmath64 .
( b ) total cross section at @xmath65 . ( c )
transport cross sections at @xmath74 , @xmath75 , and @xmath76 .
( d ) polar angle averaged differential cross section scaled by @xmath67 at @xmath64 for @xmath68 ( solid line ) and that in the normal state scaled by @xmath69 ( dashed line ) . ,
width=321 ] the depth dependence of the transport cross section is shown in fig .
[ fig1](c ) .
the demonstrated positions of the electron bubble , @xmath74 , @xmath75 , and @xmath76 , are included within the experimentally accessible range @xmath4 @xcite . as the electron bubble approaches to the free surface , the transport cross section becomes smaller .
this depth dependence is due to the competition of the surface bound state and the bound state around the electron bubble .
the surface bound state is embedded in the quasiclassical green s function @xmath32 as terms proportional to @xmath77 $ ] @xcite . for @xmath78 , since @xmath79 approaches zero , the cross section has the bulk spectrum originating from the bound state around the electron bubble , as shown in fig .
[ fig1](b ) . _
mobility. _ the qp energy integral of the transport cross section in eq . provides the drag coefficient @xmath80 on the electron bubble by the majorana fermions in the surface bound state . on mobility
, however , the drag force also acts on the electron bubble by the qps in the continuum state with the energy above the superfluid gap .
when we assume that the transport cross section for the qps in the continuum state corresponds to that in the bulk @xmath0he - b , @xmath81 , the drag coefficient is given by @xcite @xmath82 from the drag coefficients , we can obtain the mobility as @xmath83 , where @xmath84 is the electron charge .
the temperature dependence of the mobility comes from the amplitude of the gap @xmath85 and the derivative of the fermi distribution @xmath86 . in low temperatures , therefore , the mobility is dominated by the majorana fermions in the surface bound state because the low energy qps have a major contribution to the drag force .
then , the assumption that the transport cross section in the continuum state is replaced by that in the bulk is relevant in low temperatures . in fig .
[ fig2 ] , we compare the temperature dependence of the calculated mobility ( solid lines ) and the experimental result ( circles ) in ref . .
the calculation is performed for positions of the electron bubble at @xmath74 , @xmath75 , and @xmath76 which correspond to @xmath87 nm , 40 nm , and 60 nm , respectively .
the obtained mobility is strongly suppressed in low temperatures from the bulk value of the mobility ( dashed line ) . up to @xmath88 ( above @xmath89 ) , the depth independent mobility quantitatively agrees with the experimental result . on the other hand , above @xmath88 ( below @xmath89 ) ,
the value of the mobility at @xmath64 is slightly larger than that at @xmath74 and @xmath75 .
it is expected that the difference of the mobility of the electron bubbles at @xmath90 nm and in @xmath91 nm will be observed .
the depth independent mobility above @xmath89 results from the cancellation of the spatial variation of the squared local dos of the surface bound state proportional to @xmath92 by the depth dependence of the transport cross section @xmath63 in fig .
[ fig1](c ) . for the electron bubble
far away from the free surface , since the transport cross section has the depth independent spectrum in fig .
[ fig1](b ) , the drag force @xmath80 from the majorana fermions decreases proportionally to the squared dos and the mobility approaches the value in the bulk @xmath0he - b as shown in fig .
[ fig2 ] for @xmath93 .
( color online ) temperature dependence of calculated mobility ( solid lines ) compared with the experimental result ( circles ) in ref . .
the solid lines indicate the mobility at @xmath74 , @xmath94 , @xmath64 , and @xmath93 from the small value in low temperatures .
the dashed line indicates the mobility in the bulk @xmath0he - b .
inset : the mobility in @xmath95 with a logarithmic scale .
, width=321 ] _ summary. _ we have formulated the scattering theory on majorana fermions in the surface bound state by an impurity in @xmath0he - b . by applying the theory to the electron bubble trapped below a free surface of @xmath0he - b , we have quantitatively reproduced the observed depth independent mobility in ref . .
the depth independence is due to the cancellation of the dos of the surface bound state and the suppression of the scattering cross section by the competition between the surface bound state and the bound state around the electron bubble .
therefore , we conclude that the experiment has succeeded in observing majorana fermions in the surface bound state .
the detected majorana fermions should have ideal linear dispersion owing to the perfect agreement with the theoretical expectation . according to the present result , the depth dependence of the mobility will be observed below @xmath89 .
i thank h. ikegami , k. kono , j. a. sauls , y. kato , and y. masaki for helpful discussions .
i also acknowledge financial support from the japan society for the promotion of science ( jsps ) .
this work was supported by jsps kakenhi grants no .
15k17715 and no . 15j05698 .
41ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop @noop * * , ( ) @noop @noop |
the density matrix is a very useful tool in the numerical study of interacting systems . besides being used in the density - matrix renormalization group ( dmrg)@xcite and its higher - dimensional generalizations,@xcite
the density matrix is also used as a diagnostic tool in the contractor renormalization ( core ) method for numerical renormalization group in two dimensions,@xcite and forms the basis of a method to identify the order parameter related to a quasi - degeneracy of ground states.@xcite in previous work,@xcite we extended the results of chung and peschel@xcite to write the density matrix ( dm ) of a cluster of @xmath3 sites cut out from a system of noninteracting spinless fermions in @xmath4 dimensions as the exponential of a quadratic operator , called the pseudo - hamiltonian , as it resembles the hamiltonian of a noninteracting system .
that result was then applied in numerical studies of noninteracting spinless fermions in one dimension , to better understand how the distribution of cluster dm eigenvalues scale with @xmath3 , and to explore the possibility of designing truncation schemes based on the pseudo - hamiltonian.@xcite we believe truncation schemes such as that described in ref .
will be helpful to the choice of basis states in renormalization groups such as core .
thus , some questions motivating the present paper were : ( i ) does the density matrix of an interacting fermi - liquid system resembles that of a noninteracting one ?
( ii ) can we apply our exact result in ref . to two dimensions as well as for one dimension ?
( iii ) is it numerically practical to compute this sort of density matrix in a fermion system . to answer these questions
, we investigated a spinless analog of the extended hubbard model , given by the hamiltonian @xmath5 in the limit of @xmath6 , so that fermions are not allowed to be nearest neighbors of each other .
this model is chosen for two reasons : ( i ) for a given number of particles , the @xmath6 hilbert space is significantly smaller than the @xmath7 hilbert space , and we can work numerically with larger system sizes ; and ( ii ) the model , in spite of its simplicity , has a rich zero - temperature phase diagram,@xcite where we find practically free fermions in the limit @xmath8 , and an inert solid at half - filling @xmath9 .
as the filling fraction approaches quarter - filling from below , @xmath10 , the system becomes congested , highly correlated , but is nonetheless a fermi liquid , perhaps with additional orders that are not clear in small systems . slightly above quarter - filling , the dense fluid and inert solid coexists , while slightly below half - filling , the system is expected to support stable arrays of stripes . to probe this rich variety of structures in the ground state at different filling fraction @xmath2 ,
we describe in section [ sect : formulation ] how the reduced dm of a small cluster , with the appropriate symmetry properties , can be calculated from a finite non - square system subject to twist boundary conditions .
then in section [ sect : ni ] , we investigate in great details the cluster dm spectra of the noninteracting system , particularly on how to handle finite size effects in the numerics , for comparison with the cluster dm spectra of a strongly - interacting system , presented in section [ sect : strong ] .
finally , in section [ sect : summary ] , we summarized our findings , and discuss the prospects of designing an operator - based dm truncation scheme for interacting systems , at some , if not at all , filling fractions .
in this section , we give the theoretical formulations and describe the numerical tools needed to investigate the cluster dm spectra of noninteracting and strongly - interacting systems of spinless fermions in two dimensions . in section [
sect : cdmf ] , we give the matrix elements of the dm of a small cluster embedded within a larger , but still finite , system .
these matrix elements are obtained by tracing out degrees of freedom external to the cluster , starting from the ground - state wave function of the system , obtained through exact diagonalization . in section [
sect : sysdef ] , we describe how our finite systems can be defined with nonsquare periodic boundary conditions , and how we make use of the translational invariance of both noninteracting and strongly - interacting models to reduce the computational efforts in exact diagonalization . in section [ sect : averaging ] , we describe several averaging apparatus required to obtain a handle on the infinite - system spectra of the cluster dm , and then in section [ sect : states ] , we describe a classification scheme for the one - particle and multi - particle eigenstates of the cluster dm that makes the symmetry of the underlying square lattice explicit .
the dm @xmath11 of a cluster cut out from a larger system is a density operator which gives the expectation @xmath12 for any observable @xmath13 local to the cluster , when the larger system is in its ground state @xmath14 .
the cluster dm @xmath11 can be calculated from the ground - state dm @xmath15 of the system , by tracing out degrees of freedom outside of the cluster .
we write this as @xmath16 where the subscript @xmath17 denotes a trace over environmental degrees of freedom .
since a cluster is a collection of sites identified in real space , it is natural to choose as a many - body basis the real - space configurations . for a finite two - dimensional system with @xmath18 sites , we label the sites @xmath19 through @xmath20 , so that for any pair of sites @xmath21 and @xmath22 , we have @xmath23 and @xmath24 if @xmath25 .
we then distinguish between sites within the cluster , of which there are @xmath3 of them , @xmath26 , @xmath27 , , @xmath28 , and sites outside of the cluster , of which there are @xmath29 of them , @xmath30 , @xmath31 , , @xmath32 .
we think of the @xmath33 sites outside the cluster as constituting the _ environment _ to the cluster .
we work with the configuration basis states @xmath34 , where @xmath35 are the @xmath36 occupied sites in the system .
these can be thought of as a direct product of the configuration basis states of the cluster @xmath37 , where @xmath38 are the @xmath39 occupied sites within the cluster , and the configuration basis states of the environment @xmath40 , where @xmath41 are the @xmath42 occupied sites in the environment . here , we have the occupied sites of the system @xmath43 being the union of the occupied sites in the cluster and in the environment , with the site indices @xmath44 and @xmath45 resorted in ascending order to give the site indices @xmath46 . in terms of the configuration basis of the system , the ground - state wave function of the system can be written as @xmath47 where @xmath48 is the amplitude associated with configuration @xmath49 , and @xmath50 , @xmath51 are fermion annihilation and creation operators acting on the site @xmath52 .
we can also write the ground - state wave function as @xmath53 in terms of the direct product of configuration bases of the cluster and the environment , where @xmath54 and @xmath55 are fermion annihilation and creation operators acting on site @xmath56 within the cluster , and @xmath57 and @xmath58 are fermion annihilation and creation operators acting on site @xmath59 within the environment . in , the amplitude @xmath60 is taken directly from the expansion in , while the factor @xmath61 accounts for the fermion sign incurred when we reorder the operator product @xmath62 to get the operator product @xmath63 .
similarly , the ground - state dm in can be written as @xmath64 using the system - wide configuration basis , or as @xmath65 using the direct - product basis between cluster configurations and environment configurations . performing the trace over the environment
as prescribed in , we find the fermion cluster dm @xmath11 to be @xmath66 its matrix elements are @xmath67 these matrix elements can be computed naively by running over all possible pairs of cluster states @xmath68 and @xmath69 , and performing the sums over @xmath70 and @xmath71 as they appear in , looking up the fermion - sign factors @xmath61 and amplitudes @xmath72 as and when they are needed .
we call this the _ naive algorithm_. alternatively , we can also reorganize the fermion - sign - factor - adjusted amplitudes @xmath73 into a matrix @xmath74 , whose rows are associated with the cluster configurations @xmath68 , and whose columns are associated with the environment configurations @xmath75 , the matrix @xmath11 can be computed directly by matrix multiplication as @xmath76 representing a collection of inner products .
we call this the _ pre - sorted inner - product algorithm_. we compare and analyze the computational complexity of these two algorithms in appendix [ appendix : computational ] . in the numerical studies presented in section [
sect : strong ] , we use the pre - sorted inner - product algorithm exclusively . from
, we see that the cluster dm @xmath11 is manifestly hermitian , and thus all its eigenvalues @xmath77 are real ( and in fact nonnegative ) . when obtained from the wave function of a state with definite particle number , @xmath11 has no matrix elements between cluster states containing different number of particles , and thus the eigenstates @xmath78 of @xmath11 can be organized into _ sectors _ , corresponding to @xmath79 particles within the cluster . for the rest of this paper
, we would refer to the eigenvalues of @xmath11 generically as its _ weights _ , since these have a natural probabilistic interpretation .
a @xmath39-particle weight of @xmath11 is therefore an eigenvalue corresponding to an eigenstate containing @xmath39 particles within the cluster . for noninteracting spinless fermions on an infinite square lattice ,
it is possible to compute the cluster dm @xmath11 starting from the fermi sea ground state , through the evaluation and diagonalization of @xmath80 . for an interacting system , we need to compute @xmath11 starting from @xmath81 in , the latter we obtain through exact diagonalization on a finite system .
we define the a finite system relative to the infinite square lattice in terms of the lattice vectors @xmath82 and @xmath83 , as shown in fig.[fig : system ] , such that @xmath84 is the number of lattice sites within the system . and @xmath83 .
we shall denote such a system as @xmath85 . in this example
shown , the system is @xmath86 . ] if we impose periodic boundary condition such that @xmath87 , then in the exact diagonalization to obtain @xmath14 we can take advantage of translational invariance through the use of the bloch states @xmath88 as our computational basis@xcite . in this bloch state ,
the configurations @xmath89 are all related to the generating configuration @xmath90 by the lattice translations @xmath91 associated with displacement @xmath92 , while @xmath93 are wave vectors allowed by the boundary conditions .
any configuration within the collection of translationally - related configurations @xmath89 can serve as the generating configuration , but we pick the one with the least sum of indices of occupied sites .
working with finite non - square systems introduces several complications .
first of all , we sometimes end up with degenerate ground states which suffer from symmetry - breaking not found in the true infinite - system ground state . however , because the point symmetry group of our non - square finite system is only a subgroup of the square lattice point symmetry group , the finite - system ground - state manifold is not invariant under all square lattice symmetry operations .
thirdly , when working with finite systems , we introduce systematic deviations which are collectively known as _ finite size effects_. we identify the three primary sources of finite size effects as ( i ) finite domain effect , which has to do with the fact that the small set of discrete wave vectors allowed are not adequately representative of the continuous set of wave vectors on the infinite square lattice ; ( ii ) shell effect , which has to do with the fact that the set of discrete wave vectors allowed are organized by symmetry into shells in reciprocal space , each of which can be partially or fully filled in the many - body ground state ; and ( iii ) shape effect , which has to do with the detailed shape of the non - square system we introduced . to eliminate these numerical artefacts , we adopted three averaging devices .
first , we average over the @xmath94-fold degenerate ground - state manifold .
our first motivation for doing so is as follows : if @xmath95 is the point symmetry group of the square lattice , and its subgroup @xmath96 is the point symmetry group of the @xmath85 system , then we will find that the cluster density matrices @xmath97 one for each wave function @xmath98 within the @xmath94-fold degenerate ground - state manifold , are not invariant under @xmath96 , much less @xmath95 .
we remove this artificial symmetry breaking by calculating the degeneracy - averaged cluster dm @xmath99 over the cluster density matrices @xmath100 within the ground - state manifold .
this degeneracy - averaged cluster dm is invariant under @xmath96 .
a second motivation for such a mode of averaging over the degenerate ground - state manifold of the finite system is that thermodynamically , given the pure state density matrices @xmath101 , with energy eigenvalue @xmath102 , we typically construct the canonical ensemble dm as @xmath103 where @xmath104 is the canonical partition function .
states within a degenerate manifold have the same energy , and therefore contribute equally to the thermodynamic dm @xmath105 . in the limit of @xmath106 ,
the usual thermodynamic argument is that pure states decouple from one another , and we treat their respective density matrices independently , except for those states which are degenerate .
because they appear with the same boltzmann weight whatever the inverse temperature is , we should still treat the uniform combination instead of the individual density matrices in the limit of @xmath106 , the second averaging device involves an average over the orientation of the finite non - square system relative to the underlying square lattice .
this averaging restores the @xmath95-symmetry to the averaged ground state . in principle , this requires us to compute @xmath11 for a group of four systems : @xmath107 , @xmath108 , @xmath109 and @xmath110 .
however , if the cluster whose dm we are calculating is invariant under the action of @xmath95 , this averaging can be achieved by computing @xmath111 where @xmath112 is a point group transformation of the square lattice , @xmath113 is the unitary transformation of the cluster hilbert space associated with @xmath114 , and @xmath115 is the order of @xmath95 .
after these two averagings , the cluster dm has the full symmetry ( including translations ) of the underlying square lattice , but finite size effects remain .
we eliminate these as much as we can@xcite with the third averaging device , twist boundary conditions averaging.@xcite the usual way to implement twist boundary conditions is to work in the _ boundary gauge _ , keeping the hamiltonian unchanged , and demanding that @xmath116 where @xmath82 and @xmath83 are the lattice vectors defining our finite system , @xmath117 is a site within the system , and @xmath118 is the twist vector parametrizing the twist boundary conditions . in choice of gauge , the hamiltonian is not manifestly invariant under translations
. however , we can continue to block - diagonalize it using the bloch basis states defined in , provided the set of allowed wave vectors @xmath93 are shifted relative to @xmath119 for the usual periodic boundary conditions by the twist vector @xmath120 , i.e. @xmath121 the other natural way to implement twist boundary conditions is in the _ bond gauge _
, where we make the substitution @xmath122 in the hamiltonian , but demand that @xmath123 where @xmath82 and @xmath83 are the lattice vectors defining our finite system .
now the hamiltonian is manifestly invariant under translations , and we can bloch - diagonalize it using the bloch basis states defined in , with the same set of allowed wave vectors @xmath124 as for the usual periodic boundary conditions .
exact diagonalization can be performed in any gauge , but we chose to do it in the bond gauge , because in this gauge , the bloch basis states @xmath125 defined in can be used as is to block diagonalize the hamiltonian at all twist vectors @xmath120 .
this gives us the ground - state wave function @xmath126 in the bond gauge .
in the boundary gauge , or any other gauges , appropriate gauge transformations must be applied to @xmath127 before we can use this bloch basis to block diagonalize the hamiltonian .
because of this , the computational cost for exact diagonalization incurred in the bond gauge is fractionally lower than in other gauges .
we can also calculate the correlation functions @xmath128 ( of which the cluster dm is a function of ) in any gauge , with appropriately - defined covariant observables @xmath129 , where @xmath130 are the ` physical ' observables we would use when there is no twist in the boundary conditions . in the boundary gauge , these covariant observables @xmath131 and @xmath132 are particularly simple , except when the displacement vector @xmath133 crosses the boundaries of our system . for our purpose of calculating the cluster
dm , this situation occurs only when the cluster itself straddle the system boundaries .
therefore , with the cluster properly nested within the system , we chose to perform twist boundary conditions averaging in the boundary gauge .
in the boundary gauge , the cluster dm is obtained by tracing down the ground - state wave function @xmath134 .
we can get this wave function from @xmath126 by applying the gauge transformation @xmath135 where @xmath136 is an occupation number basis state , with occupation @xmath137 on site @xmath92 .
now , averaging over twist vectors @xmath120 is the same as integrating over the brillouin zone , so we perform twist boundary conditions averaging over a uniform grid of monkhorst - pack points with order @xmath138.@xcite for rectangular systems @xmath139 , the first brillouin zone is sampled by varying the two independent twist angles between @xmath140 and @xmath141 .
for non - square systems , the two independent twist angles @xmath142 and @xmath143 are defined by @xmath144 the twist vector @xmath120 is then related to the independent twist angles @xmath145 and @xmath146 by @xmath147 where @xmath148 are the primitive reciprocal lattice vectors of our non - square system . for such systems ,
the first brillouin zone is a parallelogram on the @xmath149-@xmath150 plane , so the uniform grid of monkhorst - pack points are imposed on the square domain @xmath151 on the @xmath142-@xmath143 plane instead .
the point symmetry group of the square lattice is the dihedral group @xmath152 , which has eight elements.@xcite the five irreducible representations of this group are @xmath153 , @xmath154 , @xmath155 , @xmath156 and @xmath17 . for the cross - shaped cluster shown in fig.[fig : cross ]
, there is a one - to - one correspondence between these five irreducible representations and the one - particle states , but instead of labeling these one - particle states as @xmath157 , @xmath158 , @xmath159 , @xmath160 and @xmath161 , we adopt an angular momentum - like notation , @xmath162 that would make clear the structure of these one - particle states . we find it more convenient to work with the one - particle states @xmath163 for a cluster dm possessing the full point group symmetry of the square lattice , the one - particle state @xmath164 is constrained by symmetry to always be an eigenstate of @xmath11 .
we call the associated eigenvalue @xmath165 the _ weight _ of @xmath164 .
furthermore , the one - particle states @xmath166 and @xmath167 are also equivalent under the square lattice symmetry , and hence their weights @xmath168 and @xmath169 are equal .
we call this doubly - degenerate one - particle weight @xmath170 . on the other hand , the @xmath171 one - particle eigenstates of @xmath11 are in general not @xmath172 , @xmath173 or @xmath174 , @xmath175 , but some admixture of the form @xmath176 we call their corresponding weights @xmath177 and @xmath178 respectively
. we can then extend this angular - momentum - like notation to multi - particle states of the cluster .
though the quantum numbers used to label the one - particle states are , strictly speaking , not angular momentum quantum numbers , we apply the rules of angular momenta addition as if they were to write down the angular - momentum - like quantum numbers for the multi - particle states .
for example , for the two - particle states of the cluster , we have @xmath179
in preparation for our main calculations on the strongly - interacting system , we investigated in great details the cluster dm spectra of a system of noninteracting spinless fermions described by the hamiltonian @xmath180 an analytical formula for the cluster dm @xmath11 is known for this system,@xcite using which we can obtain the spectrum of @xmath11 for any system size .
we take advantage of this analytical formula to calculate @xmath11 for an infinite system of noninteracting spinless fermions .
however , our goal in calculating the cluster dm spectrum of noninteracting spinless fermions is to compare it against the cluster dm spectrum of interacting spinless fermions . for the latter system
, we can only calculate sans approximations the cluster dm from the exactly - diagonalized ground - state wave function of finite systems . to make this comparison between noninteracting and interacting spinless fermions more meaningful , we compute their cluster dms for the same series of finite systems . in section [
sect : nicdm ] , we describe how the infinite - system and finite - system cluster dms for noninteracting spinless fermions are calculated .
in section [ sect : nifse ] , we observe that the noninteracting finite - system cluster dm spectra are contaminated by various finite - size effects .
these finite - size effects also arise when we compute the interacting finite - system cluster dm spectra , so we want to learn how to deal with them .
clearly , the effectiveness of various techniques in reducing finite - size effects can be best gauged by applying them to finite systems of noninteracting spinless fermions , since we will then be able to compare the results from the various techniques against the infinite - system limit .
the simplest antidote to the various finite - size effects is to use a larger finite system .
as expected , finite - size effects do become less and less important as the size of the system is increased .
unfortunately , based on comparisons of the finite - system cluster dm spectra with the infinite - system cluster dm spectrum , we realized that we would need to go to system sizes of a few hundred sites in order for the finite - system cluster dm spectra to be decent approximations of the infinite - system cluster dm spectrum .
since such system sizes are not practical for exactly diagonalizing the strongly - interacting system given by , we look into the method of twist boundary conditions averaging in section [ sect : tbca ] .
this method involves averaging the cluster dm spectra over various phase twists introduced into the periodic boundary conditions imposed on a given finite system . for noninteracting spinless fermions , we find that twist boundary conditions averaging reduces finite domain and shell effects in the cluster dm spectra , which then approximate the infinite - system cluster dm very well . as a matter of standardization ,
we apply the method of twist boundary conditions averaging to both noninteracting and interacting spinless fermions , and compare their twist boundary conditions - averaged cluster dm spectra . instead of the general formalism presented in section [ sect : cdmf ] ,
for the system of noninteracting spinless fermions we calculate the cluster dm weights using the exact formula @xmath181_{ij } { { c_{i}^{\dagger}}}{{c_{j}^{\phantom{\dagger}}}}\biggr\}\ ] ] obtained in ref . , which relates the dm @xmath11 of a cluster of sites and the cluster green - function matrix @xmath80 .
the matrix elements of @xmath80 are given by @xmath182 where @xmath14 is the ground state of the system , and @xmath92 , @xmath183 are sites within the cluster .
the corollary of is that , if @xmath184 is an eigenvalue of the cluster green - function matrix @xmath80 , the corresponding one - particle weight of @xmath11 is @xmath185 to calculate the infinite - system spectra of @xmath11 , we convert the sum over @xmath186 in into an integral @xmath187 over those wave vectors @xmath93 bounded by the fermi surface @xmath188 , where @xmath189 is the fermi energy . on an infinite square lattice with unit lattice constant
, the dispersion relation is given by @xmath190 we then obtain the infinite - system cluster green - function matrix eigenvalues @xmath191 , @xmath192 , @xmath193 and @xmath194 as functions of the fermi energy @xmath189 by numerically integrating , and diagonalizing the resulting infinite - system cluster green - function matrix @xmath195 . for the same set of fermi energies , we also integrate @xmath196 over the fermi surfaces to find the corresponding filling fractions .
the one - particle infinite - system cluster dm weights @xmath197 , @xmath198 , @xmath199 , and @xmath200 are then calculated using . to calculate the cluster dm spectra for a finite system of @xmath18 sites with @xmath36 noninteracting particles
, we determine the set of wave vectors @xmath201 with the lowest single - particle energies @xmath202 we then evaluate the finite - system cluster green - function matrix elements in by summing over these occupied wave vectors @xmath201 . following this
, we diagonalize the finite - system cluster green - function matrix @xmath203 to obtain the eigenvalues @xmath204 , @xmath205 , @xmath206 , and @xmath207 , and therefrom the one - particle finite - system cluster dm weights @xmath208 , @xmath209 , @xmath210 , and @xmath211 using . by varying the number @xmath36 of noninteracting particles in the finite @xmath18-site system
, we determined the finite - system cluster dm one - particle spectra for the filling fractions @xmath212 accessible to the finite system . imposing the usual periodic boundary conditions , we calculated the finite - system spectra of @xmath11 for several small systems ranging from @xmath213 to @xmath214 sites . for these small systems sizes
, we find that it is impossible to say anything meaningful about the dependence on the filling fraction @xmath212 for the cluster dm weights because of the finite size effects .
using the relation for noninteracting systems , we investigated the effect of system size on the spectrum of @xmath11 for a series of system @xmath215 , @xmath216 , with the same shape .
as the system size is increased , we find that the cluster dm spectrum approaches the infinite - system limit , as shown in figure [ fig : free0n1 ] . for this series of systems ,
the infinite - system limit is more or less reached at around @xmath217 ( 272 sites ) , based on comparison with the infinite - system limit itself .
we can also arrive at this estimate by looking at the convergence of the one - particle weights alone .
more importantly , we find that the shell effect affects weights of different symmetry differently : @xmath177 is almost unaffected , while @xmath165 is the most severely affected .
shell effect persists in @xmath165 even up to a system size of 1088 sites ( for @xmath218 ) .
( ) and @xmath219 ( @xmath220 ) systems have the same shape but different sizes .
the solid curves are the zero- and one - particle weights for the infinite system . for @xmath177 , we see the finite domain effect deviations for the @xmath221 system is practically gone by the time we get to the @xmath219 system . for the rest of the one - particle weights ,
the finite domain effect is largely removed in the @xmath219 system , but shell effect persists .
in fact , shell effect in @xmath165 is still visible when we go to the @xmath222 system ( not shown ) , which has @xmath223 sites . ]
we also looked at @xmath224 , which is almost unaffected by shell effect , for several systems with between 200 to 300 sites of different shapes . for systems of these sizes ,
the finite domain effect is negligible , but we find @xmath224 from finite systems of different shapes differing very slightly from the infinite - system limit , and also from each other . since we expect systems of different shapes to all approach the same infinite - system limit , we attribute these very small deviations to the shape effect . based on more extensive numerical studies ( see chapter 4 of ref . )
not reported in this paper , we know that shape effect deviations are not effectively removed by the three averaging devices we have introduced in section [ sect : averaging ] , but fortunately these deviations are very small . for a system of interacting spinless fermions , we can not directly compute the exact infinite - system cluster dm .
we can , however , choose to work with ( i ) an approximate ground - state wave function of an infinite system , or ( ii ) the exact ground - state wave function of a finite system , or ( iii ) an approximate ground - state wave function of a finite system .
as reported in section [ sect : strong ] , we chose option ( ii ) , where the exact ground - state wave function is obtained through numerical exact diagonalization
. exact diagonalization severely limits the sizes of the finite systems we can work with ( see appendix of ref . for formula on size of hilbert space for the strongly - interacting model given by ) . with aggressive memory reduction measures ,
it is possible ( but not necessarily feasible ) to exactly diagonlize finite systems with up to 30 sites for all filling fractions .
however , as we have illustrated in section [ sect : nifse ] , for systems so small , the numerical cluster dm spectra would be plagued by strong finite size effects , most notably by the shell effect .
this is where twist boundary conditions averaging comes in . very crudely speaking
, we can think of averaging numerical observables over @xmath225 twist boundary conditions for a finite system with @xmath18 sites as being equivalent to computing these numerical observables for a single finite system of @xmath226 sites , subject to only periodic boundary conditions . in the best - case scenario
, the effective system size @xmath227 might be as large as @xmath228 , though @xmath227 will typically grow slower than @xmath229 .
the computational cost of performing exact diagonalization for a system with @xmath18 sites over @xmath225 twist boundary conditions is on the order of @xmath230 , whereas the computational cost of performing exact diagonalization just once for a system with @xmath226 sites is @xmath231 .
so long as the effective system size @xmath227 grows faster than @xmath232 , it would be computational cheaper to employ the method of twist boundary conditions averaging , instead of exactly diagonalizing a single large finite system , to reduce finite size effects .
the detail dependence of @xmath227 on @xmath225 will of course depend on the nature of the observable of interest . from the detailed study undertaken in appendix d of ref .
, we know that there are cuts and cusps on the twist surface @xmath233 of a generic observable @xmath130 , where @xmath234 is the many - body ground state of a finite @xmath18-site system subject to twist boundary conditions with twist vector @xmath235 . for non - square systems , these cusps and cuts demarcate features with a hierarchy of sizes on the twist surface .
the ` typical ' twist surface feature has a linear dimension of @xmath236 .
these are decorated by fine structures with linear dimension @xmath237 , which are in turn decorated by hyperfine structures with linear dimensions @xmath238 . the number of integration points we must use is therefore determined by what feature size we want to integrate faithfully .
monkhorst - pack special - point integration , in reducing finite size effects for the @xmath239 system ( ) is checked against the @xmath239 ( ) system with periodic boundary conditions imposed .
the solid curves are the cluster dm weights from the infinite system . ] for the purpose of this study , we decided to integrate the fine structure on the twist surface faithfully .
therefore , we chose to average the spectrum of @xmath11 over a @xmath240 monkhorst - pack grid ( which consists of 256 integration points in the first brillouin zone ) for the @xmath221 system with @xmath241 sites .
we find that twist boundary conditions averaging does indeed result in an averaged spectrum which approximates the infinite - system limit well ( see fig.[fig : free1tbcav ] ) .
this averaging device , however , does not completely eliminate shell effects , as can be seen from the twist boundary conditions - averaged @xmath242 . to reduce the bias this creates for one particular choice of finite system , we combined the twist boundary conditions - averaged spectrum of @xmath11 for various finite systems .
this is shown in fig.[fig : free1diffsys ] .
we will overlay the cluster dm spectra from several finite systems in the same way , to derive a twist boundary conditions - averaged approximation to the infinite - system cluster dm spectrum of strongly - interacting spinless fermions .
( ) , @xmath243 ( ) , @xmath244 ( @xmath245 ) , @xmath239 ( @xmath220 ) , @xmath246 ( @xmath247 ) , @xmath248 ( @xmath249 ) and @xmath250 ( @xmath251 ) systems of noninteracting spinless fermions subject to twist boundary conditions averaging , using @xmath240 monkhorst - pack special - point integration . also shown are the cluster dm weights of the infinite system . ]
in this section , we compute the cluster dm for interacting spinless fermions . as with the case of noninteracting spinless fermions ,
the cluster dm evaluated directly from the ed of various finite systems are severely affected by finite size effects . in section [
sect : ni ] , we saw how finite size effects can be significantly reduced when the cluster dm is averaged over various twist boundary conditions , in the sense that the averaged cluster dm weights from different finite systems at various filling fractions fall close to their respective infinite - system limits . applying twist boundary conditions averaging onto the interacting cluster dm
, we find that finite size effects are reduced , but not as dramatically as for the noninteracting cluster dm .
nevertheless , the averaged cluster dm weights from different finite systems with different number of particles are sufficiently consistent with each other that we can plot a smooth curve interpolating the averaged cluster dm weights .
as explained in section [ sect : intro ] , our interest in studying the strongly - interacting model of spinless fermions with infinite nearest - neighbor repulsion is to understand how the cluster dm evolves with filling , given that we expect in this model crossovers between regimes of qualitatively different states .
furthermore , we had proposed an operator - based method of truncation which was justified by the fact that the noninteracting cluster dm is generated from a set of single - particle operators . since we proposed the use of this truncation scheme for interacting systems , we are interested to know whether the structure of the interacting cluster dm is such that it can also be generated , perhaps approximately , from a set of single - particle operators . to this end , we present in sections [ sect : strongzero ] , [ sect : strongone ] and [ sect : strongtwo ] , the zero- , one- and two - particle cluster dm weights of the strongly - interacting system of spinless fermions .
we check whether it is possible to : ( i ) write the two - particle eigenstates as the product of one - particle eigenstates ; and ( ii ) predict the relative ordering of the two - particle weights based on the relative ordering of the one - particle weights .
we then discuss in section [ sect : fldm ] whether these two criteria are met , and whether it is feasible to design an operator - based dm truncation scheme , similar to the one described in ref .
, for the strongly - interacting system .
the zero - particle cluster dm weight calculated for various finite strongly - interacting systems is shown in fig.[fig : w0strong ] .
also shown in the figure is the zero - particle cluster dm weight of the infinite system of noninteracting spinless fermions . as we can see , the zero - particle weights of the respective systems only start differing significantly from each other for @xmath252 . with repulsive
interacting between spinless fermions , it is more difficult in a congested system ( @xmath253 ) to form an empty cluster of sites from quantum fluctuations . as a result , the strongly - interacting @xmath254 falls below the noninteracting @xmath254 .
however , this fact alone does not tell us anything more about the correlations in the strongly - interacting ground state , and so we move on to consider the one - particle cluster dm weights .
( ) , @xmath255 ( ) , @xmath244 ( @xmath245 ) , @xmath239 ( @xmath220 ) , @xmath246 ( @xmath247 ) and @xmath256 ( @xmath249 ) systems of strongly - interacting spinless fermions subject to twist boundary conditions averaging , using @xmath257 monkhorst - pack special - point integration . at @xmath258 and @xmath259 ,
we know analytically that @xmath260 and @xmath261 respectively , and the solid ` curve ' interpolates between these two known limits and the equally weighted data points at finite filling fractions @xmath262 . also shown as the dashed curve is the zero - particle weight of the infinite system of noninteracting spinless fermions . ] in our operator - based dm truncation scheme developed in ref . for a noninteracting system ,
the one - particle cluster dm weights play a very important role , since we select which one - particle operators to keep or discard based on the negative logarithm of these numbers .
we expect the one - particle cluster dm weights , though not entirely sufficient by themselves , would also play an important role in defining an operator - based truncation scheme .
therefore , in this section , we present results for a series of calculations to determine the infinite - system limit of the one - particle cluster dm spectra for our strongly - interacting system as a function of filling fraction @xmath2 , and discuss their implications for an operator - based dm truncation scheme .
though we really do need to worry about the evolution of the structure of @xmath263 and @xmath264 as a function of @xmath2 in both the noninteracting and strongly - interacting systems , the one - particle weights are ordered by their magnitudes as @xmath265 for both systems .
but while the noninteracting one - particle weights go down by roughly one order of magnitude as we go through the sequence @xmath266 , we see from fig.[fig : w1stronginterpolate ] that the interacting one - particle weights decay more slowly along this same sequence .
we studied the finite @xmath267 ( ) , @xmath243 ( ) , @xmath244 ( @xmath245 ) , @xmath239 ( @xmath220 ) , @xmath246 ( @xmath247 ) and @xmath248 ( @xmath249 ) systems subject to twist boundary conditions averaging , using @xmath257 monkhorst - pack special - point integration . at a filling fraction of @xmath258 ,
the system approaches the noninteracting limit , and thus all the one - particle weights are zero . at half - filling @xmath259 ,
the two - fold degenerate checker - board ground state is unaffected by twist boundary conditions averaging .
we can thus perform degeneracy averaging analytically , to find that @xmath268 and @xmath269 .
the solid ` curves ' in figure [ fig : w1stronginterpolate ] interpolate between these two known limits and the equally weighted data points at finite filling fractions @xmath270 . also shown in figure [ fig : w1stronginterpolate ] as the dashed curves are the one - particle weights of the infinite system of noninteracting spinless fermions .
when only periodic boundary conditions are imposed , there is significantly more ` scatter ' in the one - particle weights as a function of filling fraction @xmath2 , for interacting systems of different sizes , than for noninteracting systems of different sizes . averaging the one - particle weights of the interacting systems over various twist
boundary conditions visibly reduces this ` scatter ' , even though the remnant ` scatter ' seen in figure [ fig : w1stronginterpolate ] is still rather large , compared to case for the noninteracting system ( figure [ fig : free1diffsys ] ) . from our own detailed study of the method of twist boundary conditions averaging for noninteracting systems,@xcite we know that twist boundary conditions averaging effectively removes finite size effects from some observables , but not for others .
we have no reason to expect twist boundary conditions averaging to apply equally effectively over the same set of observables , when we go from noninteracting systems to interacting systems .
conversely , observables for which twist boundary conditions averaging is ineffective in noninteracting systems , might be effectively twist - boundary - conditions - averaged in interacting systems . with only input from the exact diagonalization of finite systems , and without employing system - size extrapolations , the method of twist boundary conditions averaging offers us the best hope of gaining insight into the infinite - system properties we seek .
we expect that the remnant ` scatter ' in the twist - boundary - conditions - averaged one - particle weights will be reduced , if we had not made the nearest - neighbor repulsion infinite .
there are two reasons why we did not also study the case of @xmath271 .
first , for a fixed system size @xmath18 and particle number @xmath36 , the hilbert space for the @xmath7 system would be much larger than that of the @xmath6 system . a parallel study for @xmath7 of the @xmath6 system sizes and particle numbers reported in this paper , with twist boundary conditions averaging , will require unacceptably long total computation time .
second , and more importantly , we believe that as long as @xmath272 is large , the qualitative implications on the operator - based dm truncation scheme would be similar to the case where @xmath273 , and therefore it suffices to examine the latter case , which is computationally much more manageable . in any case , we do not believe the remnant ` scatter ' in the twist - boundary - conditions - averaged dm weights will hamper our efforts in drawing qualitative conclusions regarding the applicability , or otherwise , of the operator - based dm truncation scheme for interacting fermi liquids . in the operator - based dm truncation scheme described in ref .
, we discard one - particle cluster dm eigenstates with very small weights , and keep only the many - particle cluster dm eigenstates built from the retained one - particle eigenstates .
the sum of weights of the truncated set of cluster dm eigenstates will then be very nearly one , _ if _ the discarded one - particle weights are all very small compared to the maximum one - particle weight .
as we can see from fig.[fig : w1stronginterpolate ] , the ratio of the largest one - particle weight , @xmath177 , to the smallest one - particle weight , @xmath178 , is not large enough for us to justify keeping @xmath263 and discarding @xmath264 , except when the system is very close to half - filled .
we believe that the one - particle cluster dm weights are so close to each other in magnitude , because of the net ` strength ' of interactions straddling the cluster and its environment is strong compared to the net ` strength ' of interactions strictly within the cluster . unfortunately for our five - site cluster , which was chosen because it is the smallest non - trivial cluster having the full point group symmetry of the underlying square lattice , cluster is the smallest cluster having @xmath152 as its point symmetry group .
however , it is trivial as a cluster , because we can have at most @xmath274 particles within this cluster , as opposed to a maximum of @xmath275 particles within the five - site cluster . ]
the sites within the cluster are poorly connected , i.e. a cluster site is on average connected to more environmental sites than to other cluster sites .
to have more of the total interactions of the cluster with the system be confined within the cluster , a cluster significantly larger than the five - site cluster studied in this paper will be needed .
this large cluster must then be embedded in a finite system that is larger still , making exact diagonalization studies unfeasible .
while it is desirable to have a broader distribution of one - particle weights , our more important task is to examine how closely the many - particle cluster dm eigenstates can be approximated as products of one - particle cluster dm eigenstates .
in particular , we look at the two - particle cluster dm eigenstates , and find that of the two - particle states listed in , the only states which are allowed by the no - nearest - neighbor constraint to appear in the cluster hilbert space are @xmath276 , @xmath277 , @xmath278 and @xmath279 .
we know therefore that the two - particle sector of @xmath11 comprises a @xmath280 @xmath281-diagonal block ( with weight @xmath282 ) , a @xmath280 @xmath283-diagonal block ( with weight @xmath284 ) , and two degenerate @xmath285 @xmath36-diagonal blocks ( with weights @xmath286 and @xmath287 ) .
the two - particle weights are shown as a function of filling @xmath2 in fig.[fig : w2strong ] .
for the finite @xmath267 ( ) , @xmath243 ( ) , @xmath244 ( @xmath245 ) , @xmath239 ( @xmath220 ) , @xmath246 ( @xmath247 ) and @xmath248 ( @xmath249 ) systems studied , subject to twist boundary conditions averaging , using @xmath257 monkhorst - pack special - point integration , all the two - particle weights are zero at @xmath258 as the systems approach the noninteracting limit . at half - filling @xmath9 , we again perform degeneracy averaging analytically on the two - fold degenerate checker - board ground state to find that all the two - particle weights are zero . in figure
[ fig : w2strong ] , the solid ` curves ' interpolates between these two known limits and the equally weighted data points at finite filling fractions @xmath262 .
we realized that there are significantly fewer nontrivial two - particle eigenstates of @xmath11 than predicted from the combination of one - particle eigenstates .
from the point of view of implementing the operator - based dm truncation scheme , this poses no problem _ if _ the non - occurring two - particle states are predicted to have small enough weights that they will be excluded by the truncation scheme .
however , we find that this is not the case .
for example , the two - particle state @xmath288 , which does not occur , is predicted by simple combination of the one - particle states @xmath174 and @xmath289 to have a weight comparable to that of the two - particle state @xmath276 , which does occur . of the two - particle weights that are allowed by the no - nearest - neighbor constraint
, we expect their weights to follow the sequence @xmath290 , if they can indeed to thought of products of one - particle states . from fig.[fig : w2strong ] , we indeed observe this sequence of two - particle weights , even though their actual magnitudes ( calculated as the product of one - particle weights divided by the zero - particle weight ) do not come out right .
this observation is encouraging , because we might yet be able to push a variant of the operator - based dm truncation scheme through , by introducing constraints on how many - particle cluster dm states can be built up from one - particle cluster dm states . over broad ranges of filling fractions , the ground state of our strongly - interacting model of spinless fermions
is expected to be an interacting fermi liquid .
while we understand the cluster dm structure of a noninteracting fermi liquid completely , we do not yet understand how an interacting fermi liquid will manifest itself in the structure of its cluster dm .
unlike a noninteracting fermi liquid , the ground - state dm of an interacting fermi liquid will not simply be the exponential of a noninteracting pseudo - hamiltonian .
nevertheless , we expect that the interacting pseudo - hamiltonian @xmath291 appearing in the interacting fermi - liquid ground - state dm @xmath292 can be made to look like the sum of a noninteracting pseudo - hamiltonian @xmath293 , and a much weaker interaction term @xmath294 , by a canonical transformation . from landau s fermi liquid theory ,
we know that such a canonical transformation ( similar to the one which relates landau quasi - particles to bare fermions ) works by burying much of the bare interactions within the quasi - particles . in tracing down the ground - state dm , our hope then is that the cluster dm can also be written , after a canonical transformation local to the cluster , as the exponential of the sum of a noninteracting pseudo - hamiltonian @xmath295 ( which is perhaps related to @xmath293 in the same way as for the noninteracting fermi liquid ) , and a weak interaction term @xmath296 .
we suspect that the criterion for this to be possible is that we must be able to construct approximate quasi - particles using _ only _ cluster states to absorb the bare interactions .
however , we also believe that the requirement that the canonical transformation act strictly within the cluster will fail to completely incorporate interactions that straddle the cluster and its environment .
in section [ sect : strongtwo ] , we found that for our strongly - interacting system , the two - particle cluster dm eigenstates look nothing like simple products of two one - particle cluster dm eigenstates each .
in fact , many combinations of two one - particle cluster dm eigenstates give invalid two - particle cluster states that violate the condition of no - nearest - neighbor occupation .
it is tempting , based on this observation , to say then that the cluster dm is not the exponential of an approximately noninteracting pseudo - hamiltonian .
however , it must be remembered that in an interacting fermi liquid , the quasi - particles are also not single bare particles .
instead , they are superpositions of states containing different number of bare particles , which leads us to think of a quasi - particle as a bare particle being screened by other bare particles in its vicinity . with this in mind , we realized that to identify the quasi - particle structure of the cluster dm , we need to construct appropriate linear combinations of the @xmath36-particle cluster dm eigenstates , so that the cluster dm , when written in terms of these ` quasi - particles ' , look like the exponential of a noninteracting pseudo - hamiltonian @xmath297 plus a weak interaction term @xmath298 .
this involves writing the pseudo - hamiltonian @xmath299 as a sum of terms , representing the independent quantum fluctuations associated with each of the quasiparticles .
this can be accomplished by defining an operator singular value decomposition of the cluster dm with respect to an appropriate operator norm , which forms the basis for judging whether the quantum fluctuations associated with two linear combination of bare operators are independent .
details of such an operator singular value decomposition will be reported in a future paper.@xcite
to summarize , we have calculated numerically the cluster dm for a cross - shaped cluster of five sites within both a system of noninteracting spinless fermions described by the hamiltonian , and a system of strongly - interacting spinless fermions described by the hamiltonian . for the noninteracting system
, the cluster dm was obtained from the cluster green - function matrix using the exact formula obtained in ref . , whereas for the interacting system , the cluster dm was obtained from the exact - diagonalization ground - state wave function by tracing down degrees of freedom outside of the cluster . for the purpose of making the comparison of the cluster dm spectra more straightforward , we worked with the same collection of finite non - square systems for interacting and noninteracting spinless fermions . to make the results of this comparison less dependent on the geometry of the finite systems chosen , degeneracy averaging followed by orientation averaging of the cluster
dm spectra were carried out . when combined , these two averaging apparatus has the effect of restoring full square - lattice symmetry to the cluster dm .
we also analyzed in detail the finite size effects not removed by degeneracy and orientation averaging , by inspecting the numerical cluster dm spectra coming from finite noninteracting systems of various sizes . by increasing the system size systematically
, we find visually that the infinite - system limit of the cluster dm is ` attained ' when the system size reaches a couple of hundred sites .
noting that these are forbiddingly large system sizes to work with for interacting systems , where the ground - state wave functions have to be obtained via exact diagonalization , we then tested the apparatus of twist boundary conditions averaging on finite systems with between 10 and 20 sites . for noninteracting spinless fermions , we find that the twist boundary conditions - averaged cluster dm weights for different finite systems and different filling fractions indeed fall close to the various infinite - system limits .
since we do not perform system - size extrapolations , we interpolate between the degeneracy- , orientation- , and twist boundary conditions - averaged cluster dm weights for the various finite interacting systems and their respective accessible filling fractions , and take the result curves to be our best approximation of the cluster dm spectrum of the infinite interacting system . comparing the twist boundary conditions - averaged cluster dm spectra for the noninteracting and strongly - interacting systems ,
we find similar qualitative behavior in the zero - particle weights as functions of filling fraction , and qualitatively different behaviours in the one - particle weights as functions of filling fraction .
however , the relative ordering @xmath265 is the same at all @xmath300 for both systems .
quantitatively , we find for noninteracting spinless fermions that the one - particle weights go down by roughly one order of magnitude each time as we go through the sequence @xmath301 . for strongly - interacting spinless fermions ,
the one - particle weights decay much more slowly along the sequence . the implications
this observation have for the operator - base dm truncation scheme developed in ref .
is that , for a small fixed fraction of one - particle eigenstates retained , the total cluster dm weight of eigenstates retained would be much smaller for the strongly - interacting system compared to the noninteracting system , since the ratio of the smallest to the largest one - particle weights , @xmath302 , is not very much smaller than one .
this narrow distribution of one - particle cluster dm weights aside , we observed that the relative ordering @xmath303 of the two - particle cluster dm weights , predicted based on the combination of one - particle cluster dm weights , is confirmed numerically even though the predicted weights are off .
this suggests that we might be able to push a variant of the naive operator - based dm truncation scheme through , by introducing additional rules on how one - particle cluster dm eigenstates can be combined to give only the valid many - particle cluster dm eigenstates .
we did not attempt to implement such a truncation scheme , and test how badly its numerical accuracy is affected by the ratio @xmath302 ( by calculating the dispersion relation , for example , as was done in ref . ) , because we feel that such a naive scheme was not in the spirit of finding an appropriate ` quasi - particle ' description for the low - energy excitations of our system of interacting spinless fermions given by .
finally , we realized that our numerical studies do not allow us to conclude whether the cluster dm of the interacting system furnishes a good ` quasi - particle ' description for the strongly - interacting system . to be able to check this
, we must be able to construct appropriate superpositions of cluster dm eigenstates with different particle numbers , so that the pseudo - hamiltonian @xmath304 looks like the sum of a noninteracting pseudo - hamiltonian @xmath297 and a weak interacting pseudo - hamiltonian @xmath298 . instead of a simple eigenvalue problem for the cluster dm
, the problem of discovering what ` quasi - particle ' operators make up the cluster dm is an operator singular value decomposition problem .
we will carefully define this operator singular value decomposition in a future paper , and describe how it can be applied to the density matrix of two disjoint clusters to systematically extract operators associated with independent quantum fluctuations within each cluster , and their inter - cluster correlations.@xcite this research is supported by nsf grant dmr-0240953 , and made use of the computing facility of the cornell center for materials research ( ccmr ) with support from the national science foundation materials research science and engineering centers ( mrsec ) program ( dmr-0079992 ) .
sac would like to thank garnet chan for illuminating discussions on the numerical implementation of the trace - down algorithm .
in this appendix we compare the _ naive algorithm _ and the _ pre - sorted inner product algorithm _ , based on and respectively , for numerically computing the cluster density matrix , and determine their computational complexities .
to begin , we denote by @xmath305 the size of the system hilbert space with @xmath36 particles , @xmath306 the size of the cluster hilbert space with @xmath39 particles , and @xmath307 the size of the environment hilbert space with @xmath42 particles . noting that there can be no matrix elements between cluster configurations with different number of particles , we calculate each @xmath39 sector of the cluster dm separately . to keep our notations compact , let us drop the @xmath36 and @xmath39 dependences in @xmath305 , and @xmath308 respectively from this point onwards , and reinstate these dependences only when necessary .
readers are referred to appendix a in ref . for more technical details on the computational implementation of this trace - down calculation of the cluster dm . in the naive algorithm based on
, the cluster dm matrix elements are computed by starting nested ` for ' loops in @xmath309 and @xmath310 , each running over @xmath311 indices . for each pair of cluster configurations @xmath68 and @xmath69 , one would need to then determine which of the @xmath36-particle configurations @xmath49 contain the two cluster configurations .
this involves running through the @xmath283 configurations in the system hilbert space , and for each configuration , comparing the @xmath36 occupied sites with the @xmath39 occupied sites in the cluster configurations @xmath68 and @xmath69 .
the computational effort incurred for this matching is thus @xmath312 .
two vectors of indices , @xmath313 , whose entries are the indices of system configurations @xmath49 giving cluster configuration @xmath68 , and @xmath314 , whose entries are the indices of system configurations @xmath49 giving cluster configuration @xmath69 , are obtained .
the lengths of these index vectors vary , but are of @xmath315 .
one can then compare the two index vectors , at a computational cost of @xmath316 , to find which pairs of system configurations giving cluster configurations @xmath68 and @xmath69 share the same environment configuration .
following this , one can sum over the amplitude of such pairs , at a computational cost of @xmath315 , to obtain the cluster dm matrix element @xmath317 .
for this naive algorithm , the net computational effort is on the order of @xmath318 . in the pre - sorted inner - product algorithm based on
, we need to first run through @xmath283 system configurations to pre - sort the amplitudes in the ground - state wave function . for each system configuration @xmath49 , we determine at a computational cost of @xmath36 comparisons what cluster configuration @xmath68 and environment configuration @xmath75 it contains .
we then search through the cluster and environment hilbert spaces to determine what the indices of @xmath68 and @xmath75 are in their respective hilbert spaces , which incurs a computational effort on the order of @xmath319 and @xmath320 respectively .
once these indices are determined , the amplitudes in the ground - state wave function are organized into a @xmath321 matrix .
the net computational expenditure is thus on the order of @xmath322 . after sorting the ground - state wave function , we
can then start nested for loops in @xmath309 and @xmath310 , each running over @xmath311 indices , to evaluate the matrix element @xmath317 as the inner product between two vectors of length @xmath323 .
this trace - down stage incurs a computational cost of @xmath324 .
overall , the computational cost is on the order of @xmath325 . for models allowing nearest - neighbor occupation
, the system hilbert space is the direct product of the cluster hilbert space and the environment hilbert space , i.e. @xmath326 .
since the number @xmath36 of particles is small in any reasonable exact diagonalization , we can treat it as a @xmath327 constant . for small clusters , the size @xmath311 of the cluster hilbert space
will also be small , so that the size @xmath323 of the environment hilbert space will be comparable in magnitude to @xmath283 . with these considerations ,
we find that the computational cost for the naive algorithm is @xmath328 , while the computational cost for the inner - product algorithm with pre - sorting is @xmath329 . the efficiency of the two algorithms therefore depend on the prefactors , the estimation of which requires more thorough analyses of the two algorithms . for a model such as , where nearest - neighbor occupation is forbidden , the system hilbert space is smaller than the direct product of the cluster hilbert space and the environment hilbert space , i.e. @xmath330 . given again that @xmath36 and @xmath311 are small numbers ,
the computational cost for the two algorithms are essentially determined by the ratio @xmath331 .
this ratio is strongly dependent on the dimensionality of the problem : the superfluous configurations generated by the direct product of the cluster hilbert space and the environment hilbert space are invalid because they contain nearest - neighbor sites , right at the boundary between the cluster and its environment , which are occupied . in one dimension , the number of superfluous configurations is small , because the boundary between the cluster and its environment consists only of two bonds , whatever the size @xmath3 of the cluster . in two dimensions ,
the boundary between the cluster and its environment is a line cutting roughly @xmath332 bonds .
the number of superfluous configurations is then proportional to @xmath333 , where @xmath334 is a constant prefactor which depends on the shape of the cluster . in @xmath4 dimensions ,
the number of boundary bonds is on the order of @xmath335 , and the number of superfluous states is proportional to @xmath336 .
therefore , in dimensions greater than one , @xmath337 become increasingly larger than @xmath283 as @xmath3 is increased , and the inner - product algorithm with pre - sorting , which involves only one power of @xmath337 , is more efficient than the naive algorithm , which involves @xmath338 .
c. morningstar and m. weinstein , phys .
d * 54 * , 4131 ( 1996 ) ; e. altman and a. auerbach , phys .
b * 65 * , 104508 ( 2002 ) ; e. berg , e. altman , and a. auerbach , phys . rev . lett . *
90 * , 147204 ( 2003 ) ; r. budnik and a. auerbach , phys .
* 93 * , 187205 ( 2004 ) ; s. capponi , a. luchli , and m. mambrini , phys .
b * 70 * , 104424 ( 2004 ) .
g. spronken , r. jullien and m. avignon , phys .
b * 24 * , 5356 ( 1981 ) ; x. zotos , p. prelovek and i. sega , phys . rev .
b * 42 * , 8445 ( 1990 ) ; d. poilblanc and e. dagotto , phys . rev .
b * 44 * , r466 ( 1991 ) ; d. poilblanc , phys . rev .
b * 44 * , 9562 ( 1991 ) ; c. gros , z. phys .
condensed matter * 86 * , 339 ( 1992 ) ; j. t. gammel , d. k. campbell and e. y. loh , jr . , synth . met . * 57 * , 4437 ( 1993 ) ; c. lin , f. h. zong , and d. m. ceperley , phys .
e * 64 * , 016702 ( 2001 ) .
eric w. weisstein .
`` dihedral group d4 . '' from _ mathworld _ a wolfram web resource , http://mathworld.wolfram.com/dihedralgroupd4.html ; m. lax , _ symmetry principles in solid state and molecular physics _
( dover ( mineola , new york ) , 2001 ) ; j. m. hollas , _ modern spectroscopy , third edition _
( john wiley & sons ( chichester ) , 1996 ) . | the reduced density matrix of an interacting system can be used as the basis for a truncation scheme , or in an unbiased method to discover the strongest kind of correlation in the ground state . in this paper
, we investigate the structure of the many - body fermion density matrix of a small cluster in a square lattice .
the cluster density matrix is evaluated numerically over a set of finite systems , subject to non - square periodic boundary conditions given by the lattice vectors @xmath0 and @xmath1 .
we then approximate the infinite - system cluster density - matrix spectrum , by averaging the finite - system cluster density matrix ( i ) over degeneracies in the ground state , and orientations of the system relative to the cluster , to ensure it has the proper point - group symmetry ; and ( ii ) over various twist boundary conditions to reduce finite size effects .
we then compare the eigenvalue structure of the averaged cluster density matrix for noninteracting and strongly - interacting spinless fermions , as a function of the filling fraction @xmath2 , and discuss whether it can be approximated as being built up from a truncated set of single - particle operators . |
as two molecules are brought together , the coulomb interaction of electrons and nuclei in one molecule with those in the other increases , meaning that the eigenstates of the isolated molecules are not eigenstates of the full hamiltonian . however ,
if the intermolecular interaction is weak , it is often useful to think of the system as two interacting molecules as opposed to one supermolecule . to do so , one expands the full hamiltonian all the interactions between particles in either molecule in the basis of molecular states , and the off - diagonal elements of that expansion are the couplings @xcite .
excitonic couplings are interactions between excited states localized on different molecules . in the single - exciton manifold
relevant to weak illumination , the frenkel hamiltonian of a system of two interacting two - level molecules is @xmath3 where @xmath4 is the coupling between the donor and the acceptor , whose excitation energies are @xmath5 and @xmath6 .
the coupling is often described as containing short - range and long - range contributions .
short - range effects include exchange , overlap of donor and acceptor wavefunctions , and exciton transfer mediated by charge - transfer states @xcite . because they depend on the spatial overlap between donor and acceptor wavefunctions , short - range terms decrease exponentially with distance and are consequently often neglected .
neglecting the short - range couplings leaves only the long - range coulomb interaction , @xmath7 where @xmath8 and @xmath9 are spatial coordinates and @xmath10 is the transition density between the ground ( @xmath11 ) and excited ( @xmath12 ) states of molecule @xmath13 ( either the donor @xmath14 or the acceptor @xmath15 ) .
we return to the appropriate choice of the relative permittivity @xmath16 below . because eq . [ eq : coul ] resembles the interaction of two charge distributions , the integral
can be simplified using elementary approximations from electrostatics .
the various approximations summarized in table [ tbl : methods]differ in how they condense all the information in the continuous transition densities into something more manageable and discrete .
the simplest approximation , useful for intermolecular separations much larger than the sizes of the molecules , is the point - dipole approximation ( pd ) , obtained as the lowest - order term in the multipole expansion of eq .
[ eq : coul ] @xcite , @xmath17 where @xmath18 is the separation between the molecules ( i.e. , the centers of their transition densities ) and @xmath19 and @xmath20 are their transition dipole moments , however , pd is not accurate even at intermediate molecular separations @xcite .
a sequence of more accurate approximations can be obtained by representing the transition density as originating from an array of suitably chosen transition charges , @xmath21 the charges @xmath22 and their positions @xmath23 are chosen once and for all by fitting them to the _ ab initio _ transition density , allowing each subsequent coupling calculation to be much faster .
the different methods that have been used extended dipole , tresp , and tdc differ only in the number of transition charges used . in the extended dipole approximation ( ed )
, two transition charges are used per molecule , so that the coulomb interaction becomes @xmath24 where @xmath25 are the positions of the positive and negative charges and @xmath26 is the magnitude of the charges . to be consistent with the point - dipole approximation at large separations , the charge @xmath27 and the distance @xmath28
must be chosen so that @xmath29 , essentially making @xmath27 an additional free parameter , whose tunability ensures ed agrees with the exact results better than pd .
the opposite extreme is the transition density cube ( tdc ) method @xcite , where the transition charges are located on a cartesian grid , making the method a direct numerical integration of eq .
[ eq : coul ] .
the main difficulty is that the grid needs to be fine @xcite , containing many charges even in areas where the transition density is negligible . for bacteriochlorophylls ,
the method converges when the number of charges in each molecule is around 500000 @xcite , making the calculation of their pairwise interactions slow and prone to rounding errors . between the two extremes of ed and tdc
lies the transition monopole approximation ( tma ) @xcite , which assigns one transition charge to each atom .
if hydrogens are excluded , this leads to about 50 charges per bchl , a happy medium between 2 and 500000 .
the atomic transition charges were initially assigned using a mulliken or hirshfeld population analysis @xcite , but this approach is not uniquely defined and was found to not accurately reproduce the shape of the transition density @xcite .
the method of transition charges from electrostatic potentials ( tresp ) @xcite solves this problem by fitting the charges to best represent the transition density . for molecules with few atoms ,
additional fitting parameters can be supplied by placing multipoles at each atom , which slightly improves the accuracy at the shortest separations @xcite .
however , 50 charges offer enough free parameters that tresp is as accurate as tdc for chlorophylls @xcite , which is why we do not include tdc results in this study .
an alternative to transition charges and eq .
[ eq : charges ] is to expand the transition densities in a convenient chemical basis set and compute the resulting integrals in eq .
[ eq : coul ] using optimized quadrature techniques of quantum chemistry @xcite . like tdc
, this approach gives the exact coulomb coupling within the chosen basis , but is more expensive than tresp because it still requires numerical integration every time .
as with tdc , we do not consider this approach here because tresp is sufficiently accurate . once long - range couplings have been calculated using one of the methods surveyed above
, any further improvement must come from including short - range effects .
short - range couplings are particularly important in polycyclic aromatic hydrocarbons ( pah ) @xcite and other flat molecules @xcite whose lack of steric hindrance allows for tight packing . among photosynthetic complexes , large short - range couplings have been reported in lh2 @xcite and in the special pair of reaction centers @xcite .
we return to these cases below .
the simplest situation is the coupling between a donor and an acceptor that are identical molecules . in that case , the eigenenergies of eq .
[ eq : matrixham ] are , meaning that @xmath4 can be obtained by halving the difference between the energies of the two excitonic states , @xmath30 which can be obtained from an electronic - structure calculation of the entire dimer .
the same approach can be extended to heterodimers @xcite .
scholes et al .
used the eigenstates @xmath31 of @xmath32 ( from the quantum - chemical treatment of the dimer ) and the site energies @xmath5 and @xmath6 ( from the quantum - chemical treatment of the monomers ) to calculate @xmath4 @xcite .
doing so assumed that the effective shifts in the site energies of the two molecules ( induced by the presence of the other ) are equal , an approximation that was removed when the method was refined by madjet et al @xcite .
they used the fact that heterodimer eigenstates are not fully delocalized , and that site - energy shifts can be computed from the extent of delocalization , which they obtained by comparing the monomer and dimer transition dipole moments . in this work ,
we use the closely related fragment excitation difference ( fed ) method @xcite , which measures the delocalization in the eigenstates more directly , by determining the difference between the excitation densities on the two molecular fragments . along with @xmath31 ,
this suffices to calculate @xmath4 exactly , given a particular electronic - structure method and basis set .
all coupling and energy calculations were performed on bacteriochlorophyll _ a _ ( bchl__a _ _ ) . in each case , the starting point was bchl__a _ _ taken from the @xmath33 position in the crystal structure of the lh2 complex of the purple bacterium _
_ @xcite .
an important preliminary question is which molecular geometry to use .
it has been argued that the crystal structure is not reflective of the molecular configuration _ in vivo _ and that , consequently , the geometry should be computationally optimized before calculating the excitonic couplings @xcite .
this optimization has generally been carried out using either hf or dft . to the naked eye
, the differences between the crystal structure and the optimized geometry appear small ( figure [ fig : optimization ] ) and one might think that this would result in only a minor correction to the couplings .
however , the electronic properties of chlorophyll molecules are known to vary substantially with the molecular geometry @xcite . to test the influence of geometry optimization on bchl couplings , we compared the electronic - structure results obtained with different geometry optimizations and with experiment .
we calculated the @xmath34 transition energy at the crystal structure geometry and following three different geometry optimizations : hf/ and dft / b3lyp/ ( gradient convergence criterion of in each case , 4.0 @xcite ) , as well as molecular mechanics with allinger s mm3 forcefield @xcite ( gradient convergence criterion of @xmath35kcal / mol / , tinker 7.1 @xcite ) .
the transition energy itself was calculated using both cis/ and tddft / b3lyp/ , as shown in table [ tbl : energies ] .
we found that the cis calculation performed on the crystal structure gave the best agreement with the experimentally measured value for this transition , 770 nm @xcite .
all geometry optimizations introduced errors much larger than the shifts in the @xmath34 peak if the spectrum is measured in different solvents or following aggregation .
this suggests that geometry optimization of this molecule prior to electronic - structure calculations can introduce large and unnecessary errors .
consequently , all the following calculations were performed at the crystal - structure geometry using cis/. table [ tbl : energies ] also compares the @xmath34 energies of the full bchl molecule and of the molecule with its phytyl tail removed .
the removal of the tail changes the transition by at around 1 nm and can therefore be safely performed in the interest of computational speed . )
.,width=302 ] consistency is important to ensure a fair comparison of different methods . in particular
, all the underlying electronic - structure calculations must use the same method and basis set , because different methods can give substantially different results @xcite .
we used cis/ , but this choice is not essential to our argument , since there is no reason to expect that other methods would be less affected by the geometric errors that are central to this paper .
two parameters in coupling calculations are difficult to determine _ ab initio _ : the relative permittivity @xmath16 and the magnitude @xmath36 of the transition dipole moment .
the vacuum value @xmath37 is an appropriate choice for nearest - neighbor couplings in tightly packed aggregates , but for more distant molecules , the coulomb interaction is screened by the intervening medium .
the appropriate choice of @xmath16 for light - harvesting environments has been discussed extensively @xcite , and values from 1 to 2 have been used .
this debate is beyond the scope of this work ; instead , we report vacuum couplings , which can easily be adjusted by multiplication with @xmath38 .
our conclusions about the sizes of relative errors are unaffected , because all the couplings are scaled in the same proportion .
the second uncertain parameter is the magnitude of the @xmath34 transition dipole moment , which tends to be overestimated by cis calculations @xcite . for the crystal - structure geometry ,
cis/ predicts a @xmath34 dipole moment of @xmath39d , significantly larger than the best experimental estimate of 6.1d ( in vacuum ) @xcite . here
again we make the simplest choice , reporting all results with the theoretically predicted value @xmath40 , which can be corrected by multiplication with @xmath41 , where @xmath36 is the desired magnitude of the dipole moment @xcite .
again , the conclusions about the relative errors are unaffected because all couplings would be scaled by the same factor .
having made these preliminary choices , we compared the couplings predicted by pd , ed , tresp , and fed , calculated between two identical bchl molecules , displaced perpendicular to their bacteriochlorin rings by a separation ranging from 5 to 20 ( figure [ fig : separation]a ) .
.the influence of geometry on the calculated wavelength of the @xmath34 transition of bchl__a__. the transition was calculated using both cis and tddft based on the crystal structure and on the geometries obtained by optimizing the crystal structure using different routines .
the best agreement with experiment is obtained at the crystal structure geometry , indicating that geometry optimization may induce large and uncontrolled errors .
in addition , removing the phytyl tail has a negligible effect , as expected . [ cols="<,^,^,^,^ " , ] the means and standard deviations in figure [ fig : orients]b indicate that the difference between tresp and fed is less than the geometric uncertainty in the vast majority of cases .
however , focusing on the mean and standard deviation risks ignoring the outliers , which may indicate additional failures of tresp .
of particular concern are configurations where the two bchls are parallel but offset from each other .
the offset can give rise to a relatively large mg
mg distance even though portions of the two molecules might be close to each other .
the offset - parallel arrangement arises in some natural complexes , giving rise to some of the most strongly coupled bchls in nature . to check the applicability of tresp to those cases , we calculated the couplings between three such pairs , as shown in table [ tab : closest ] . for the two nearest - neighbor couplings in the lh2 complex of purple bacteria ,
the error of ed with respect to fed is over 50% on average , confirming its poor performance seen in figure [ fig : orients]b .
the error of tresp is , as expected , larger than the average in figure [ fig : orients]b , at 14% and 24% for @xmath42 and @xmath43 , respectively , a result consistent with previous work finding short - range corrections of 17% and 24% in the two cases @xcite .
however , the geometric errors are also larger , because small displacements of particular atoms can have an outsized influence on the coupling when those atoms are close together . in the two cases from lh2 ,
the average geometric uncertainty in the tresp couplings is 43% , substantially more than the 19% error with respect to fed .
thus , we can again conclude that the uncertainty in the atomic coordinates is a larger source of error than excluding the short - range contributions .
special pairs in reaction centers are even more strongly coupled , and the results in table [ tab : closest ] indicate that all the methods fail .
in particular , fed shows a considerable admixture of higher excited states , indicating that it is inappropriate to model this case as two coupled q@xmath44 transitions ( eq . [ eq : matrixham ] ) .
madjet et al .
@xcite checked the consistency of the same calculation by comparing the values of the @xmath45 and @xmath46 parameters of their theory ; these values were unequal by a large margin , especially for certain electronic - structure methods , indicating the failure of the effective two - state hamiltonian .
the difference between our coupling and that of madjet et al .
is probably caused by their optimization of the geometry .
the failure of fed indicates that the special pair can not be considered as two coupled molecules but should be seen as one unit .
this can also be diagnosed from the severe failure of tresp , whose geometric uncertainty is over 100% .
therefore , the geometric uncertainty indicates not only the inherent uncertainty of tresp , but also the applicability of the two - state approximation . in difficult cases such as those just discussed ,
the close separation between certain atoms will increase the uncertainty above the typical 20% , indicating that the calculation should be viewed with suspicion .
we have presented a detailed comparison of methods for excitonic coupling calculations of bacteriochlorophylls , with a focus on the uncertanties that arise due to the uncertainties in the atomic positions . for tresp , these geometric uncertainties are much larger than the disagreement between tresp and the more accurate fed , indicating that the short - range contributions to the couplings are almost always a minor correction in comparison with the error bars , making the computational cost of obtaining them ( 67 orders of magnitude more than for tresp ) unjustified . furthermore , tresp is clearly preferrable over other the pd and ed approximations , both of which lead to errors much larger than the geometric uncertainty .
therefore , we recommend the use of tresp for the calculation of excitonic couplings between bacteriochlorophylls at all separations and orientations , with the warning that particularly large geometric uncertainties may indicate a failure of the two - state approximation . our error analysis can be extended to site energies and other components of eet simulations in order to determine the overall sensitivity to uncertainties in the experimental data .
this will allow us to identify the theoretical predictions that do not depend sensitively on microscopic details and are thus more likely to apply to a wide range of pigment - protein complexes .
this work was supported by the australian research council through a discovery early career researcher award ( de140100433 ) and the centres of excellence for engineered quantum systems ( ce110001013 ) and quantum computation and communication technology ( ce110001027 ) .
the molecular structures were drawn using vmd @xcite .
52ifxundefined [ 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 , ed .
( , ) , ed .
( , ) @noop * * , ( ) ( , ) @noop * * , ( ) | excitonic couplings between ( bacterio)chlorophyll molecules are necessary for simulating energy transport in photosynthetic complexes .
many techniques for calculating the couplings are in use , from the simple ( but inaccurate ) point - dipole approximation to fully quantum - chemical methods .
we compared several approximations to determine their range of applicability , noting that the propagation of experimental uncertainties poses a fundamental limit on the achievable accuracy .
in particular , the uncertainty in crystallographic coordinates yields an uncertainty of about 20% in the calculated couplings . because quantum - chemical corrections are smaller than 20% in most biologically relevant cases , their considerable computational cost is rarely justified .
we therefore recommend the electrostatic tresp method across the entire range of molecular separations and orientations because its cost is minimal and it generally agrees with quantum - chemical calculations to better than the geometric uncertainty .
we also caution against computationally optimizing a crystal structure before calculating couplings , as it can lead to large , uncontrollable errors . understanding the unavoidable uncertainties can guard against striving for unrealistic precision ; at the same time , detailed benchmarks can allow important qualitative questions which do not depend on the precise values of the simulation parameters to be addressed with greater confidence about the conclusions .
photosynthesis begins when ( bacterio)chlorophyll molecules in an antenna complexes absorb light @xcite , creating molecular excited states .
this triggers excitonic energy transfer ( eet ) @xcite , the migration of the excited - state energy through the network of ( bacterio)chlorophyll until it decays or reaches the photosynthetic reaction center .
the most common approach to modelling eet is to assume that each ( bacterio)chlorophyll is a site which can be in either the ground or excited states so that the transfer of the excitation from one site to another is mediated by the coupling between them . in principle , the coupling can be calculated if the electronic structures and the relative positions of the two molecules are known . in practice
, a full quantum - chemical treatment is often too expensive , which has led to various approximate methods for calculating excitonic couplings .
this work is about determining the accuracy of these approximations and their range of applicability when they are applied to aggregates of bacteriochlorophylls . at separations that are much larger than the molecular dimensions ,
the leading term in the excitonic coupling is the dipole - dipole interaction of the two transition dipoles .
this motivates the point - dipole ( pd ) approximation , which truncates all the higher - order contributions .
because of its simplicity , the pd approximation has been widely used , even for calculating nearest - neighbor couplings where the small distance between the molecules would indicate that the approximation is inappropriate .
various methods discussed below
go beyond the pd approximation and include more information about the electronic structure , but in a way that keeps the computation tractable . by comparing their performance with accurate quantum - chemical calculations
, one might expect that a hierarchy of methods with defined distance cutoffs could be developed , so that the suitable method would be known for any particular intermolecular distance .
the difficulty with that approach is that setting cutoffs depends on a subjective opinion about what is a tolerable error .
however , an objective error threshold follows from the propagation of errors in the experimentally measured values that enter eet calculations .
it is sufficient to use less accurate methods if the corrections would be drowned out by unavoidable uncertainties already in play .
a particular inescapable error in eet calculations arises from the uncertainties in the atomic positions within each pigment .
we show that a substantial uncertainty in the coupling is obtained by propagating the geometric uncertainties even if high - resolution crystal structures are used . in most cases , this uncertainty exceeds the quantum - mechanical corrections to the coupling , meaning that the substantially cheaper classical calculations are equally reliable .
p10mmp42mmp19mmp28mmp14mmp24 mm method & & short - range coupling & representation oftransition density & cost & benefit + pd & point dipole & no & point dipole & @xmath0 & lowest cost + ed & extended dipole & no & two charges & @xmath0 & ( 7ex,1.7ex ) to + + ( 0,-11.5ex ) ; + tresp & transition charges from electrostatic potentials & no & charges on atoms & @xmath1 & + tdc & transition density cubes & no & charges on a grid & hours@xmath2 & + fed & fragment excitation difference & yes & full & 20h & highest accuracy + |
selective modification of band gaps ( band - gap engineering ) by atomic - scale design of materials is a powerful concept in electronic and photonic development.@xcite band gaps can be altered by , for example , introducing dopants , defects or by exploiting finite size effects.@xcite the physical origin of band - gap variations is a modification in the charge distribution in concert with wavefunction hybridization and modification .
dispersive or van der waals ( vdw ) interactions@xcite also alter the distribution of electronic charges and hence the electron band - structure .
a _ direct effect _ is evident , for example , by considering the formation of the double - dipole configuration@xcite which is the electrostatic signature and inherent nature of a pure vdw binding.@xcite in addition , there are also _ indirect , geometry - induced effects _ of vdw binding on electron behavior .
these indirect effects arise when two material fragments come close to one another , thereby changing the local electron environments as compared to isolated fragments .
for example , vdw - binding can cause smaller amounts of net charge transfer within individual vdw - bonded fragments.@xcite also , wavefunction hybridization will certainly arise when material fragments approach one another , even if this hybridization does not significantly contribute to the binding itself ( in purely dispersive interaction ) .
wavefunction hybridization and pauli exclusion cause scattering of the surface - state electrons in physiosorption of acenes and quinones on cu@xmath4 , even if there is no net charge transfer.@xcite the strength of vdw interactions in surface / adatom systems ( including adclusters or complete overlayers ) can be controlled by the precise choice of the surface material@xcite and by the surface morphology ( flat , stepped , pyramidal , _
etc._).@xcite it is therefore important to quantify the extent to which van der waals ( vdw ) bonding can modify the electron dispersion , that is , the band structure .
such a study is now possible , since recent development of the vdw density functional ( vdw - df ) method @xcite enables systematic ( theoretical ) explorations of bonding in sparse materials within density functional theory ( dft).@xcite [ cols="^ " , ]
the most pronounced electronic difference between bulk graphane and ml or bl graphane is the fact that the a point takes over the role of the @xmath5 point , see fig .
[ fig : bs1 ] . at a
, the band gap is smallest in the bulk and the size of that gap coincides essentially with the size of the gap at @xmath5 in the ml or bl .
also , in the bulk , the deviations in the band - gap variations can be considerably larger than those in the bl , see bottom panel of fig .
[ fig : bs1 ] . here , we give a more detailed analysis of the overall band structure effects associated with graphane assembly into bulk . in the the left panel of fig .
[ fig : bs3 ] , we show the band diagram along @xmath6 , @xmath7 , and @xmath8 .
focusing on the cb , the bulk band structure is very different from that of the bl along the same path , see top panel of fig .
[ fig : bs1 ] . in particular at the @xmath5 point , that gap is approximately twice as large as in the bl .
also , the band energy at k is considerably higher than in the bl . only at m
, we find similar band energies .
thus , the effect of vdw - bonding on local features of the band structure and related observable properties can be dramatic . in the mid - panel of fig .
[ fig : bs3 ] , we show show the band diagram along @xmath9 , @xmath10 , and @xmath11 . again , the the bulk band structure is very different from that of the bl along the parallel path along @xmath6 , @xmath7 , and @xmath8 . the band energy at h is lower than at l , while it is the other way around at k and at m in the bl . only around the @xmath12 point (
corresponding to @xmath5 in the bl and ml cases ) do the bulk and ml band diagram show similar features . in the right panel of fig .
[ fig : bs3 ] , we effectively combine the two diagrams from the left and mid - panel , calculating the band diagram for a would - be 2-layer unit cell at @xmath13 .
the resulting band diagram shows the features of the bl band diagram and of the vb there is also a very good quantitative agreement .
therefore , the bl band - diagram can partly be understood as a zone - folded version of the bulk band - diagram .
this paper predicts and characterizes band - structure modifications produced indirectly as a geometry - induced effect of pure dispersive binding between two macromolecules in possible new material systems : bilayer and bulk graphane.@xcite using non - selfconsistent vdw - df calculations,@xcite we determine the binding separation ( @xmath14-lattice parameter ) and binding energies in vdw - bonded graphane bilayers ( bulk ) .
we use the calculated separations to obtain corresponding gga - dft band diagrams .
our results demonstrate that vdw interactions can significantly alter electron behavior , at least locally in the brillouin zone . in graphane ,
the direct band gap is reduced ( increased ) by up to 1.2 ev ( 4 ev ) .
our analysis shows that the origin of the band - gap modifications in this system is the concerted action of two geometry - induced effects .
the first cause is the hybridization between _ unoccupied _ wave functions in the lowest conduction band .
this effect generally leads to lower ( higher ) energy state for the bonding - type ( antibonding - type ) hybrid wave function , but not always , since the graphane wavefunctions can possess a more complex hybridization nature than what applies in the hydrogen - like case .
the second cause is the modified electrostatic interaction between the hybrid wave functions and the electron density .
this cause can either increase or decrease the energy gain ( loss ) in a bonding - type ( antiboning - type ) hybridization .
moreover , it may result into a modified hybrid wave - function where a pure bonding - type ( antibonding - type ) character is lost .
our analysis also shows that selfconsistent charge rearrangements ( on the gga level ) with respect to the ml density have no significant impact on the bandstructure .
additional charge rearrangements described by self - consistent vdw - df calculations@xcite are expected to be even smaller , justifying our use of non - selfconsistent vdw - df .
the nature of conduction and optical absorption in bl graphane would be determined by the region around the @xmath5 point .
there , the band structure of the bilayer essentially coincides with that of the monolayer .
therefore , we expect graphane multilayers to behave electronically similar to a graphane monolayer , at least for properties defined by a simple response . for bulk graphane ,
the behavior is more complicated .
we emphasize that other vdw - bonded systems may exist where significant band - gap modifications arise at brillouin - zone points having higher relevance for the electronic nature of the material .
our results for bl graphane suggest that vdw forces can have non - negligible _ indirect _ effects on the overall band structure in layered or macromolecular materials . a similar effect can be found in v2o5,@xcite where traditional gga severely overestimate the @xmath14 lattice constant and where vdw - df provides a more accurate description.@xcite furthermore
, we notice that vdw binding of the intrinsic semimetal graphene to metal@xcite or semiconductor@xcite surfaces seems to generally lead to a shift of the fermi level , rendering graphene a true metal . these observations together with the fact that the vdw binding strength in surface / adatom systems ( and thus presumably the strength of the corresponding effect on the band structure ) depends on the choice of the substrate material@xcite and on the substrate morphology@xcite implies a possibility to exploit dispersive interactions also in band - gap engineering
. finally , a comment on the accuracy of the predicted band - gap modifications is in order .
gga dft typically severely underestimates the band gap in semiconductors or insulators .
a possible remedy is the use of the gw method where the kohn - sham orbitals are used to construct the self - energy operator .
the gw method also contains non - local correlations and therefore provides an alternative description of vdw interaction .
gw calculations come at a considerably higher computational cost than that of vdw - df , however .
the procedure illustrated here , using vdw - df calculations to determine binding morphologies followed by band - structure calculations might well also be adapted for gw , pursuing characterizations of band - structure effects in sparse matter .
we thank g. d. mahan for encouragement and discussions . support by the swedish national graduate school in materials science ( nfsm ) , the swedish research council ( vr ) and the swedish national infrastructure for computing ( snic )
is gratefully acknowledged .
langreth , b.i .
lundqvist , s.d .
chakarova - kck , v.r .
cooper , m. dion , p. hyldgaard , a. kelkkanen , j. kleis , lingzhu kong , shen li , p.g .
moses , e. murray , a. puzder , h. rydberg , e. schrder , and t. thonhauser , _ j. phys . :
condensed matter _ * 21 * , 084203 ( 2009 ) .
e. ziambaras , j. kleis , e. schrder , and p. hyldgaard , phys .
b 76 , 155425 ( 2007 ) .
s. d. chakarova - kck , e. schrder , b. i. lundqvist , and d. c. langreth , _ phys .
lett . _ * 96 * , 146107 ( 2006 ) .
k. johnston , j. kleis , b. i. lundqvist , and r. m. nieminen , _ phys .
b _ * 77 * , 121404(r ) ( 2008 ) .
j. kleis , e. schrder , and p. hyldgaard , _ phys .
b _ * 77 * , 205422 ( 2008 ) .
we have also mapped out the vdw - df potential energy landscape ( pes ) that arises when one of the graphane sheets is slightly displaced in various directions from this @xmath15 high - symmetry configuration .
this stability test was performed for various layer separations around the binding separation predicted by vdw - df2 . on the pbe level ,
the displaced configurations essentially do not relax , reflecting the essentially flat pbe pes at these separations . however , adding the nonlocal vdw correction by evaluating eq .
( [ eq : ebind ] ) , we find that the high - symmetry configuration possesses the lowest total energy of all considered systems .
we notice that full relaxation of the considered high - symmetry configurations using lda also predicts the @xmath15 to be most favorable .
also , the predicted lda binding separation of @xmath16 is in fair agreement with those predicted here by the various versions of vdw - df ( whereas the lda - predicted layer - binding energy of @xmath17 mev / cell is considerably smaller ) .
however , we stress that this correspondence with vdw - df results coincidentally .
lda binding in systems which ( like multilayer graphane ) are bound by vdw forces has been assigned to unphysical long - range exchange interactions and lda does not , by any means , contain vdw interactions , see j. harris , phys .
b * 31 * , 1770 ( 1985 ) ; e. d. murray , k. lee and d. c. langreth , j. chem . theory comput . * 5 * , 2754 ( 2009 ) and references therein . in the present case
the charge rearrangements are not accompanied by any significant charge transfer .
our bader analysis @xcite shows that the difference in charge before and after the charge rearrangement is @xmath18 e / atom . in cases where charge transfer takes place ,
see for example ref .
, such rearrangements may have more significance . | we use density functional theory and the van der waals density functional ( vdw - df ) method to determine the binding separation in bilayer and bulk graphane and study the changes in electronic band structure that arise with the multilayer formation . the calculated binding separation ( distance between center - of - mass planes ) and binding energy are @xmath0 ( @xmath1 ) and @xmath2 mev / cell ( @xmath3 mev / cell ) in the bilayer ( bulk ) , depending on the choice of vdw - df version .
we obtain the corresponding band diagrams using calculations in the ordinary generalized gradient approximation for the geometries specified by our vdw - df results , so probing the indirect effect of vdw forces on electron behavior .
we find significant band - gap modifications by up to -1.2 ev ( + 4.0 ev ) in various regions of the brillouin zone , produced by the bilayer ( bulk ) formation . |
the saddle - point method is one of the most essential tools in statistical physics @xcite .
when comparing different statistical ensembles , it is used with overwhelming success in both fundamental theoretical considerations and practical calculations @xcite . yet
, the conventional form of this usually easy - to - handle approximation fails in the case of condensed ideal bose gases @xcite ; for instance , it does not yield the correct fluctuation of the number of condensate particles @xcite . to elucidate the reason for this failure we consider an ideal bose gas with single - particle energies @xmath0 , with @xmath1 labeling the individual energy eigenvalues . since the grand canonical partition function @xmath2 generates the canonical partition functions @xmath3 by means of the expansion @xmath4
each @xmath5-particle partition function @xmath3 can be represented , according to cauchy s theorem , by a contour integral in the complex @xmath6-plane , @xmath7 where the path of integration encircles the origin counter - clockwise . as usual
, @xmath8 is the inverse temperature . denoting the negative logarithm of the integrand as @xmath9 , i.e. , writing @xmath10 or @xmath11 the saddle - point @xmath12
is determined by the requirement that this function becomes stationary , @xmath13 giving @xmath14 apart from the appearance of one extra particle on the left hand side , this is just the grand canonical relation between particle number @xmath5 and fugacity @xmath12 .
proceeding according to folk wisdom , one then expands the logarithm @xmath9 quadratically around @xmath12 and leads the path of integration parallel to the imaginary axis over the saddle , relying on the fact that for large @xmath5 the main contribution to the integral ( [ zin ] ) is collected in the immediate vicinity of the saddle - point @xcite , so that the quadratic expansion should prove sufficient . doing the remaining gaussian integral , one arrives at the standard saddle - point approximation @xmath15 to the canonical partition functions , @xmath16 here and in the following we write @xmath17 for the @xmath18-th derivative of a function @xmath19 at a saddle - point ; we have used @xmath20 .
hence , one finds @xmath21 from which the canonical occupation number @xmath22 of the state @xmath23 is obtained by differentiating once with respect to @xmath24 , @xmath25 as long as the gas is not condensed , the fourth term in the approximation ( [ lsp ] ) remains small in comparison to the third , and therefore may be neglected
. then the partial derivative @xmath26 vanishes as a consequence of the saddle - point equation ( [ sa1 ] ) , and one is left with @xmath27 so that within these approximations the canonical occupation numbers equal their grand canonical counterparts . however , this reasoning breaks down for temperatures below the onset of bose
einstein condensation , where eq .
( [ sa1 ] ) requires that @xmath28 be on the order of @xmath29 , so that the third and the fourth term on the right hand side of eq .
( [ lsp ] ) are of comparable magnitude , namely of the order @xmath30 .
hence , the argument that led to the familiar formula ( [ won ] ) becomes invalid .
moreover , inspecting the higher derivatives of @xmath9 at the saddle - point , @xmath31 \ ; , \label{der}\ ] ] one finds that in the condensate regime these derivatives grow dramatically with increasing @xmath18 , @xmath32 , casting doubt on the validity of the saddle - point approximation even if higher - order terms are included @xcite . however , for the derivation of an asymptotic series the convergence properties of the formal taylor series of @xmath9 are irrelevant @xcite ; what actually endangers the approximation scheme ( [ wsa ] ) in the condensate regime is the narrow approach of the saddle - point to the ground - state singularity @xmath33 of the grand canonical partition function ( [ gpf ] ) .
namely , the gaussian integral ( [ wsa ] ) can well represent the exact expression ( [ zin ] ) only if the function @xmath9 is free of singularities at least in those intervals where the gaussian is still large , i.e. , where @xmath34 is on the order of unity .
since @xmath35 , this observation translates into the requirement that _ the function @xmath9 should be regular at least in an interval of order @xmath36 around @xmath12_. on the other hand , _ in the condensate regime the singularity at @xmath33 falls within a distance of order @xmath36 from the saddle - point @xmath12 _ , again as a consequence of eq .
( [ sa1 ] ) .
it is this conflict , _ not _ the poor behavior of the taylor expansion of @xmath9 , which necessitates an approach to the contour integral ( [ zin ] ) that is essentially more careful than the standard scheme ( [ wsa ] ) . interestingly ,
though , the dilemma does not appear to be overly severe the magnitude of both conflicting intervals being of the _ same _ order @xmath36 , and one might wonder already at this point whether the usual procedure can be saved by simple means . the failure of the saddle - point approximation ( [ wsa ] ) for condensed bose gases has been the subject of a long debate in the literature @xcite , with notable early contributions by dingle @xcite and fraser @xcite .
various schemes have been designed for computing the number of condensate particles , and its fluctuation , for gases with a fixed number @xmath5 of particles , _ without _ resorting to the saddle - point method .
quite recently , navez _ et al .
_ have suggested a statistical ensemble within which one regards the condensate as an infinite particle reservoir for the excited - states subsystem @xcite , thus putting into shape an idea already expressed by fierz @xcite . in a mathematical setting , each excited single - particle level @xmath37 then becomes formally equivalent to a harmonic oscillator with frequency @xmath38 , which allows one to derive elegant integral representations for canonical and microcanonical expectation values : setting @xmath39 for convenience , restricting oneself to temperatures below the onset of condensation , and introducing the spectral zeta function @xmath40 with the sum running over the excited states only , the canonical number of condensate particles can be written as @xcite @xmath41 where @xmath42 is the riemann zeta function .
the path of integration up the complex @xmath43-plane lies to the right of all poles of the integrand , so that the residues of these poles , taken from right to left , yield a systematic expansion of the integral @xcite .
similarly , the canonical mean - square fluctuation of the number of condensate particles takes the form @xmath44 and the difference between canonical and microcanonical fluctuations is given by @xmath45 ^ 2 } { \frac{1}{2\pi i } \int_{\tau - i\infty}^{\tau+i\infty } \ !
{ \mbox d } t \ , \gamma(t ) z(\beta , t-2 ) \zeta(t-1 ) } \ ; . \label{idf}\ ] ] yet , there are at least two reasons not to be content with this state of affairs .
firstly , since eqs .
( [ icn ] ) , ( [ icf ] ) , and ( [ idf ] ) rely on the presence of a reservoir of condensate particles , they are blind to the onset of bose einstein condensation , that is , to the sudden appearance of this reservoir ; as a consequence of this underlying `` oscillator approximation '' , they do not allow one to discuss just how such a sharp feature can emerge in a bose gas with a large , fixed number of particles . secondly ,
for a given single - particle spectrum the integrals may not always be straightforward to evaluate ; already the treatment of an anisotropic harmonic oscillator potential , in which case @xmath46 is related to zeta functions of the barnes type @xcite , requires quite some analytical skills .
therefore , one desires a tool that works , in principle , for all temperatures , and is easy to use in practical calculations .
the development of such a tool is the objective of the present paper . following a trail pioneered by dingle @xcite
, we work out and test a variant of the saddle - point method that fulfills the above two requirements .
our guiding maxim is the same which already governed london s classic analysis @xcite of the condensation phenomenon : if the ground state is causing trouble , single it out and give it a special treatment which , in our case , means to exempt the ground - state factor of the grand canonical partition function ( [ gpf ] ) from the quadratic expansion performed in the usual scheme ( [ wsa ] ) , and to treat that factor exactly .
as will be demonstrated in detail , this natural strategy leads to an approach to the canonical and the microcanonical statistics of condensed bose gases which is both extremely accurate and unsurpassed in computational ease .
we proceed as follows : in the next section we will first explain _ why _ and _ how _ this proper saddle - point approximation works for the calculation of the canonical partition functions , and how known limiting cases are recovered , concentrating key technical details in appendices [ aa ] and [ ab ] .
we then apply the method to computing canonical occupation numbers and their fluctuations , illustrating the analytical results by numerical calculations for an ideal bose gas confined by an isotropic harmonic potential . in section [ s3 ]
we turn to the microcanonical ensemble , and show how the very same refined saddle - point approach allows one to obtain the interesting quantities almost without further effort .
the final section [ s4 ] summarizes the most important findings .
although our discussion is led along the lines of the ideal bose gas , and although some interesting physical insights will turn up in particular , it will become clear why the canonical occupation numbers are well described by the expression ( [ won ] ) , and thus essentially equal to their grand canonical analogues , even for temperatures below the onset of condensation , where the previous argument had failed
, this is not primarily a work on the ideal bose gas as such .
rather , major emphasis lies on the mathematics of saddle - point integrals with a singular integrand , of which condensed bose gases provide perhaps the most prominent examples ; we hope that the detailed exposition presented here will prove fruitful also in other areas of mathematical physics where similar problems arise .
since the conventional approximation ( [ wsa ] ) is thwarted by the fact that the saddle - point @xmath12 , i.e. , the solution to eq .
( [ sa1 ] ) , approaches the ground - state singularity @xmath33 of the integrand ( [ int ] ) within order @xmath36 in the condensate regime , we now exclude the ground - state contribution from the quadratic expansion of the logarithm ( [ log ] ) .
that is , we define @xmath47 and write the canonical @xmath5-particle partition function as @xmath48 the key idea for treating integrals of this kind , due to dingle @xcite , is to let the potentially dangerous denominator stand as it is , and to expand only the tempered function @xmath49 around the saddle - point @xmath12 .
the resulting approximation to @xmath3 will then be valid for _ all _ temperatures : for high @xmath50 , when the saddle - point moves away from the singularity , it is of no concern whether or not the denominator is included in the gaussian approximation ; in the condensate regime its exclusion is crucial .
let us first check the behavior of the derivatives @xmath51 . when omitting the ground - state term ( @xmath52 ) from the sum in eq .
( [ der ] ) , the behavior of the remaining sum is governed in the condensate regime by the following terms ( @xmath53 ) , each of them being about proportional to the @xmath18-th power of temperature . hence , @xmath54^{1/n } \propto k_b t \ ; , \label{pr1}\ ] ] with approximate @xmath50-proportionality holding the better , the larger @xmath18 , since large @xmath18 emphasize the low - lying states .
now we focus on systems with single - particle energies of the form @xcite @xmath55 where @xmath56 are integer quantum numbers .
the dimensionless anisotropy coefficients @xmath57 should be of comparable magnitude , the lowest of them equaling unity .
this class of systems contains , e.g. , a gas of @xmath5 ideal bose particles confined by a @xmath58-dimensional harmonic oscillator potential ( @xmath59 ) , or by a @xmath58-dimensional hard box ( @xmath60 ) . for @xmath61 and large @xmath5
there is a sharp onset of bose einstein condensation , with @xmath62 in the condensate regime ; @xmath63 is the total number of excited particles @xcite .
thus , for largish @xmath18 the eqs .
( [ pr1 ] ) and ( [ pr2 ] ) give @xmath64 resulting , by virtue of eq .
( [ der ] ) , in @xmath65 as long as , besides @xmath66 , also @xmath67 . to confirm this relation , the case @xmath68
of which will be of particular interest later , fig .
[ f_1 ] shows the numerically computed quantities @xmath69 ( which , in the general case , should approach @xmath70 for large @xmath18 ) for a gas of @xmath71 ideal bose particles kept at temperature @xmath72 in a three - dimensional isotropic harmonic oscillator potential , with @xmath73 denoting the condensation temperature in the large-@xmath5-limit @xcite ; @xmath74 is the oscillator frequency .
as expected , @xmath75 approaches the value @xmath76 fairly rapidly with @xmath18 . for comparison ,
1 also shows the corresponding data for a gas with the same number of particles which is stored in a one - dimensional harmonic potential .
its temperature is @xmath77 , where @xmath78 this is a borderline case : for @xmath79 there is no sharp onset of bose
einstein condensation , so that @xmath80 merely plays the role of a characteristic temperature below which the ground - state population becomes significant .
there are logarithmic corrections @xcite which keep the ratios @xmath75 below the value @xmath81 also for large @xmath18 , but , as seen in the figure , even now @xmath75 rapidly approaches a constant not too far from unity .
we conclude that for systems of the type ( [ gro ] ) , with @xmath61 , the @xmath82-growth of @xmath83 in the condensate regime is replaced by the somewhat milder @xmath84-growth of @xmath51 when going from @xmath9 to its ground - state - amputated descendant @xmath49 , so that also the formal taylor expansion of @xmath49 around the saddle - point @xmath12 is ill - behaved . but as already indicated , the properties of the taylor series of @xmath49 are only of secondary importance .
what really matters is that this function does not share the ground - state singularity ; the singular point to be watched now is the one at @xmath85 . since @xmath86 ,
the saddle - point remains separated from that singularity by at least the @xmath5-independent gap @xmath87 .
this guarantees that when quadratically expanding the amputated function @xmath49 , rather than @xmath9 , around @xmath12 , an interval of the required order @xmath88 becomes singularity - free for sufficiently large @xmath5 ; the higher the temperature ( while remaining in the condensate regime ) , the smaller the gap , and the larger the particle number has to be .
then the gaussian approximation to @xmath89 is safe . as shown in detail in appendix [ aa ] , the subsequently emerging saddle - point integral for the canonical partition function ( [ cpf ] ) can be done exactly , yielding ( cf .
( [ csa ] ) with @xmath90 ) @xmath91 where @xmath92 and @xmath93 is a parabolic cylinder function , employing whittaker s notation @xcite . for discussing this unfamiliar - looking expression ( [ czn ] ) , which , as already remarked above , is valid for _ all _ temperatures , we observe @xcite that @xmath94 is related to the complementary error function @xmath95 , @xmath96 hence @xmath97 for high temperatures , well above the condensation point , @xmath12 approaches zero , so that the parameter @xmath98 , and as a consequence also @xmath99 , grows without bound when @xmath5 becomes large .
then we may replace the complementary error function by the leading term of its asymptotic expansion for large _ positive _ arguments @xcite , @xmath100 this is a special case of the approximation ( [ hta ] ) introduced in appendix [ ab ] , and implies , together with the further approximations ( [ hts ] ) ( [ htf ] ) , that the general expression ( [ czn ] ) correctly approaches the standard saddle - point result ( [ wsa ] ) outside the condensate phase , @xmath101 this was to be expected , since when the denominator in the integrand ( [ cpf ] ) does not become small , it does nt matter whether it is included in the quadratic approximation , as in the scheme ( [ wsa ] ) , or treated exactly , as in the derivation of eq .
( [ czn ] ) . in the condensate regime , where @xmath102 and @xmath103 as specified by eq .
( [ ord ] ) , the definition ( [ cxn ] ) of the parameter @xmath98 implies either @xmath104 or @xmath105 , whichever is larger .
since we require @xmath61 , we conclude that in either case @xmath98 approaches zero for large @xmath5 , so that now @xmath106 is a large _ negative _ number .
therefore , we may safely use the approximation @xmath107 , and arrive at @xmath108 even when the particle number @xmath5 is merely moderately large , the decrease of the parameter @xmath99 with temperature from large positive to large negative values can be fairly rapid , so that the complementary error function in the canonical @xmath5-particle partition function ( [ cef ] ) acts as a switch , meaning that the transition from the high - temperature asymptotics ( [ wsa ] ) to the condensate asymptotics ( [ coa ] ) becomes quite sharp .
this is confirmed by fig .
[ f_2 ] , which depicts @xmath109 as function of temperature for a gas of @xmath110 ideal bosons in a three - dimensional isotropic harmonic potential . in
the borderline case of the one - dimensional harmonic potential , also indicated in the figure , there is no such sharp transition .
it is worthwhile to discuss the condensate partition function ( [ coa ] ) a little further . anticipating that , despite the incorrect reasoning
, the expression ( [ won ] ) for the canonical occupation numbers will remain valid approximately even in the condensate regime , we have @xmath111 hence , we may eliminate the saddle - point @xmath12 by setting @xmath112 yielding @xmath113 ) } \ ; . \label{opf}\ ] ] the infinite product on the right hand side , describing the excited - states subsystem , equals the canonical partition function of a collection of infinitely many , distinguishable harmonic oscillators with frequencies @xmath114 ( where @xmath53 ) , thus leading back to the `` oscillator approximation '' which has been the starting point for the derivation of the integral representations ( [ icn ] ) , ( [ icf ] ) , and ( [ idf ] ) in ref . @xcite .
in contrast , the value of the present approximation ( [ opf ] ) lies in the fact that it is not restricted to the excited states , but also contains the ground state explicitly . taking the derivatives with respect to @xmath115 , with @xmath116 ,
one then finds occupation numbers @xmath117 and mean - square fluctuations @xmath118 of the excited states , while differentiating the logarithm of eq .
( [ opf ] ) with respect to @xmath119 produces first the obvious identity @xmath120 and then the important equation @xmath121 stating that , in the condensate regime and subject to the above approximations , within the canonical ensemble the occupation numbers of the excited states are _ uncorrelated _ stochastic variables @xcite .
remarkably , the error one would have committed had one naively employed the standard approximation ( [ wsa ] ) in the condensate regime , and which can be quantified only now , is not devastating . as explained in appendix [ ab ] ( cf .
( [ lte ] ) ) , _ in the low - temperature , large-@xmath5-regime the incorrect partition function @xmath122 exceeds the correct partition function by merely the temperature - independent factor @xmath123 , regardless of the single - particle spectrum _ ,
that is , regardless of the trapping potential implying that the error might even go unnoticed when carelessly taking derivatives of @xmath124 .
this finding is illustrated in fig .
[ f_3 ] , again for an ideal bose gas with @xmath125 particles in a three - dimensional isotropic harmonic potential .
the figure shows the ratios of @xmath122 , computed according to the scheme ( [ wsa ] ) , of the correct approximation ( [ czn ] ) , and of its condensate descendant ( [ coa ] ) , to the exact canonical partition function , the latter having been obtained from the one - particle partition function @xmath126 by means of the familiar recursion formula @xcite @xmath127 as witnessed by fig .
[ f_3 ] , the quality of the proper approximation ( [ czn ] ) is outstanding for all temperatures , its low - temperature variant ( [ coa ] ) performs bravely where it is expected to , and the standard approximation ( [ wsa ] ) is good at high temperatures , but fails by the predicted factor @xmath128 in the condensate regime . for comparison , fig .
[ f_4 ] shows the corresponding data for the one - dimensional case ; here the exact @xmath5-particle partition function is known in closed form @xcite .
we find features that are qualitatively similar to those in the preceding figure , although , with @xmath125 , the approximation ( [ czn ] ) is not quite as good at intermediate @xmath50 . the standard scheme ( [ wsa ] ) again is off by the same , universal factor @xmath128 at low temperatures .
the derivation of the expressions ( [ oan ] ) and ( [ oag ] ) from the condensate approximation ( [ coa ] ) to the canonical partition function serves to render that partition function plausible , but it is not the best one can do , since we have simply _ assumed _ the validity of eq .
( [ nzc ] ) .
an accurate and fully consistent computation of these occupation numbers starts from the identity @xmath129 with @xmath130 the equation that determines the saddle - point @xmath131 for the new contour integral ( [ con ] ) , namely @xmath132 formally looks like the grand canonical relation between particle number @xmath5 and fugacity @xmath131 for a system with an extra energy level @xmath37 .
now we have to distinguish two cases : if @xmath116 , we merely have to copy the steps made in the derivation of the proper canonical partition function ( [ czn ] ) .
that is , we separate the ground - state contribution from the exponent @xmath133 by defining the tempered function @xmath134 and obtain @xmath135 proceeding at once to temperatures below the onset of condensation .
then eq .
( [ con ] ) , together with the previous result ( [ coa ] ) for @xmath3 , yields the expression @xmath136 for the canonical occupation numbers of the excited states in the condensate regime .
this accurate result is well approximated by the previous formula ( [ oan ] ) : since the level that has artificially been doubled differs from the ground state , we may set @xmath137 in eq .
( [ csn ] ) , and obtain @xmath138 using the definitions ( [ tef ] ) and ( [ teg ] ) of the functions @xmath139 and @xmath140 .
if , however , @xmath141 , the exponent @xmath133 in eq . ( [ con ] ) corresponds to a system with a doubled ground state .
hence , we have to temper this function accordingly , and define @xmath142 ^ 2\right ) \ ; .\ ] ] this leads to @xmath143 where we have employed the condensate approximation ( [ lta ] ) to the general saddle - point formula ( [ csa ] ) , with @xmath144 .
thus , in the condensate regime the canonical ground - state occupation number is given by @xmath145 for showing that this cumbersome expression actually is consistent with the familiar grand - canonical result in the large-@xmath5-limit , we now have to carefully keep track of the two different parameters @xmath12 and @xmath131 .
that is , we may set @xmath146 whereas @xmath147 reflecting the double appearance of the ground state in eq .
( [ nsp ] ) .
hence , the argument of the exponential in the ground - state occupation formula ( [ csz ] ) should be approximately equal to zero .
this follows by observing @xmath148 which , upon inserting @xmath149 as obtained from eqs .
( [ pl1 ] ) and ( [ pl2 ] ) , indeed gives the required relation @xmath150 for demonstrating the accuracy of the canonical formulas ( [ csn ] ) and ( [ csz ] ) , we resort once more to a gas of @xmath125 ideal bose particles in a three - dimensional harmonic oscillator potential . figure [ f_5 ] shows the occupation number @xmath151 as a function of temperature , computed according to eq .
( [ csn ] ) , and compares these data to those that are obtained if @xmath3 and @xmath152 are naively calculated from the standard saddle - point formula .
the inset quantifies the ratios of these approximate occupation numbers to the exact ones , which have again been computed recursively . in the condensate regime ,
the naive approximation to _ both _ @xmath3 and @xmath152 is off by the same universal factor @xmath128 derived in appendix [ ab ] , so that this error cancels when forming their ratio : the standard saddle - point scheme accidentally yields the correct canonical occupation numbers of the excited states both above and below the onset of condensation . on the other hand ,
if one uses even at high temperatures the approximations ( [ coa ] ) and ( [ dzc ] ) , each of which is correct in the condensate regime only , it follows from eq .
( [ hte ] ) that the ratio of the individual errors is given by the square root of @xmath153 . since this ratio approaches unity in the large-@xmath5 , high-@xmath50-regime , eq .
( [ csn ] ) actually is correct also at high temperatures .
this accidental correctness is no longer met in the case of the ground - state occupation number @xmath154 , depicted in fig .
the error introduced when computing @xmath155 at low temperatures within the standard scheme is given by the factor @xmath156 ( see appendix [ ab ] ) , so that the resulting value of @xmath154 is too small by the constant factor @xmath157 in the condensate regime , whereas eq .
( [ csz ] ) yields the correct data . outside the condensate regime
the standard approximation becomes correct , whereas eq .
( [ hte ] ) reveals that the condensate approximation ( [ csz ] ) is wrong by the factor @xmath158 .
needless to say , if one deduces the canonical occupation numbers directly from the ratios of the saddle - point approximations ( [ csa ] ) to the respective contour integrals , instead of invoking their high- and low - temperature limits ( [ usr ] ) and ( [ lta ] ) , one obtains expressions that are valid for all temperatures .
the calculation of the canonical mean - square fluctuations @xmath159 of the occupation numbers now directly parallels that of the occupation numbers themselves : starting from the identity @xmath160 we are left with the task to evaluate the further contour integral @xmath161 ^ 2 } \ ; , \ ] ] where @xmath162 as a consequence of the second derivative performed in eq .
( [ cof ] ) , the state with energy @xmath37 now has formally been tripled , giving the equation @xmath163 for the new saddle - point @xmath164 .
thus , for @xmath116 the tempered version of the function @xmath165 becomes @xmath166 implying @xmath167 in the condensate regime ; hence @xmath168 setting @xmath169 , similar to the reasoning underlying eq .
( [ zez ] ) , it can be seen that the third term on the right hand side is close to @xmath170 , so that this expression ( [ cfn ] ) properly reduces to the familiar eq .
( [ oaf ] ) .
if @xmath171 , we disentangle the three - fold ground - state contribution from @xmath165 by defining @xmath172 ^ 3\right ) \ ; , \ ] ] and get @xmath173 employing the condensate approximation ( [ lta ] ) with @xmath174 . therefore
, the canonical fluctuation of the number of condensate particles now takes the form
@xmath175 we check the results ( [ cfn ] ) and ( [ cfz ] ) again for @xmath125 ideal bosons in an isotropic harmonic oscillator potential .
figure [ f_7 ] shows the root - mean - square fluctuation @xmath176 as obtained from eq .
( [ cfn ] ) , and from the standard saddle - point approximations to the three individual terms on the right hand side of the identity ( [ cof ] ) . for each term
we have the same accidental correctness of the standard saddle - point result in the condensate regime , and of the condensate approximation at high temperatures , as already described for the occupation numbers of the excited states ; the inset , which depicts the ratios of the two approximations to the recursively calculated exact fluctuation , confirms that either approximation is correct at both high and low temperatures . in the case of the condensate fluctuation
, however , the situation is quite different . as witnessed by the inset in fig .
[ f_8 ] , our formula ( [ cfz ] ) gives the correct fluctuation of the number of ground - state particles in the condensate regime . since the canonical mean - square fluctuation @xmath177 has to vanish for zero temperature , when all @xmath5 particles occupy the ground state , the third term on the right hand side of eq .
( [ cfz ] ) approaches @xmath178 for @xmath179 .
when naively using the standard saddle - point scheme , the results for the individual terms in eq .
( [ cof ] ) are incorrect by factors @xmath180 , @xmath181 , and @xmath182 , respectively , with the universal numbers @xmath183 determined in appendix [ ab ] .
hence , the naive scheme yields spurious mean - square ground - state fluctuations @xmath184n^2 + [ r_1/r_2 - r_1/r_3]n \nonumber \\ & \approx & 0.02438 \ , n^2 + 0.01304 \ , n \label{err}\end{aligned}\ ] ] for @xmath179 , giving , for instance , the incorrect r.m.s .- fluctuation @xmath185 for @xmath125 , in precise agreement with what is observed in fig . [ f_8 ] . in the opposite regime , that is , for high temperatures , the standard scheme becomes correct . then , since @xmath186 , the condensate approximation ( [ cfz ] ) is off by roughly the same factor @xmath187 that also determines the error of @xmath154 itself .
it is now also illuminating to compare the saddle - point scheme developed here to the oscillator approximation that has led in ref .
@xcite to the integral representations ( [ icn ] ) ( [ idf ] ) .
this latter approximation can not cope with the bose
einstein transition , that is , its validity remains restricted to the condensate regime , since it rests on the fiction of an infinite reservoir of condensate particles @xcite .
this very feature , however , is what allows one to derive closed expressions for the condensate fluctuations , provided the pole structure of the spectral zeta function ( [ szf ] ) in the complex @xmath43-plane is known . for an ideal bose gas in a three - dimensional isotropic harmonic potential , this function can be written in terms of riemann zeta functions , @xmath188 \ ; , \label{ozf}\ ] ] having shifted the ground - state energy to @xmath39 . taking into account the three rightmost poles of the product @xmath189 , located at @xmath190 , @xmath191 , and @xmath192 , eq . ( [ icf ] ) then gives @xmath193 \nonumber \\ & + & \;\ ; \frac{k_bt}{\hbar\omega } \left(-\frac{1}{2}\right ) \label{oap}\end{aligned}\ ] ] for @xmath194 , where @xmath195 is euler s constant . in contrast
, the saddle - point approach requires the numerical determination of the three saddle - point parameters @xmath12 , @xmath131 , and @xmath164 from eqs .
( [ sa1 ] ) , ( [ nsp ] ) , and ( [ sa3 ] ) , respectively , but this effort is rewarded by the possibility to monitor the fluctuations for all temperatures , including the transition regime . as detailed in fig
. [ f_9 ] , sufficiently below the onset of condensation both the saddle - point approximation ( [ cfz ] ) and the oscillator approximation ( [ oap ] ) yield excellent agreement with the exact condensate fluctuation , even for particle numbers as low as @xmath125 .
for extending the techniques developed in the previous section to the microcanonical ensemble , we write the grand canonical partition function ( [ gpf ] ) as @xmath196 where the microcanonical partition functions @xmath197 denote the number of microstates accessible to a thermally isolated @xmath5-particle system with total excitation energy @xmath198 . setting the ground - state energy equal to zero
, we now assume that all single - particle energies @xmath0 , and hence also all possible excitation energies , can be represented reasonably well as integer multiples of a basic quantum @xmath199 . introducing the variable @xmath200 and writing @xmath201 , eq .
( [ gpm ] ) takes the form of a double power series , @xmath202 with @xmath203 indicating the degree of degeneracy of the single - particle energy level @xmath204 , and @xmath205 . with the help of two suitable contours which encircle the origin of the complex @xmath206- and @xmath6-plane , respectively , the desired partition functions can be isolated from this series by means of the identity @xmath207 where @xmath208 the saddle - point @xmath209 for the double contour integral ( [ dci ] ) now is determined by the simultaneous solution of the two equations @xmath210 reading explicitly @xmath211 as long as the gas is not condensed , we may safely use the standard approximation scheme @xcite .
that is , we may expand the function ( [ mff ] ) quadratically around the saddle - point , @xmath212 with @xmath213 substitute @xmath214 and @xmath215 , and get @xmath216 where @xmath217 is the functional determinant @xmath218 for temperatures below the onset of condensation , however , this procedure is invalid , because then the second of the saddle - point equations ( [ mse ] ) dictates that @xmath12 differs from the ground - state singularity @xmath219 of the grand canonical partition function ( [ dse ] ) merely by a quantity of order @xmath36 , exactly as in the canonical case .
hence , we have to proceed according to the insight accumulated there for computing @xmath220 in the condensate regime , and have to exempt the ground - state factor from the the quadratic expansion : defining @xmath221 we have to start from the representation @xmath222 the expansion of the tempered function @xmath223 around @xmath209 then yields @xmath224\!(z - z_0 ) + \frac{1}{2}f^{(0,2)}(z - z_0)^2 \ ; , \label{qua}\end{aligned}\ ] ] where we have used @xmath225 it is now of key interest to compare the magnitude of the two terms in the square brackets of eq .
( [ qua ] ) , which constitute the coefficient of @xmath226 . on the one hand , we have @xmath227 hence @xmath228 in the condensate regime . on the other hand ,
relevant contributions to the integral over @xmath206 are collected from an interval of order @xmath229 around @xmath230 .
thus , the relevant @xmath231 are on the order of @xmath232 , with @xmath233 if we focus again on systems of the type ( [ gro ] ) i.e. , if both the quantum @xmath199 and the weights @xmath203 have been adjusted accordingly , we may repeat the reasoning that has led to the canonical estimate ( [ sum ] ) , and conclude that both @xmath234 and @xmath235 are of the order @xmath236 .
therefore , @xmath237 for relevant @xmath206 . to be honest ,
this can be taken as a rather crude guideline only , in the same sense as the quantities @xmath75 displayed in fig .
[ f_1 ] have not yet approached the expected value @xmath76 for @xmath238 .
in fact , numerical calculations for the three - dimensional isotropic harmonic oscillator potential reveal that for reasonably large @xmath5 the expression @xmath232 is about proportional to @xmath239 , the square root of the total number of excited particles in a microcanonical setting , instead of being proportional to @xmath240 .
nonetheless , the above estimate indicates that for @xmath61 and large @xmath5 we may neglect @xmath231 against @xmath241 .
this implies a drastic simplification of the analysis , because then the remaining saddle - point integral factorizes : leading both contours parallel to the respective imaginary axis over the saddle , we are left with @xmath242 the first of these integrals is standard , the second is precisely of the type worked out in appendices [ aa ] and [ ab ] .
thus , without any further labor we obtain @xmath243 the saddle - point approximation to the microcanonical partition functions in the condensate regime .
the computation of microcanonical occupation numbers @xmath244 in the condensate regime is a matter of routine by now , so we merely need to sketch the main steps . denoting the number of microstates of an isolated @xmath5-particle system with total excitation energy @xmath245 and with exactly @xmath246 particles occupying a given single - particle state with energy @xmath247 as @xmath248 , we have @xmath249 and @xmath250 introducing the symbol @xmath251 , where the overbar is meant to indicate that the partial derivative acts on only _ one _ of the @xmath252-fold degenerate states with energy @xmath247 , the first microcanonical moments are generated from the grand canonical partition function ( [ dse ] ) by means of the identity @xmath253 hence , @xmath254 where @xmath255 the evaluation of the integral ( [ mon ] ) first requires the knowledge of its saddle - point @xmath256 , obtained by simultaneously solving the two equations @xmath257 if then @xmath116 , we define the tempered function @xmath258 yielding @xmath259 in direct analogy to eq .
( [ mpf ] ) , with @xmath260 if , however , @xmath141 , we have to account for ground - state doubling . in this case
we define @xmath261 and invoke eq .
( [ lta ] ) with @xmath144 , resulting in @xmath262 to give at least one application of these formulas , fig .
[ f_10 ] shows the microcanonical ground - state occupation number @xmath263 as a function of the microcanonical temperature for @xmath125 ideal bose particles in an isotropic harmonic oscillator potential , as computed from eqs .
( [ mpf ] ) and ( [ mfm ] ) according to eq .
( [ rat ] ) .
the microcanonical temperature does not differ significantly from the canonical one @xcite ; the inset quantifies the ratio of canonical to grand canonical , and of microcanonical to grand canonical occupation numbers .
as expected , the relative differences between the ground - state occupation numbers in the three ensembles are on the order of @xmath29 . for calculating the corresponding microcanonical mean - square fluctuations @xmath264 within the saddle - point approximation
, we exploit the identity @xmath265 which immediately leads to the analogue of eq .
( [ cof ] ) , namely @xmath266 the newly appearing integral is defined by @xmath267 giving @xmath268 hence , its saddle - point @xmath269 is found by simultaneously solving the two equations @xmath270 the usual distinction follows : if @xmath116 , extracting the ground - state contribution from the function ( [ mfh ] ) means introducing @xmath271 resulting in @xmath272 with @xmath273 if @xmath141 , we define instead @xmath274 and invoke eq .
( [ lta ] ) once more , now for @xmath174 , to arrive at @xmath275 collecting the results ( [ mpf ] ) , ( [ mfa ] ) , and ( [ msa ] ) for @xmath116 , or ( [ mpf ] ) , ( [ mfm ] ) , and ( [ msm ] ) for @xmath141 , eq .
( [ ana ] ) then allows one to determine the fluctuations .
an example for such a calculation is depicted in fig .
[ f_11 ] : the heavy solid line is the root - mean - square fluctuation @xmath276 as obtained from the above saddle - point scheme for a gas of @xmath277 ideal bosons in the usual isotropic oscillator potential . for comparison ,
when evaluating for the same system the formula ( [ idf ] ) for the difference @xmath278 up to terms of the order @xmath279 , one finds @xmath280 \nonumber \\ & + & \;\ ; \frac{k_bt}{\hbar\omega } \frac{\zeta(3)}{2\,\zeta(4 ) } \left[\ln\!\left(\frac{k_bt}{\hbar\omega}\right ) + \gamma + \frac{5}{24 } + \frac{27}{32 } \frac{\zeta^3(3)}{\zeta^2(4 ) } - \frac{5}{2 } \frac{\zeta(2)\zeta(3)}{\zeta(4 ) } + \frac{3}{2 } \frac{\zeta^2(2)}{\zeta(3)}\right ] \ ; . \label{dof}\end{aligned}\ ] ] this , together with eq .
( [ oap ] ) , yields a closed expression for the microcanonical condensate fluctuations of an ideal bose gas in a three - dimensional isotropic harmonic trap in the oscillator approximation , that is , under the assumption of an infinite reservoir of condensate particles : @xmath281 \left(\frac{k_bt}{\hbar\omega}\right)^2 \nonumber \\ & - & \left[0.55531 \ln\!\left(\frac{k_bt}{\hbar\omega}\right ) + 0.96969 \right ] \frac{k_bt}{\hbar\omega } \ ; . \label{mfl}\end{aligned}\ ] ] the dashed line in fig .
[ f_11 ] corresponds to the leading term of this approximation @xcite , whereas the thin line also takes the next - to - leading order into account . even for @xmath277 particles ,
finite - size effects are still visible in the condensate fluctuation ; the leading - order term of eq .
( [ mfl ] ) alone yields only modest agreement with the saddle - point result . after accounting for the dominant corrections ,
the agreement becomes close to perfect : even on the scale of the inset , the saddle - point result is indistinguishable from the oscillator approximation ( [ mfl ] ) .
the necessity to abandon the usual saddle - point scheme when exploring canonical or microcanonical statistics of condensed bose gases with @xmath5 particles is brought about by a characteristic dilemma . on the one hand ,
the approach of the saddle - point to the ground - state singularity at @xmath33 of the grand canonical partition function within order @xmath36 may be taken as the very hallmark of bose
einstein condensation ; on the other , the customary gaussian approximation requires that intervals of order @xmath36 around the saddle - point stay clear of singularities .
the solution to this problem almost suggests itself : if one exempts the ground - state factor of the grand canonical partition function from the gaussian expansion and treats that factor exactly , but proceeds as usual otherwise , then the singular point that now decides the fate of the approximation is the one produced by the first excited state at @xmath85 . since the saddle - point remains pinned below @xmath282 , it remains separated from the decisive singularity at @xmath85 by an @xmath5-independent gap .
this gap is wide enough to get the approximation going if the particle number is sufficiently large , because the required interval of regularity shrinks with increasing @xmath5 .
the representation ( [ cpf ] ) of the canonical partition functions can be viewed as the prototype integral expressing this strategy ; the other canonical and microcanonical quantities computed in this work constitute nothing but variations of the same mathematical theme .
the success of this amended saddle - point method hinges on the fact that the emerging integrals with singular integrands can be done exactly ; as explained in appendix [ aa ] , they lead directly to parabolic cylinder functions .
thus , we have accomplished the two goals set in the introduction : the results ( [ csa ] ) are easy to use , and provide bona fide approximations to partition functions , occupation numbers , and their fluctuations which are valid at all temperatures not only in the high - temperature limit or in the condensate regime , but also in the critical temperature range that witnesses the onset of condensation . in particular , the sharpness of the drop of the complementary error function contained in the canonical partition function ( [ cef ] ) allows one to precisely assess the sharpness of this onset in a bose gas with a finite , fixed number of particles .
the observation that the interval of regularity claimed by the gaussian expansion around the saddle - point is of the same order @xmath36 as the distance of the saddle - point from the ground - state singularity , meaning that the original conflict is not too large , reflects itself in the noteworthy fact that the error of the conventional scheme in the condensate regime is merely a temperature- and system - independent multiplicative constant .
the discovery , made in appendix [ ab ] , that this constant approaches unity when the multiplicity of the ground - state pole is increased fits in nicely : increasing that multiplicity amounts to considering a bose gas with a multiple - degenerate ground state and thus drives the saddle - point away from the ground - state singularity the occupation number of each individual of the degenerate states is lowered , thereby lessening the error of the naive approach .
it is also of interest to compare the workload implied by the proper saddle - point method to that required by other techniques aiming at canonical or microcanonical statistics .
exact recursion relations like eq .
( [ frf ] ) or their microcanonical analogues are invaluable for treating relatively small samples with not substantially more than about 1000 particles @xcite , but the computation of , e.g. , microcanonical condensate fluctuations by such means for a gas with @xmath277 bosons , as presented in fig .
[ f_11 ] , is entirely out of the question
. the integral representations ( [ icn ] ) ( [ idf ] ) , on the other hand , immediately yield analytical expressions for condensate occupation numbers and fluctuations , _ provided _ the pole structure of the spectral zeta function ( [ szf ] ) is known , but they do not allow one to monitor the system at the onset of condensation .
in contrast , saddle - point techniques always require one numerical step finding the saddle - point as the root of the respective saddle - point equation , but once this has been done , the formalism yields all quantities of interest , by means of the ever - same formulas , without further hardship .
therefore , we may conclude that despite all reservations @xcite piled up in more than half a century since schubert s incisive comments @xcite , it really is the saddle - point method which , if executed properly , provides the most powerful approach to the statistical mechanics of isolated , condensed ideal bose gases .
having an instrument that reliable and flexible at one s disposal is certainly not merely of mathematical value , but may also have some bearing on experiments with bose
einstein condensates of dilute atomic vapors which are now becoming routine .
these experiments are mainly done in isolated harmonic traps , in the regime of vapor densities where the atomic interactions , quantified by the @xmath283-wave scattering length @xmath284 , can be considered as weak : denoting the atom mass as @xmath285 and the characteristic trap frequency as @xmath74 , and defining the oscillator length @xmath286 which quantifies the extension of the trap s ground state , one has @xmath287 under typical conditions . for example , @xmath288 nm for @xmath289rb @xcite , giving @xmath290 for a sample of @xmath71 condensate atoms in a shallow trap with @xmath291 s@xmath292 , while @xmath293 .
these two relations place the system in the bogoliubov regime , traditionally associated with the notion of a weakly interacting bose gas .
however , it is feasible to prepare even more weakly interacting samples , either by tuning the scattering length with the help of external magnetic fields @xcite , or by working with spin - polarized atomic hydrogen @xcite , which features the unusually low triplet scattering length @xmath294 nm @xcite .
thus , @xmath295 in a harmonic trap with @xmath291 s@xmath292 , so that even for @xmath296 hydrogen atoms one finds @xmath297 , indicating the crossover regime from the ideal to the bogoliubov gas .
this crossover should manifest itself , in a non - trivial manner , in the behavior of the condensate fluctuations @xcite , which also determine what one may aptly term `` the minimum linewidth of an atom laser '' @xcite .
it would therefore be of substantial importance to study condensate fluctuations of _ very weakly interacting bose gases _ , that is , of systems intermediate between the ideal gas and the bogoliubov gas , and to probe whether the difference ( [ dof ] ) between the canonical and the microcanonical ensemble remains visible there ; in general , this difference should show a pronounced dependence on the trap type @xcite . in this way , an old , apparently purely academic issue the non - equivalence of statistical ensembles in the condensate regime
suddenly pops up at the forefront of topical research , in the theory of the atom laser .
seen from the experimental angle , such an enterprise is on the verge of becoming possible ; on the theoretical side , the first requirement is a tool for routinely computing ideal condensate fluctuations within the different ensembles , for traps with various geometries .
this tool is available now .
in section [ s2 ] we have met contour integrals of the form @xmath298 with positive integer @xmath299 , and a saddle - point lying too close to the singularity at @xmath33 for the standard approximation ( [ wsa ] ) to be viable . in this appendix
we derive the proper saddle - point approximation to these integrals , following a suggestion by dingle @xcite .
writing , in accordance with our previous notation , the negative logarithm of the full integrand as @xmath300 , @xmath301 the saddle - point @xmath302 is determined by the equation @xmath303 in the large-@xmath5-limit , this equation corresponds to the grand canonical relation between particle number @xmath5 and fugacity @xmath302 for a system with a @xmath299-fold degenerate ground state . substituting @xmath304 , so that the saddle - point is found at @xmath305 , and writing the difference between the singular point and the saddle - point as @xmath306 we have @xmath307 and @xmath308 defining @xmath309 , eqs .
( [ spe ] ) and ( [ fbu ] ) immediately yield @xmath310 moreover , we require @xmath311 . expanding @xmath312 up to second order around the saddle - point with a first derivative @xmath313 which does not , as in the conventional approximation ( [ wsa ] ) , vanish , but instead becomes _ large _ when @xmath314 is small , as in the condensate regime , then leading the path of integration over the saddle at @xmath315
, we obtain the approximation @xmath316 which , upon substituting @xmath317 becomes @xmath318 with @xmath319 the integral occurring here is closely related to whittaker s parabolic cylinder function @xmath320 , namely @xcite @xmath321 where @xmath322 runs from @xmath323 to @xmath324 , passing @xmath325 from above .
thus , the saddle - point approximation to the integrals ( [ isi ] ) finally takes the form @xmath326
the approximation ( [ csa ] ) is valid regardless whether or not the saddle - point @xmath302 lies close to @xmath282 , that is , for all temperatures .
therefore , it should adopt a more simple form in the condensate regime , where @xmath327 is of the order @xmath328 , and should merge into the standard saddle - point formula in the high - temperature limit , where @xmath302 approaches zero . in the case of high temperatures , the saddle - point moves away from the ground - state singularity ; the distance @xmath314 defined in eq .
( [ dos ] ) approaches unity , @xmath329 .
hence , both parameters @xmath98 and @xmath330 introduced in eqs .
( [ xnb ] ) and ( [ xwb ] ) adopt large positive values .
then the asymptotic expansion of the parabolic cylinder functions gives @xcite @xmath331 moreover , we have @xmath332 observing that , as a consequence of @xmath329 , @xmath333 we find @xmath334 this is just the expected result : for high temperatures , i.e. , when @xmath302 stays sufficiently far away from @xmath282 , we recover the formula provided by the usual approximation scheme ( [ wsa ] ) . in the opposite limit , that is , in the condensate regime
, we infer @xmath335 from eq .
( [ ord ] ) , whereas @xmath314 is of order @xmath328 .
since @xmath336 for systems of the type ( [ gro ] ) , and we have required @xmath337 , we find that @xmath314 goes to zero faster than @xmath338 increases when @xmath5 becomes large . hence ,
@xmath339 approaches zero for large @xmath5 .
this , in turn , implies that the argument @xmath99 of the parabolic cylinder functions now is a large negative number , so that @xcite @xmath340 since @xmath341 the low - temperature limit of the approximation ( [ csa ] ) becomes @xmath342 which has been used heavily in secs .
[ s2 ] and [ s3 ] .
it is of interest to recast this expression ( [ lta ] ) into a form which lends itself to a direct comparison with the now incorrect standard saddle - point formula . to this end
we exploit that , since @xmath343 , the second derivative of the full function @xmath344 at the saddle - point will be dominated by the singular ground - state contribution , giving @xmath345 with this additional approximation , and utilizing the identity @xmath346 , eq .
( [ lta ] ) yields @xmath347 this is a most intriguing observation : in the large-@xmath5 , low - temperature limit , i.e. , in the condensate regime , the result of the properly executed saddle - point approximation to the bose - type integral ( [ isi ] ) differs from the standard saddle - point formula by a _
temperature- and system - independent _ factor @xmath183 , namely @xmath348 some numerical values of @xmath183 are listed in the following table .
recalling stirling s formula for @xmath349 , one immediately realizes that these universal renormalization factors @xmath183 approach unity when the singularity index @xmath299 is increased .
this finding , which might appear paradoxical at first , has a simple explanation : in a system with a @xmath299-fold degenerate ground state , each individual of these states takes only @xmath350-th of the particles that a non - degenerate ground state would have to carry .
therefore , @xmath327 is of the order @xmath328 : the larger @xmath299 , the farther away is the saddle - point @xmath302 from the singularity at @xmath33 , and the better is the standard procedure . on the other hand , in the high - temperature regime , where @xmath329 , the left hand side of eq .
( [ lta ] ) can be written as @xmath351 thus , when using the low - temperature approximation ( [ lta ] ) in the high - temperature regime , the result is incorrect by the factor @xmath352 c. g. darwin and r. h. fowler , proc .
* 21 * , 391 , 730 ( 1923 ) .
e. schrdinger , _ statistical thermodynamics _ ( cambridge university press , cambridge , 1946 ) .
k. huang , _ statistical mechanics _ ( wiley , new york and london , 1963 ) .
r. k. pathria , _ statistical mechanics _ ( pergamon press , oxford , 1985 ) .
g. schubert , z. naturforsch . * 1 * , 113 ( 1946 ) ; _ ibidem _ * 2*a , 250 ( 1947 ) . i. fujiwara , d. ter haar , and h. wergeland , j. stat .
* 2 * , 329 ( 1970 ) .
r. m. ziff , g. e. uhlenbeck , and m. kac , phys .
32 * , 169 ( 1977 ) .
m. gajda and k. rzaewski , phys .
. lett . * 78 * , 2686 ( 1997 ) .
s. grossmann and m. holthaus , phys .
lett . * 79 * , 3557 ( 1997 ) .
r. b. dingle , _
asymptotic expansions : their derivation and interpretation _
( academic press , new york and london , 1973 ) ; chap .
r. b. dingle , proc .
cambridge phil .
* 45 * , 275 ( 1949 ) .
a. r. fraser , phil . mag . * 42 * , 165 ( 1951 ) .
p. navez , d. bitouk , m. gajda , z. idziaszek , and k. rzaewski , phys .
lett . * 79 * , 1789 ( 1997 ) .
m. fierz , helv .
acta * 29 * , 47 ( 1956 ) . m. holthaus , e. kalinowski , and k. kirsten , ann .
( n.y . ) * 270 * , 198 ( 1998 ) .
f. london , phys . rev . * 54 * , 947 ( 1938 ) .
s. r. de groot , g. j. hooyman , and c. a. ten seldam , proc .
london a * 203 * , 266 ( 1950 ) .
s. grossmann and m. holthaus , optics express * 1 * , 262 ( 1997 ) .
c. weiss and m. wilkens , optics express * 1 * , 272 ( 1997 ) .
e. t. whittaker and g. n. watson , _ a course of modern analysis _
( cambridge university press , cambridge , 1962 ) ; chap .
. i. m. ryshik and i. s. gradstein , _ tables of series , products , and integrals _ ( veb deutscher verlag der wissenschaften , berlin , 1963 ) .
m. abramowitz and i. a. stegun ( eds . ) , _ handbook of mathematical functions _
( dover , new york , 1972 ) .
p. t. landsberg , _ thermodynamics _ ( interscience publishers , new york , 1961 ) . p. borrmann and g. franke , j. chem . phys . * 98 * , 2484 ( 1993 ) .
f. brosens , j. t. devreese , and l. f. lemmens , solid state commun .
* 100 * , 123 ( 1996 ) .
n. l. balazs and t. bergeman , phys .
a * 58 * , 2359 ( 1998 ) .
f. c. auluck and d. s. kothari , proc .
. soc . * 42 * , 272 ( 1946 ) .
m. toda , r. kubo , and n. sait , _ statistical physics i _ ( springer verlag , berlin 1992 ) .
j. m. vogels , c. c. tsai , r. s. freeland , s. j. j. m. f. kokkelmans , b. j. verhaar , and d. j. heinzen , phys .
a * 56 * , r1067 ( 1997 ) .
s. inouye , m. r. andrews , j. stenger , h .- j .
miesner , d. m. stamper - kurn , and w. ketterle , nature * 392 * , 151 ( 1998 ) .
d. g. fried , t. c. killian , l. willmann , d. landhuis , s. c. moss , d. kleppner , and t. j. greytack , phys .
81 * , 3811 ( 1998 ) .
m. j. jamieson , a. dalgarno , and m. kimura , phys .
a * 51 * , 2626 ( 1995 ) .
s. giorgini , l. p. pitaevskii , and s. stringari , phys .
* 80 * , 5040 ( 1998 ) .
r. graham , phys .
lett . * 81 * , 5262 ( 1998 ) .
w. ketterle , _ atom laser _ , in _ mcgraw - hill 1999 yearbook of science & technology _ ( mcgraw - hill , new york , 1998 ) ; p. 43 . this integral representation can be obtained from eq .
( 19.5.4 ) in ref .
@xcite , observing that the parabolic cylinder function @xmath353 equals the whittaker function @xmath354 . | the application of the conventional saddle - point approximation to condensed bose gases is thwarted by the approach of the saddle - point to the ground - state singularity of the grand canonical partition function .
we develop and test a variant of the saddle - point method which takes proper care of this complication , and provides accurate , flexible , and computationally efficient access to both canonical and microcanonical statistics .
remarkably , the error committed when naively employing the conventional approximation in the condensate regime turns out to be universal , that is , independent of the system s single - particle spectrum .
the new scheme is able to cover all temperatures , including the critical temperature interval that marks the onset of bose
einstein condensation , and reveals in analytical detail how this onset leads to sharp features in gases with a fixed number of particles
. in particular , within the canonical ensemble the crossover from the high - temperature asymptotics to the condensate regime occurs in an error - function - like manner ; this error function reduces to a step function when the particle number becomes large .
our saddle - point formulas for occupation numbers and their fluctuations , verified by numerical calculations , clearly bring out the special role played by the ground state .
-20 mm |
particle - in - cell ( pic ) simulations can shed light on the physical mechanism of particle acceleration that occurs in the complicated dynamics within relativistic shocks .
recent pic simulations of relativistic electron - ion and electron - positron jets injected into an ambient plasma show that acceleration occurs within the downstream jet @xcite . in general
, these simulations have confirmed that relativistic jets excite the weibel instability , which generates current filaments and associated magnetic fields @xcite , and accelerates electrons @xcite .
therefore , the investigation of radiation resulting from accelerated particles ( mainly electrons and positrons ) in turbulent magnetic fields is essential for understanding radiation mechanisms and their observable spectral properties . in this report
we present a new numerical method to obtain spectra from particles self - consistently traced in our pic simulations . at @xmath0 .
1a shows jet electrons ( red ) , ambient electrons ( blue ) , and the total electron density ( black ) .
1b shows electric field energy ( red ) and magnetic field energy ( blue ) divided by the total kinetic energy . ]
we have performed simulations using a system with ( @xmath1 @xmath2 ( @xmath3 : grid size ) and a total of @xmath4 billion particles ( 12 particles@xmath5cell@xmath5species for the ambient plasma ) in the active grid zones @xcite . in the simulations the electron skin depth , @xmath6 , where @xmath7 is the electron plasma frequency and the electron debye length @xmath8 is half of the grid size . here
the computational domain is six times longer than in our previous simulations @xcite .
the electron number density of the jet is @xmath9 , where @xmath10 is the ambient electron density and @xmath11 .
the electron / positron thermal velocity of the jet is @xmath12 , where @xmath13 is the speed of light . ) and larger perpendicular velocity ( @xmath14 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons at the rest frame ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] ) and larger perpendicular velocity ( @xmath14 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons at the rest frame ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] ) and larger perpendicular velocity ( @xmath14 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons at the rest frame ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] figure 1 shows the averaged ( in the @xmath18 plane ) electron density and electromagnetic field energy along the jet at @xmath19 .
the resulting profiles of jet ( red ) , ambient ( blue ) , and total ( black ) electron density are shown in fig .
the ambient electrons are accelerated by the jet electrons and pile up towards the front part of jet . at the earlier time the ambient plasma density increases linearly behind the jet front .
at the later time the ambient plasma shows a rapid increase to a plateau behind the jet front , with additional increase to a higher plateau farther behind the jet front .
the jet density remains approximately constant except near the jet front .
plane at @xmath20 for @xmath21 the colors indicate the x - component of current density generated by the weibel instability , with the x- and z - components of magnetic field represented by arrows ( a ) .
phase space distributions as a function of @xmath22 plotted for the jet ( red ) and ambient ( blue ) electrons at the same time.,title="fig : " ] plane at @xmath20 for @xmath21 the colors indicate the x - component of current density generated by the weibel instability , with the x- and z - components of magnetic field represented by arrows ( a ) .
phase space distributions as a function of @xmath22 plotted for the jet ( red ) and ambient ( blue ) electrons at the same time.,title="fig : " ] the weibel instability remains excited by continuously injected jet particles and the electromagnetic fields are maintained at a high level , about four times that seen in a previous , much shorter grid simulation system ( @xmath23 ) . at the earlier simulation time a large electromagnetic structure is generated and accelerates the ambient plasma . as shown in fig .
1b , at the later simulation time the strong magnetic field extends up to @xmath24 .
these strong fields become very small beyond @xmath25 in the shocked ambient region @xcite .
plane at @xmath20 for @xmath26 the colors indicate the x - component of current density generated by the weibel instability , with the x- and z - components of magnetic field represented by arrows ( a ) .
phase space distributions as a function of @xmath22 plotted for the jet ( red ) and ambient ( blue ) electrons at the same time.,title="fig : " ] plane at @xmath20 for @xmath26 the colors indicate the x - component of current density generated by the weibel instability , with the x- and z - components of magnetic field represented by arrows ( a ) .
phase space distributions as a function of @xmath22 plotted for the jet ( red ) and ambient ( blue ) electrons at the same time.,title="fig : " ] the acceleration of ambient electrons becomes visible when jet electrons pass about @xmath27 .
the maximum density of accelerated ambient electrons is attained at @xmath28 .
the maximum density gradually reaches a plateau as seen in fig .
the maximum electromagnetic field energy is located at @xmath29 as shown in fig .
let a particle be at position @xmath30 at time @xmath31 @xcite .
at the same time , we observe the electric field from the particle from position @xmath32 .
however , because of the finite velocity of light , we observe the particle at an earlier position @xmath33 where it was at the retarded time @xmath34 . here
@xmath35 is the distance from the charge ( at the retarded time @xmath36 ) to the observer . after some calculation and simplifying assumptions the total energy @xmath37 radiated per unit solid angle per unit frequency from a charged particle moving with instantaneous velocity @xmath38 under acceleration @xmath39
can be expressed as @xcite the observer s viewing angle is set by the choice of @xmath42 ( @xmath43 ) .
the choice of unit vector @xmath42 along the direction of propagation of the jet ( hereafter taken to be the @xmath44-axis ) corresponds to head - on emission . for any other choice of @xmath42 ( e.g. , @xmath45 ) ,
off - axis emission is seen by the observer . in order to calculate radiation from relativistic jets propagating along the @xmath44 direction
@xcite we consider a test case which includes a parallel magnetic field ( @xmath46 ) , and jet velocity of @xmath47 .
two electrons are injected with different perpendicular velocities ( @xmath48 ) .
a maximum lorenz factor of @xmath49 is calculated with the larger perpendicular velocity .
figure 2 shows electron trajectories in the @xmath50 plane ( red : @xmath51 , blue : @xmath52 ) ( a : left panel ) , the radiation ( retarded ) electric field ( b : middle panel ) , and spectra ( right panel ) for the case @xmath53 .
the two electrons are propagating left to right with gyration in the @xmath54 plane ( not shown ) .
the gyroradius is about @xmath55 for the electron with the larger perpendicular velocity .
the seven curves show the power spectrum at viewing angles of 0@xmath56 ( red ) , 10@xmath56 ( orange ) , 20@xmath56 ( yellow ) , 30@xmath56 ( moss green ) , 45@xmath56 ( green ) , 70@xmath56 ( light blue ) , and 90@xmath56 ( blue ) .
the higher frequencies become stronger at the @xmath57 viewing angle .
the critical angle for off - axis radiation @xmath58 for this case is 4.25@xmath56 .
as shown in this panel , the spectrum at a larger viewing angle ( @xmath59 ) has smaller amplitude .
since the jet plasma has a large velocity @xmath44-component in the simulation frame , the radiation from the particles ( electrons and positrons ) is strongly beamed along the @xmath44-axis ( jitter radiation ) @xcite .
equations 6.30a and 6.30b show that the radiation with the viewing angle @xmath60 disappears ( see fig .
6.5 in the textbook of rybicki and lightman @xcite ) .
however , based on two other textbooks , radiation at the viewing angle @xmath61 should not vanish @xcite .
this aspect is shown in fig .
2c , and at the higher frequency the amplitude at the viewing angle @xmath57 is stronger than that with viewing angle @xmath61 . a synchrotron shock model is widely adopted to describe the radiation mechanism in the external shock thought to be responsible for observed broad - band grb afterglows @xcite . associated with this model
are three major assumptions that are adopted in almost all current grb afterglow models .
firstly , electrons are assumed to be `` fermi '' accelerated at the relativistic shocks and to have a power - law distribution with a power - law index @xmath62 upon acceleration , i.e. @xmath63 .
this is consistent with recent pic simulations of the shock formation and particle acceleration @xcite and also some monte carlo models @xcite , but see @xcite .
secondly , a fraction @xmath64 ( generally taken to be @xmath65 ) of the total electrons associated with ism baryons are accelerated , and the total electron energy is a fraction @xmath64 of the total internal energy in the shocked region .
thirdly , the strength of the magnetic fields in the shocked region is unknown , but its energy density ( @xmath66 ) is assumed to be a fraction @xmath67 of the internal energy .
these assumed `` micro - physics '' parameters , @xmath68 and @xmath69 , whose values are obtained from spectral fits @xcite reflect a lack of knowledge of the underlying microphysics @xcite . the typical observed emission frequency from an electron with ( comoving ) energy @xmath70 in a frame with a bulk lorentz factor
@xmath71 is @xmath72 .
three critical frequencies are defined by three characteristic electron energies .
these are @xmath73 ( the injection frequency ) , @xmath74 ( the cooling frequency ) , and @xmath75 ( the maximum synchrotron frequency ) . in our simulations of grb afterglows , there is one additional relevant frequency , @xmath76 , due to synchrotron self - absorption at lower frequencies @xcite .
the general agreement between the blast wave dynamics and the direct measurements of the fireball size argue for the validity of this model s dynamics @xcite .
the shock is most likely collisionless , i.e. mediated by plasma instabilities @xcite .
the electromagnetic instabilities mediating the afterglow shock are expected to generate magnetic fields .
afterglow radiation was therefore predicted to result from synchrotron emission of shock accelerated electrons @xcite .
the observed spectrum of afterglow radiation is indeed remarkably consistent with synchrotron emission of electrons accelerated to a power - law distribution , providing support for the standard afterglow model based on synchrotron emission of shock accelerated electrons @xcite . in order to determine the luminosity and spectrum of synchrotron radiation
, the strength of the magnetic field ( @xmath69 ) and the energy distribution of the electrons ( @xmath62 ) must be determined . due to the lack of a first principles theory of collisionless shocks , a purely phenomenological approach to the model of afterglow radiation was ascribed without investigating in detail the processes responsible for particle acceleration and magnetic field generation @xcite .
rather , one simply assumes that a fraction @xmath69 of the post - shock thermal energy density is carried by the magnetic field , that a fraction @xmath64 is carried by electrons , and that the energy distribution of the electrons is a power - law , @xmath77 ( above some minimum energy @xmath78 which is determined by @xmath64 and @xmath62 ) , @xmath79 , @xmath64 and @xmath62 are treated as free parameters , determined by observations .
it is important to clarify here that the constraints implied on these parameters by the observations are independent of any assumptions regarding the nature of the afterglow shock and the processes responsible for particle acceleration or magnetic field generation .
any model should satisfy these observational constraints .
the properties of synchrotron ( or jitter " ) emission from relativistic shocks will be determined by the magnetic field strength and structure and the electron energy distribution behind the shock .
the characteristics of jitter radiation may be important to understanding the complex time evolution and/or spectral structure in gamma - ray bursts @xcite .
for example , jitter radiation has been proposed as a means to explain grb spectra below the peak frequency that are harder than the `` line of death '' spectral index associated with synchrotron emission @xcite , i.e. , the observed spectral power scales as @xmath80 , whereas synchrotron spectra are @xmath81 or softer @xcite .
thus , it is essential to calculate radiation production by tracing electrons ( positrons ) in self - consistently treated small - scale electromagnetic fields . [
[ calculating - synchrotron - and - jitter - emission - from - electron - trajectories - in - self - consistently - generated - magnetic - fields ] ] calculating synchrotron and jitter emission from electron trajectories in self - consistently generated magnetic fields ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in order to obtain the spectrum of synchrotron ( jitter ) emission , we consider an ensemble of electrons selected in the region where the weibel instability has fully grown and electrons are accelerated in the generated magnetic fields . in order to validate our numerical method we performed simulations using a small system with ( @xmath82 ( @xmath3 : grid size ) and a total of @xmath83 billion particles ( 12 particles@xmath5cell@xmath5species for the ambient plasma ) in the active grid zones @xcite .
first we performed simulations without calculating radiation up to @xmath84 .
the jet front is located around about @xmath85 .
we selected 12,150 electrons for each jet and ambient electrons randomly .
recently , a similar calculation has been carried out for the radiation from accelerated electrons in laser - wakefield acceleration @xcite and in shocks @xcite . ) and small perpendicular velocity ( @xmath90 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] ) and small perpendicular velocity ( @xmath90 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] ) and small perpendicular velocity ( @xmath90 ) .
the paths of two electrons moving helically along the @xmath15direction in a homogenous magnetic field shown in the @xmath16-plane ( a ) .
the two electrons radiate a time dependent electric field .
an observer situated at great distance along the n - vector sees the retarded electric field from the moving electrons ( b ) . the observed power spectrum at different viewing angles from the two electrons ( c ) .
frequency is in units of @xmath17 .
, title="fig : " ] the spectra shown in fig .
5 are obtained for emission from jet electrons and ambient electrons separately . in this case
the spectra are calculated for head - on radiation ( @xmath61 ) and @xmath87 .
the radiation from jet electrons show bremsstrahlung - like spectra as a red line ( @xmath61 ) and orange line ( @xmath87 ) @xcite .
the spectra with jet electrons are different from the spectra shown in fig .
since the magnetic fields generated by the weibel instability are rather weak and the jet electrons are not much accelerated , the trajectories of jet electrons are almost straight with only a slight bent .
we compare these spectra with our known spectra obtained from two ( jet ) electrons , the case with a parallel magnetic field ( @xmath95 ) , and jet velocity of @xmath47 .
two electrons are injected with different perpendicular velocities ( @xmath96 ) .
a maximum lorenz factor of @xmath97 accompanies the larger perpendicular velocity .
the critical angle for off - axis radiation @xmath58 for this case is 8.05@xmath56 . comparing the spectra with figs . 5 and 6c we find similarities .
the lower frequencies have flat spectra and the higher frequencies decrease monotonically . the slope in fig .
5 is less steep than that in fig .
this is due to the fact that the spread of lorenz factors of jet electrons is larger and the average lorenz factor is larger as well .
furthermore , even the magnetic field strength is not so large , however the slope of the spectra seems to be consistent with fig . 7.16 ( left ) with the turbulent magnetic field with the red noise ( @xmath98 ) in hededal s ph .
d. thesis @xcite .
we obtained several different parameters with jet electrons and ambient magnetic field .
however , the strength of the magnetic fields generated by the weibel instability is small , therefore the spectra for these cases are very similar to fig .
5 . as shown in fig . 7.12 in hededal s ph .
d. thesis @xcite , the trajectories of jet electrons have to be chaotic to produce a jitter - like spectrum fig
. 7.22 @xcite . as shown in fig .
1b , the magnetic field energy in the region @xmath99 is small ( @xmath100 ) , therefore , as expected , the spectra look like those emitted from electrons propagating in a turbulent magnetic field with red noise ( see also figs . 3a and 4a ) . emission obtained with the method described above
is obtained self - consistently , and automatically accounts for magnetic field structures on small scales responsible for jitter emission . by performing such calculations for simulations with different parameters , we can investigate and compare the different regimes of jitter- and synchrotron - type emission @xcite .
the feasibility of this approach has already been demonstrated @xcite , and its implementation is straightforward .
thus , we should be able to address the low frequency grb spectral index violation of the synchrotron spectrum line of death @xcite .
medvedev and spitkovsky recently showed that electrons may cool efficiently at or near the shock jump and are capable of emitting a large fraction of the shock energy @xcite .
such shocks are well - resolved in existing pic simulations ; therefore , the microscopic structure can be studied in detail .
since most of the emission in such shocks would originate from the vicinity of the shock , the spectral power of the emitted radiation can be directly obtained from finite - length simulations and compared with observational data . as shown in fig .
1 , behind the trailing shock the electrons are accelerated and strong magnetic fields are generated .
therefore , this region seems to produce the emission that is observed by satellites .
we will calculate more spectra based on our rpic simulations and compare in detail with fermi data .
this work is supported by nsf - ast-0506719 , ast-0506666 , ast-0908040 , ast-0908010 , nasa - nng05gk73 g ,
nnx07aj88 g , nnx08ag83 g , nnx08al39 g , and nnx09ad16 g .
jn was supported by mnisw research projects 1 p03d 003 29 and n n203 393034 , and the foundation for polish science through the homing program , which is supported through the eea financial mechanism.simulations were performed at the columbia facility at the nasa advanced supercomputing ( nas ) . and ibm p690 ( copper ) at the national center for supercomputing applications ( ncsa ) which is supported by the nsf .
part of this work was done while k .-
n. was visiting the niels bohr institute .
support from the danish natural science research council is gratefully acknowledged .
this report was finalized during the program `` particle acceleration in astrophysical plasmas '' at the kavli institute for theoretical physics which is supported by the national science foundation under grant no .
phy05 - 51164 .
nishikawa , j. niemiec , h. sol , m. medvedev , et al . in proceedings of the 4th heidelberg international symposium on high energy gamma - ray astronomy , eds , f. a. aharonian , w. hofmann , f. rieger * 1085 * 589593 ( 2009 ) ( arxiv:0809.5067 ) | using our new 3-d relativistic electromagnetic particle ( remp ) code parallelized with mpi , we investigated long - term particle acceleration associated with a relativistic electron - positron jet propagating in an unmagnetized ambient electron - positron plasma .
the simulations were performed using a much longer simulation system than our previous simulations in order to investigate the full nonlinear stage of the weibel instability and its particle acceleration mechanism .
cold jet electrons are thermalized and ambient electrons are accelerated in the resulting shocks .
acceleration of ambient electrons leads to a maximum ambient electron density three times larger than the original value . behind the bow shock in the jet shock strong electromagnetic fields
are generated .
these fields may lead to time dependent afterglow emission .
we calculated radiation from electrons propagating in a uniform parallel magnetic field to verify the technique .
we also used the new technique to calculate emission from electrons based on simulations with a small system .
we obtained spectra which are consistent with those generated from electrons propagating in turbulent magnetic fields with red noise .
this turbulent magnetic field is similar to the magnetic field generated at an early nonlinear stage of the weibel instability .
a fully developed shock within a larger system generates a jitter / synchrotron spectrum . , , , , , , , , , , , , and acceleration of particles , galaxies , jets , gamma rays bursts , magnetic fields , plasmas , shock waves , radiation = 0.5 cm |
the classical isoperimetric problem in euclidean space @xmath1 looks for `` least perimeter '' sets among those ones with prescribed volume . a first solution of this problem was provided in 1935 by lusternik , by proving , using symmetrisation techniques , that balls minimise minkowski content ( see also burago - zalgaller ( * ? ? ?
* chapters 8 and 10 ) ) ; the complete solution was given by de giorgi @xcite , with the proof that balls are minimisers in the class of finite perimeter sets .
the isoperimetric problem has been extended to riemannian manifolds and , more recently , to the so - called _ manifolds with density_. a manifold with density ( see also ( * ? ? ?
* chapter 18 ) ) is a manifold @xmath2 endowed with a positive function @xmath3 used to weight both perimeter and volume .
more precisely , given a regular set @xmath4 , we define the ( weighted ) volume and perimeter of @xmath5 to be @xmath6 where @xmath7 and @xmath8 are volume and surface elements on @xmath2 ( see also section [ preliminaries ] for precise definitions ) .
note that such a density is not equivalent to scaling the metric of @xmath2 conformally by some factor @xmath9 , since in that case perimeter and volume would scale by different powers of @xmath9 .
one of the first and most important examples of manifolds with density , with applications to probability and statistics , is gauss space @xmath1 with density @xmath10 .
isoperimetric sets are half - spaces ( @xcite ; see also ( * ? ? ?
* chapter 18 ) or ( * ? ? ?
* section 3 ) ) .
more recently other authors ( e.g. @xcite ) have considered different examples of manifolds with density and generalised classical techniques : steiner and schwarz symmetrisations @xcite , functional versions of the isoperimetric inequality @xcite , first and second variation formulae , mean curvature , ricci curvature , and so on .
we refer the reader also to @xcite and references therein . in this paper we will consider two kinds of densities on the plane @xmath0 : continuous modifications of the gaussian density and piecewise constant densities , apparently the first examples of discontinuous ones .
these questions arose in the lectures on `` manifolds with density '' by f. morgan during the meeting `` gmtlap - geometric measure theory and least area problems '' in modena ( italy ) , 15 - 17th february 2007 , organized by g.p .
leonardi , r. monti , f. serra cassano , r. serapioni and i. tamanini .
the plan of the paper is the following .
section [ preliminaries ] contains introductory material on manifolds with density , perimeter , existence of isoperimetric sets , and first variation of the perimeter .
we stress proposition [ teofresnel ] , which provides a snell refraction law for isoperimetric boundaries for discontinuous , piecewise constant densities .
we comment on minimal surfaces in manifolds with density and introduce calibrations , which provide sufficient conditions for minimality . in section [ secpcdens ]
we investigate the isoperimetric problem in some cases of piecewise constant densities .
the first one is the half - space density , i.e. , the density on @xmath1 taking values 1 on the half - space @xmath11 and @xmath12 on @xmath13 , where @xmath9 is a function depending only on the last coordinate @xmath14 ; in this case , the isoperimetric sets are provided by round balls in the half - space @xmath15 ( theorem [ serapioni ] ) .
the second one is the `` strip '' density on @xmath1 taking values 1 if @xmath16 and a constant @xmath12 otherwise ; we show that isoperimetric sets exist and are connected for any given volume in the planar case , with some additional properties in general dimension .
a further analysis leads us to conjecture [ conj : strip ] , where we list the possible solutions in the planar setting .
this conjecture is almost completely proved in theorem [ teostrip ] .
the last density we consider is the `` ball '' density on @xmath0 with values @xmath17 outside the ball @xmath18 and constant @xmath19 inside : here isoperimetric sets exist for any given volume , are connected , and can assume several profiles , both in the case @xmath12 ( see theorem [ ballmagg ] ) and the case @xmath20 ( see theorem [ conj : ballmino1 ] ) .
section [ seccontdens ] concerns with modifications of gauss density on @xmath0 .
a question risen during the gmtlap meeting concerned the density @xmath21 on the half - plane @xmath22 .
we recently discovered that a paper of brock et al .
@xcite gives a complete answer to this problem , showing that sets bounded by vertical lines are isoperimetric .
we point out that , from that result , the same holds true also for the whole plane . during the gmtlap meeting
also the isoperimetric problem in the plane with density @xmath23 was risen : we conjecture that here the isoperimetric boundary is provided by vertical or horizontal lines . for certain volume bounds ( those close to half of the total volume ) horizontal lines are better than vertical ones , while for other bounds vertical lines beat horizontal ones ( see proposition [ propconj1 ] ) .
the last section [ secopenque ] collects several open questions : some from this paper , others from the gmtlap meeting .
a. c. is partially supported by the mcyt research project mtm2007 - 61919 ; d.v .
was partially supported by university of padova , gnampa of miur and gala project - geometric analysis in lie groups and applications , supported by the european commission .
the authors want to thank specially frank morgan , who inspired this work during the gmtlap meeting , read kindly first manuscripts and gave important ideas and improvements during the preparation of this paper .
we further want , among the others , to acknowledge : giovanni alberti , susanna ansaloni , nicola arcozzi , francesco bigolin , alessio brancolini , luca corbo esposito , thomas meinguet , laurent moonens , fabio paronetto , csar rosales , raul serapioni , francesco serra cassano , jeremy tyson .
our framework is an open subset @xmath24 ( we deal primarily with the cases of the whole space and the upper half - plane ) endowed with a lower semicontinuous , positive density function @xmath25 ; due to the positivity of @xmath26 , we shall sometimes write @xmath27 . for a given measurable set @xmath28 ,
we define ( according to @xcite ) its @xmath26volume by @xmath29 and its @xmath26perimeter measure by@xmath30 where the @xmath31 is taken among all smooth sets @xmath32 converging to @xmath5 in the @xmath33topology .
if no confusion may arise , we will just say volume and perimeter .
we say that @xmath5 has finite perimeter in @xmath34 if @xmath35 ; whenever the set @xmath34 is clear from the context , we shall simply write @xmath36 . the lower semicontinuity of @xmath26 ensures the equality @xmath37 for smooth sets .
we also notice that , by restricting to the open sets @xmath38 finite perimeter sets and finite euclidean perimeter sets in @xmath39 coincide for all @xmath40 . by well known results on finite perimeter sets ( see @xcite ) , there exists a subset @xmath41 of the topological boundary @xmath42 which is rectifiable ; that is , up to @xmath43negligible sets , is contained in a countable union of lipschitz graphs .
we point out that for a @xmath44rectifiable set @xmath45 the unit normal vector @xmath46 is defined @xmath43almost everywhere .
the set @xmath41 is called _ reduced boundary _ of @xmath5 ; for sake of simplicity , we will sometimes write @xmath45 instead of @xmath41 and , if not differently stated , @xmath46 will denote the inward unit normal vector .
mostly important , this part of the boundary is the essential part for the perimeter measure ; that is , the following representation formula holds : @xmath47 using the fact that @xmath48 , we also obtain that @xmath49 we shall use next result .
if the density @xmath26 satisfies @xmath50 on @xmath34 , then @xmath51 ; moreover , in the case @xmath52 , if @xmath5 is smooth and connected , we have @xmath53 since we allow discontinuities in @xmath26 , we introduce the following notation for @xmath54 . for fixed @xmath55 ,
let @xmath56 where @xmath57 is the open ball of center @xmath58 with radius @xmath59 ( balls , unless stated otherwise , are always assumed to be open ) .
moreover , we define the limits of @xmath26 in the positive and negative direction of the vector @xmath60 by @xmath61 the quantity @xmath62 does not change if we replace @xmath60 with @xmath63 , so we will assume , if not stated otherwise , that @xmath60 is chosen in such a way that @xmath64 . if @xmath65 is any nonzero vector , we denote by @xmath66 the previous quantities in the direction @xmath67 .
we collect in the following proposition a little list of known results in the isoperimetric problem we shall use and generalise . a very nice description of these properties is contained in @xcite .
we recall that a variation is given by a one parameter family of diffeomorphisms @xmath68 ; for such a given family , we define @xmath69 in a region @xmath34 with density @xmath26 , consider the isoperimetric problem of finding a closed `` isoperimetric set '' @xmath5 of least perimeter for prescribed volume @xmath70 . 1 . if @xmath71 , an isoperimetric set always exists
2 . if @xmath26 is lipschitz continuous , then @xmath5 is locally @xmath72 except for a singular set of hausdorff codimension @xmath73 in @xmath34 .
we sketch here the proofs ; we fix @xmath74 and a minimising sequence of sets @xmath32 with @xmath75 .
condition @xmath71 , by setting @xmath76 implies that the functions @xmath77 are equi
integrable in @xmath78 , so they are relatively compact . up to subsequences , @xmath77 converges to a function @xmath79 ; by the a.e .
convergence , @xmath80 has to be a characteristic function @xmath81 .
the fact that @xmath82 follows by continuity of the measure and @xmath83 follows by lower
semicontinuity of the perimeter measure , so @xmath5 is an isoperimetric set
. the proof of regularity can be found in morgan ( * ? ? ?
* proposition 3.5 , corollary 3.8 and remark 3.10 ) .
[ rmkmainproperties ] we collect here some necessary conditions for a set @xmath5 to be isoperimetric , under the assumption that @xmath26 is @xmath84 ( see rosales et al .
@xcite ) : 1 .
@xmath45 is stationary under volume preserving variations , i.e. @xmath85 whenever @xmath86 ; 2 .
there exists a constant @xmath87 such that for any variation @xmath88 3 .
the generalised mean curvature @xmath89 with @xmath90 the euclidean mean curvature , is constant .
in section [ secpiecconstgen ] we shall see how to generalise these conditions to the case of piecewise regular density .
we extend to manifolds with density the classical method of calibrations ( see e.g. definition 4.1 and theorem 4.2 of harvey - lawson @xcite or federer @xcite ) , giving sufficient conditions for a surface to be perimeter minimising .
we say that a set @xmath91 of finite perimeter is
_ perimeter minimising _
( without volume constraint ) in an open set @xmath92 if @xmath93 for any set @xmath94 such that @xmath95 [ calibrations ] suppose @xmath24 is an open set and @xmath91 has finite perimeter in @xmath34 .
moreover , suppose that there exists a @xmath96 function @xmath97 such that * @xmath98 for any @xmath99 ; * @xmath100-a.e . on @xmath41 ; * @xmath101 in the distributional sense .
then @xmath5 is perimeter minimising in @xmath34 .
let @xmath94 be another open set of finite perimeter in @xmath34 with @xmath102 , where without loss of generality we assume @xmath103 to have smooth boundary with finite surface measure .
since @xmath101 , by the divergence theorem @xmath104 consider @xmath0 with density @xmath105 ; then the curve @xmath106 is area minimising , in the sense that its epigraph is perimeter minimising .
in fact , a calibration is given by @xmath107 which extends the normal vector to the curve and satisfies the hypotheses of theorem [ calibrations ] .
notice that the singularity of @xmath80 in 0 does not affect the validity of our argument .
suppose @xmath108 is a hypersurface in @xmath1 ( with a @xmath96 regular density @xmath26 ) which coincides with the graph of a @xmath96 function @xmath109 , i.e. @xmath110 it is immediate to see that the surface measure of @xmath108 is given by @xmath111 for area minimising hypersurfaces the generalised mean curvature introduced in ( [ first_variation ] ) vanishes ; this means that @xmath112 satisfies the minimal surface equation @xmath113 we observe in passing that , in gauss space @xmath1 with density @xmath114 , half spaces through the origin have boundary ( seen as a graph ) satisfying ( they have vanishing mean curvature ) but are not perimeter minimising in @xmath1 , although they are isoperimetric for volume @xmath115 @xcite .
it can be proved in fact that there are no complete perimeter
minimising surfaces in gauss space .
suppose by contradiction that there exists a measurable set @xmath5 which is perimeter minimising in @xmath1 .
choose an open bounded set @xmath116 such that @xmath117 .
let us consider a sufficiently large @xmath118 such that @xmath119 and @xmath120 and set @xmath121 .
it is easily seen that @xmath122 contradicting the minimality of @xmath5 .
[ calgraphs ] suppose that @xmath26 is a regular density in @xmath123 which does not depend on the last coordinate @xmath14 , and suppose that @xmath124 satisfies the minimal surface equation .
then the graph of @xmath112 is perimeter minimising in @xmath125
. with our hypotheses on @xmath26 the minimal surface equation becomes @xmath126 where of course we interpret @xmath26 as a function of @xmath127 coordinates only .
then the normal to the graph at a point @xmath128 is given by @xmath129 and if we set @xmath130 by we obtain @xmath131 therefore the hypotheses of theorem [ calibrations ] are satisfied , and the graph of @xmath112 is perimeter minimising .
[ corhyperplanes ] suppose @xmath1 is endowed with a density independent on the last coordinate @xmath14 ; then all affine hyperplanes @xmath132 are perimeter minimising surfaces .
let us consider the plane @xmath0 endowed with the density @xmath133 introduced in @xcite , where it is proved ( see corollary 4.7 therein ) that there exist no isoperimetric regions , essentially because constant curvature curves have infinite perimeter .
as observed in corollary [ corhyperplanes ] , hyperplanes @xmath134 are perimeter minimising surfaces . using theorem [ calgraphs ] , one could check that for any @xmath135 the graphs of functions @xmath136 defined on proper half - lines @xmath137c,+\infty[$ ] , are perimeter minimising surfaces in @xmath137c,+\infty[\times\r$ ] . in the case of a piecewise regular density ,
variational formulae have to be slightly modified taking into account the jump set of the density @xmath26 . by piecewise regular density we mean a function @xmath138 with @xmath139 disjoint open lipschitz domains , @xmath140 , and @xmath141 on @xmath142 : we denote by @xmath143 the set @xmath144 using the notation introduced in ( [ applimit ] ) , for @xmath43-almost every @xmath145 we have @xmath146 for some @xmath147 , where @xmath148 is a unit normal vector to @xmath143 ; as usual , @xmath149 has the property that @xmath150 . for a given set @xmath28 with finite perimeter , the functions @xmath151 are the traces of @xmath5 on the two sides of @xmath143 .
these traces are well defined since the trace operator @xmath152 is continuous whenever @xmath153 is lipschitz ( see for instance ( * ? ? ?
* theorem 3.88 ) ) .
we can then state the following proposition .
[ propvarvolperpc ] let @xmath26 be a piecewise regular density and let @xmath5 be a set of finite perimeter , @xmath45 its reduced boundary .
let us assume that @xmath154 is a smooth one - parameter variation with @xmath155 .
set @xmath156 and @xmath157 , @xmath46 the inward unit normal vector to @xmath41 ; then the following first variation formula for the volume holds @xmath158 moreover , the function @xmath159 is differentiable at @xmath160 if for @xmath43-a.e .
@xmath161 there exists some @xmath162 such that @xmath163 for @xmath164 , where @xmath147 is such that @xmath165 ; this is the case if , for instance , @xmath166 wherever @xmath149 is defined . in this case @xmath167 for the proof of proposition [ propvarvolperpc ] , we shall make use of the following result ( * ? ? ?
* lemma 10.1 ) .
[ lemmagiusti ] let @xmath26 be any measurable lower semicontinuous density , and let @xmath168 be a diffeormorphism .
if @xmath5 and @xmath169 have finite perimeter , then the following relation holds @xmath170 where we have defined @xmath171 we can then prove proposition [ propvarvolperpc ] .
the proof of this fact is essentially the same contained in @xcite ; in fact , we have that @xmath172 then , using the almost everywhere differentiability of @xmath26 , @xmath173 notice that @xmath174 has finite perimeter and @xmath175 is made by two pieces : @xmath176 and the part of @xmath177 where @xmath5 has trace @xmath17 , that is @xmath178 ( here @xmath179 is the inward unit normal to @xmath180 ) .
therefore , the divergence theorem gives @xmath181 summing up on @xmath182 , formula ( [ firstvarvolpc ] ) follows . in order to prove ( [ firstvarper ] ) , we use ( [ changeofvar ] ) with @xmath183 and @xmath184 ; we then have @xmath185 and then @xmath186 for the second term on the right hand side , we have as in the standard case that @xmath187 with @xmath188 the tangential divergence ; since @xmath189 , we have only to compute @xmath190 we analyse separately the two cases : @xmath191 for some @xmath147 , and @xmath145 . for @xmath191 , we have that @xmath163 for small @xmath192 , and then @xmath193 whence @xmath194 for @xmath145 , we have that , since @xmath26 is lower semicontinuous , @xmath195 ( with @xmath149 unit normal vector of @xmath143 ) ; since we are assuming ( [ posscal ] ) , @xmath196 and this completes the proof . having given first variation formulae
, we discuss the stability of isoperimetric sets . in a discontinuous density setting
the latter has to be understood as follows : if the variation @xmath197 is volume preserving , @xmath198 for all @xmath199 , then perimeter has to increase , i.e. @xmath200 .
this means that if @xmath201 is differentiable at @xmath202 then @xmath203 .
this condition can be strengthened to @xmath85 when @xmath204 : in fact , in this case we have differentiability of @xmath205 also when considering the vector field @xmath206 instead of @xmath65 . as a consequence
we will obtain necessary conditions on the isoperimetric set at points far from @xmath143 and at points on @xmath143 , as summarised in the following propositions [ teofresnel ] and [ cortangential ] .
if we assume some regularity on the isoperimetric set , then the first variation of the perimeter can be rewritten in a different way . far from @xmath143 , standard regularity holds @xcite .
the additional regularity we have to require is the @xmath43almost everywhere @xmath207 regularity of @xmath143 and the regularity of traces of @xmath176 on @xmath143 , allowing an integration by parts in ( [ firstvarper ] ) in order to eliminate the term @xmath208 .
this operation can be done if we assume the following : ( r1 ) : : for every @xmath182 the part of the boundary @xmath209 is an @xmath44-rectifiable set with an @xmath210-rectifiable boundary @xmath211 of finite @xmath212 measure , i.e. @xmath213 is a normal @xmath44current : for @xmath214 , we denote by @xmath215 the unit normal vector orthogonal to @xmath216 , contained in @xmath217 and with @xmath218 pointing outside @xmath213 ; ( r2 ) : : @xmath219 , which is also an @xmath44rectifiable set , has @xmath210rectifiable boundary @xmath220 with finite @xmath212 measure , i.e. @xmath221 is a normal @xmath44current : we denote by @xmath222 the unit normal vector orthogonal to @xmath223 , contained in @xmath224 and pointing outside @xmath221 . in this case
the first variation of the perimeter can be rewritten as @xmath225 for the densities we shall consider in section [ secpcdens ] , * ( r1 ) * and * ( r2 ) * are satisfied , so we shall not enter into further details and simply say , if they hold true , that @xmath5 has _ regular trace _ on @xmath143 .
we now give some necessary condition for the isoperimetric set ; first of all , as already mentioned in remark [ rmkmainproperties ] , the generalised mean curvature @xmath226 is equal to a constant @xmath87 for any @xmath227 .
moreover , we have the next result , giving a sort of counterpart of the celebrated snell refraction law of optics .
it follows from a well known fact , that the tangent cone to an isoperimetric set must be perimeter minimising without volume constraint (
see for instance
* section 9 ) ) .
anyway , we give the proof for reader s convenience .
[ teofresnel ] in the same notation as proposition [ propvarvolperpc ] , let @xmath5 be an isoperimetric set with regular traces on @xmath143 ; assume that @xmath45 passes through @xmath143 transversally at a point @xmath228 where @xmath229 and @xmath143 have tangent spaces , i.e. the vectors @xmath218 , @xmath230 and @xmath149 exist .
then the vector @xmath231 is parallel to @xmath232 , that is @xmath233 for any @xmath234 .
in particular the snell refraction law @xmath235 holds . here , @xmath236 and @xmath237 are , respectively , the two angles at @xmath58 between @xmath143 and @xmath45 from the two opposite sides of @xmath143 ( see figure [ fresnel ] ) .
first of all , we notice that @xmath238 , since otherwise @xmath239 for any @xmath147 and then @xmath240 . let @xmath241 be as in the statement ; consider a second point @xmath242 and a sequence of functions @xmath243 , @xmath244 such that @xmath245 and then consider the vector field @xmath246 for some unit vectors @xmath247 and @xmath248 to be chosen in a convenient way .
the constant @xmath249 has to be determined by requiring that @xmath86 ; using and taking into account the continuity of the functions @xmath250 in @xmath251 , as @xmath252 we get @xmath253 multiplying the previous equation by @xmath254 and taking the limit as @xmath255 , we obtain @xmath256 we are assuming that @xmath143 has a tangent hyperplane at @xmath257 , and then there exists a normal vector @xmath258 with the usual agreement that @xmath259 and @xmath260 .
recalling proposition [ propvarvolperpc ] , we have an admissible variation for any @xmath247 with the condition @xmath261 .
the validity of ( [ blowtang ] ) for any such @xmath247 implies then that the vector @xmath262 is a positive multiple of @xmath149 and @xmath263 for each @xmath264 ; this is the snell law , since by transversality of @xmath45 at @xmath58 , we can take @xmath247 as the unique vector ( up to a sign ) orthogonal to @xmath149 and to @xmath265 .
proposition [ teofresnel ] will be of crucial importance in section [ secpcdens ] , where we will consider piecewise constant densities .
intuitively , proposition [ teofresnel ] says that , when @xmath42 crosses @xmath143 transversally , a corner is formed according to snell law .
notice also that the angle @xmath236 must be greater than @xmath266 and less than the supplementary of this angle .
we will sometimes use in an equivalent form , where the cosines of the angles @xmath267 are substituted by the sines of the complementary angles , i.e. the ones formed by the normals to @xmath143 and @xmath45 .
when the crossing between @xmath45 and @xmath143 is not transversal and @xmath45 touches @xmath143 in the region @xmath223 defined in * ( r2 ) * , we have the following result .
[ cortangential ] let @xmath5 be an isoperimetric set with regular trace on @xmath143 ; assume that for @xmath268 , @xmath222 , @xmath218 and @xmath149 exist , i.e. @xmath269 and @xmath143 have tangent planes .
if @xmath270 , then @xmath271 that is , @xmath272 meets @xmath143 at an angle at most @xmath273 ; if @xmath274 , then @xmath275 , that is , @xmath213 is tangential to @xmath143 .
the proof is similar as for proposition [ teofresnel ] ; setting again @xmath276 and considering @xmath277 with @xmath243 , @xmath244 , as before , by taking the limits as @xmath278 and @xmath279 one gets @xmath280 this inequality holds for any @xmath247 with @xmath281 ; if @xmath282 , then by taking @xmath283 we obtain ( [ snelltan ] ) . otherwise , if @xmath284 , then ( [ nunu ] ) reduces to @xmath285 for any @xmath286 .
since @xmath287 , the only possibility is that @xmath288 .
the previous proofs can be summarised and generalised as follows .
let us assume that the surface @xmath143 admits a unique tangent cone at the point @xmath289 and let us denote by @xmath290 the blow - up at @xmath289 of the region @xmath180 .
if the isoperimetric set passes through @xmath143 at @xmath289 , then there holds @xmath291 for any direction @xmath60 contained in @xmath290 where @xmath250 is minimal among the @xmath292 s such that @xmath293 .
this means that if @xmath143 has a tangent plane at @xmath289 , proposition [ teofresnel ] holds since @xmath60 can range over a half - space ; moreover , @xmath290 can not contain more than a half - space , otherwise we would have @xmath294 which is never possible if @xmath295 .
finally , if @xmath290 is less then a half - space , then the conclusion is that @xmath296 is contained in a sector .
in this section we consider some particular piecewise regular densities on euclidean space , mainly piecewise constant densities .
section [ sechalfspace ] focuses on a density in @xmath1 taking value one in a half - space , and defined in the other half - space by a certain real function with values greater than one . in section [ secstrip ]
we study the case of density one in a planar strip and a greater constant outside the strip .
finally , in section [ secballs ] we deal with a planar density taking value one outside a ball and a different constant value inside . in this section
we will focus our attention on the half - space density in @xmath1 defined by @xmath297 where @xmath298 is a bounded measurable function with @xmath299 for @xmath300 .
observe that the weighted volume and perimeter coincide with the euclidean ones in @xmath301 .
we give the following preliminary result .
[ constrlemma ] let @xmath5 be a bounded set ; then there exists a set @xmath302 entirely contained in @xmath303 with @xmath304 and @xmath305 , with equality if and only if almost all of @xmath5 is contained in @xmath301 .
given a bounded set @xmath28 , we write @xmath306 and @xmath307 ; we claim that if @xmath308 , then there exists a set in @xmath303 with the same volume and strictly less perimeter than @xmath5 .
the volume of @xmath5 is given by @xmath309 we define the transformation @xmath168 by @xmath310 , with @xmath311 for this transformation we have that @xmath312 is the identity @xmath313 on @xmath314 if @xmath315 , while for @xmath316 we have @xmath317 this means that the set @xmath318 has the lebesgue measure @xmath319 , by lemma [ lemmagiusti ] @xmath320 with equality if for @xmath43a.e . @xmath321 we have @xmath322 .
this can not happen unless either @xmath323 or @xmath324 is a vertical cylinder ; but in this second case , @xmath5 would not be bounded .
since @xmath5 is bounded , it is contained in a strip @xmath325 for some positive @xmath326 . by translating down by @xmath254 in the @xmath327th direction the set @xmath169
, we can construct a set contained in @xmath303 , with the same volume ( in view of ) as @xmath5 and less or equal perimeter ( [ fper ] ) , which proves the claim . [ serapioni ] for given volume , an isoperimetric set exists and is a round ball contained in the half space @xmath301 .
we may consider a minimising sequence @xmath328 of smooth sets , that is @xmath328 with @xmath329 and @xmath330 where @xmath331 if @xmath328 is bounded , we set @xmath332 ; otherwise , if @xmath328 is not bounded , there exists @xmath199 such that @xmath333 and @xmath334 where we used the fact that the first two equalities hold for a.e .
@xmath199 . by the euclidean isoperimetric inequality
, we have that @xmath335 we can then consider a ball @xmath336 with @xmath254 chosen in such a way that @xmath337 and @xmath289 such that @xmath336 is contained in @xmath303 and not intersecting @xmath338
. notice that @xmath339 and so the set @xmath340 has the properties @xmath75 and @xmath341 that is @xmath32 is also a minimising sequence with bounded sets . by lemma [ constrlemma ]
, we may assume that @xmath32 is contained in @xmath303 , where the density @xmath342 . by the standard isoperimetric inequality , round balls are isoperimetric . by a second application of lemma [ constrlemma ] ,
round balls in @xmath301 are uniquely isoperimetric . when @xmath343 does not depend only on the last coordinate , then it is no longer true in general that round balls in @xmath301 are isoperimetric .
for instance , for the density @xmath344 with @xmath345 , @xmath346 is isoperimetric .
we shall discuss in section [ secballs ] the case of the density ( [ denball ] ) in the plane @xmath0 . in this section we shall consider the `` strip '' density in @xmath347 given by @xmath348 with @xmath349 .
for the density @xmath350 isoperimetric sets exist and are round balls outside the slab .
indeed , lemma [ constrlemma ] shows that any set @xmath5 may be replaced by a set outside the slab with no more perimeter , with equality if and only if the boundary of @xmath5 is vertical inside the slab .
but if this fact occurs , the original set @xmath5 will be an unbounded vertical cylinder or will not have constant curvature , and so can not be isoperimetric .
this allows us to consider only sets contained in @xmath351 .
we start with an existence result for the planar case .
[ propexistn2 ] for the above strip density in the case @xmath52 , an isoperimetric set exists for any prescribed volume and is connected .
take a minimising sequence @xmath32 of smooth sets with @xmath75 and each @xmath32 consisting of finitely many components . by lemma [ constrlemma ]
we may assume that any connected component of @xmath32 either has non empty intersection with the lines @xmath352 and @xmath353 or is contained in @xmath354 .
move horizontally ( or also vertically , but remaining inside the strip , in case of a ball contained in it ) the connected components of @xmath32 until they touch each other tangentially ; using a cutting argument , replace @xmath32 with a connected one .
move horizontally the sets @xmath32 to have barycenter @xmath355 ; the sets @xmath32 are connected and so , using the diameter estimate ( [ estdiamper ] ) , they satisfy @xmath356 .
this means that all the sets @xmath32 are contained in a big ball , and a compactness argument implies the existence of an isoperimetric region : were it not connected , we could use one of the above `` components translation '' argument to diminish perimeter .
we can also state the following result that holds in any dimension @xmath327 , under the assumption that the isoperimetric set exists .
[ wsteiner ] let @xmath5 be an isoperimetric set for given volume @xmath357 , for the above strip density .
then , there exists an isoperimetric set for volume @xmath70 which is rotationally symmetric with respect to the @xmath14-axis .
moreover , the intersection of @xmath5 with almost every horizontal hyperplane is a round ball , @xmath5 has regular trace on @xmath358 and @xmath5 is either entirely contained in @xmath359 or it will touch simultaneously both sets @xmath360 and @xmath361 . the argument is based on schwarz symmetrisation ( see @xcite ) , which can be applied in this setting ; define @xmath362 with @xmath363 and @xmath364 we also define the sets @xmath365 notice that @xmath366 we can now consider the schwarz - symmetrised set of @xmath5 , i.e. the set @xmath367 such that @xmath368 is a round ball centered at @xmath369 with measure equal to @xmath370 .
one has latexmath:[\[\label{symmetrized } of @xmath5 is a round ball in case @xmath372 ( see @xcite or @xcite ) .
we notice that ( [ symmetrized ] ) holds also on the sets @xmath373 , @xmath374 . as @xmath375 where @xmath376 and @xmath377 are defined as in ( [ s ] ) with @xmath367 instead of @xmath5 ,
we can conclude , since @xmath26 is constant along horizontal directions , that latexmath:[\[\label{eq : schw } @xmath5 is isoperimetric , we have the equality in , and so @xmath367 is an isoperimetric solution for volume @xmath70 , which is in fact rotationally symmetric with respect to the @xmath14-axis .
moreover , we also deduce that the intersection of @xmath5 with almost every horizontal hyperplane is a round ball .
the regularity of the trace of @xmath5 on @xmath358 follows since the intersections of the isoperimetric set with the planes @xmath379 ( similar argument on the jump part @xmath380 ) are round balls converging in @xmath381 as @xmath382 ( see for instance giusti ( * ? ? ? * theorem 2.11 ) ) ; therefore the traces on the jump set are round balls . for the last sentence ,
if @xmath383 and @xmath384 , we can move down @xmath5 using lemma [ constrlemma ] , and modify @xmath5 into a set @xmath385 with same volume and less perimeter .
notice that proposition [ wsteiner ] implies that the isoperimetric problem for the `` strip '' density is essentially two - dimensional , meaning that the boundary of the isoperimetric set is given by the rotation on the vertical axis of the graph of a generating function @xmath386 .
let us stress that this constitutes a different problem with respect to the `` strip '' case with @xmath52 , since now the isoperimetric problem to be considered has different densities on volume and perimeter , more precisely @xmath387 we will now focus on the case @xmath52 .
in this situation , we know that isoperimetric sets exist , are connected and that almost all horizontal slices are intervals ( from propositions [ propexistn2 ] and [ wsteiner ] ) . furthermore , snell law and constant geodesic curvature condition are satisfied .
therefore , we have that the boundary of any isoperimetric set @xmath5 is a piecewise regular curve @xmath388 with regular pieces consisting of either arcs @xmath389 of some circles with the same curvature or line segments contained in @xmath390
. moreover , the part of @xmath388 contained in @xmath391 is at most one arc of a circle ; the same holds in @xmath392 , while there can be two arcs inside the strip .
this allows to classify the isoperimetric candidates which are rotationally symmetric in this case .
[ lemstripposs ] for the strip density for @xmath52 the only isoperimetric candidates with vertical reflective symmetry ( see figure [ strip1 ] ) are , up to density - preserving isometries : * balls contained in the strip ; * pieces of the strip bounded left and right by two semicircles , and top and bottom by two segments on @xmath390 ; * sets bounded by three circular arcs with the same radius and a segment contained in @xmath353 .
one of these arcs is contained in @xmath393 , and the other two are contained in the strip , meeting the segment tangentially and the other arc according to the snell law ( [ eqfresnel ] ) .
the segment can degenerate to a single point ; in this case , the set is a circle with radius 2 and center on the line @xmath352 ; * sets bounded by four circular arcs with the same radius : two of them contained in the strip ; the other two horizontally symmetric , one contained in @xmath393 and another contained in @xmath394 , satisfying the snell law .
let us briefly sketch how to restrict to such candidates .
if an isoperimetric set is completely contained in the strip it is easy to reduce to candidates ( i ) and ( ii ) .
otherwise , by symmetry we can suppose there is an arc in @xmath393 meeting the strip ; lemma [ constrlemma ] allows us to discard balls outside the strip . unless the contact angle is not @xmath395 , the arc is continued ( starting from each of its endpoints according to snell law ) by two arcs inside the strip : they have to meet the line @xmath353 , or we could use construction of lemma [ constrlemma ] . if the meeting is tangential there are two possibilities : the boundary can be continued by a segment and then a tangential arc in the strip ( case ( iii ) ) or immediately by an arc in the strip .
in this last case there would be an arc inside the strip tangent to @xmath353 plus another arc in @xmath393 : they would have the same curvature and two common endpoints , so they must form a unique complete circle or be symmetric with respect to the line @xmath352 .
this contradicts the snell law unless the meeting angle with @xmath352 is @xmath396 , i.e. , the set is a ball of radius 2 .
there is a priori another possibility ( see figure [ figcasev ] ) : the meeting with @xmath353 is tangential and the boundary is continued by a segment and then an arc in @xmath394 with angle @xmath397 .
we will exclude this possibility in step 1 below .
if the meeting with @xmath353 is not tangential there must be an arc in @xmath394 : this is case ( iv ) .
step 2 below will prove that this set must be also symmetric with respect to the horizontal axis @xmath398 .
if the upper arc meets @xmath352 with angle @xmath399 there are two possibilities for each of its endpoints : it is continued by a segment and then by a tangential arc , or immediately by a tangential arc .
again , these arcs must meet the line @xmath353 : if the meeting is not tangential there should be an arc in @xmath394 .
this configuration is symmetric to the one of figure [ figcasev ] , which we are going to exclude in step 1 .
if the meeting is tangential , the arc s radius must be 1 and one must have also a segment on @xmath353 and , possibly , another arc on @xmath394 with meeting angle @xmath399 as in figure [ fig1or2tangarcs ] : this set can not be isoperimetric as shown in step 3 .
a little problem could be given by the upper arc meeting @xmath352 with angle @xmath400 ( see figure [ figline ] in theorem [ theoline ] ) : the continuation would be a segment on @xmath352 ( no inner arc is allowed by snell law ) .
this set would be entirely contained in @xmath401 , thus not being isoperimetric .
_ we will show that the configuration of figure [ figcasev ] is geometrically impossible .
this supposed configuration consists of two segments in @xmath353 ( each of them possibly reducing to a single point ) and four circular arcs with the same radius ; two of them vertically symmetric , contained in the strip and tangent to @xmath353 but not to @xmath352 ; the other two contained in each component of @xmath402 , all satisfying the snell law . in particular , the arc in @xmath394 meets the strip with angle @xmath403 .
we can think of this configuration as obtained from a configuration ( iii ) by adding the arc in @xmath394 ; we will show that this arc is too long to fit .
the chord associated with this arc rests upon the segment on @xmath353 which appears in ( iii ) .
this chord has length @xmath404 , @xmath405 being the common radius of the arcs .
we will get a contradiction if we show that the segment of configuration ( iii ) is shorter . indeed , its length is equal to @xmath406 where we put @xmath407 ( since @xmath408 ) and used the inequality @xmath409 .
the last term in is strictly less than @xmath404 unless @xmath410 , i.e. @xmath411 . the latter case @xmath411 will be discarded in the following step 3 .
_ here we show that the four arcs configuration of type ( iv ) must be symmetric with respect to the horizontal line @xmath398 . as in figure
[ fig4nonsymm ] , let @xmath201 be the center of the arc in @xmath393 , and let @xmath412 be its endpoints on the line @xmath352 .
similarly , @xmath413 is the center of the lower arc and @xmath414 its endpoints . by @xmath415
we denote the center of the left arc inside the strip ; @xmath416 belongs to this arc and is chosen so that @xmath417 is horizontal
. it will be convenient to set @xmath418 and @xmath419 is the common radius of the arcs .
notice that @xmath420 and @xmath421 uniquely determine the configuration ; indeed , since the height of the strip is 2 , @xmath422 moreover , @xmath423 are also the angles formed by the vertical line passing through @xmath424 and the two arcs meeting at @xmath424 ; similarly for @xmath425 at @xmath426 .
thus the snell law @xmath427 holds .
when @xmath421 is negative ( i.e. @xmath415 lies above the line @xmath352 ) computations are the same .
recall that the configuration must be symmetric with respect to a vertical line , which must contain @xmath201 and @xmath413 .
equivalently , @xmath201 and @xmath413 must belong to such axis : they lie `` one above the other '' .
equivalently , the first component of the oriented segment @xmath428 is equal to the one of @xmath429 , i.e. @xmath430 .
taking into account and , we can exploit this condition to get @xmath431 where we have set @xmath432 .
let us study the function @xmath433\to\r$ ] defined by @xmath434 .
it is an even map and @xmath435 is negative for @xmath436 and positive for @xmath437 .
thus @xmath438 is even , increasing in @xmath439 and decreasing in @xmath1370,1]$ ] . in order to have @xmath440 it necessarily holds that @xmath441 , whence @xmath442 .
this condition corresponds to @xmath443 , which immediately implies symmetry with respect to @xmath398 .
_ we have to discard sets @xmath5 as in figure [ fig1or2tangarcs ] bounded by : one arc in @xmath393 and/or one in @xmath394 meeting the strip with angle @xmath403 ; one or two segments on @xmath352 and one / two on @xmath353 ( any of them can be a single point ) ; two arcs in the strip meeting tangentially the segments .
notice that the radius of any of the four arcs is 1 , since the inner ones are tangent to @xmath390 .
set @xmath444 .
cut from @xmath5 the region enclosed by one of the arcs outside the strip . in this way
we lose area @xmath445 .
since the arc disappears , but a segment is created , we lose perimeter @xmath446 . we restore area in the following way : `` enlarge '' @xmath5 horizontally until area is restored , for example by translating the right arc inside the strip on the right . two segments ( one on @xmath352 and one on @xmath353 )
are created ; their length @xmath447 must satisfy the restoring - area condition @xmath448 this operation produces a growth of perimeter equal to @xmath449 . to show that we have decreased perimeter
we must check that @xmath450 or equivalently that @xmath451 which is always satisfied .
next result assures that any isoperimetric set in this setting must be vertical reflective symmetric .
[ prop : simetrico ] let @xmath5 be an isoperimetric region for the strip density . then , @xmath5 has vertical reflective symmetry .
consider the symmetrised set @xmath367 of @xmath5 , which will be of one of the types described in lemma [ lemstripposs ] .
if @xmath367 is of type ( i ) , then @xmath5 is entirely contained in the strip , and it follows trivially that it must be a ball , thus symmetric . if @xmath367 is of type ( ii ) , then the part of @xmath42 contained in @xmath390 consists of a segment in @xmath352 and another one in @xmath353 , both of equal length .
the upper segment must be continued left and right by two tangential arcs of the same radius ( since curvature is constant and snell law holds ) .
this already gives the symmetry of @xmath5 ( we are also using that horizontal slices of @xmath5 are intervals ) .
finally , if @xmath367 is of type ( iii ) or ( iv ) , we proceed analogously : the part of @xmath42 contained in @xmath393 is an arc of a circle , which must be continued by two arcs of the same radius inside the strip , forming identical angles ( note that segments in @xmath352 are not allowed , since @xmath367 have not such segments ) ; this yields the desired symmetry .
[ rem : symm1 ] the vertical reflective symmetry of an isoperimetric set @xmath5 which produces a set @xmath367 of type ( ii ) after schwarz symmetrisation also follows from the following argument : call @xmath452 the length of the top slice , and remove an interval of length @xmath447 from the right endpoint of each horizontal slice of @xmath5 .
the resulting set has the same volume and perimeter that the biggest ball inside the strip , so it must coincide with that ball .
adding now the removed intervals , we get that the original set @xmath5 is symmetric . on the other hand , in case that @xmath367 consists of a set of type ( iii ) or ( iv ) , the portion of @xmath42 in @xmath393 must be optimal for the free boundary problem ( fixing the corresponding segment on @xmath352 ) , so it must be an arc of a circle ; for case ( iv ) , an identical arc will appear in @xmath394 . as pieces of @xmath42 contained in \{y=1 }
are not allowed in these cases , the symmetry of @xmath5 follows since inside the strip we must also have the optimal configuration for the fixed boundary segments .
[ rem : symm2 ] recall that the horizontal slices of @xmath5 are intervals in view of proposition [ wsteiner ] . consider the symmetrised set @xmath367 , which will be one of those sets described in lemma [ lemstripposs ] ; it is clear that the length of the horizontal slices of @xmath367 is a continuous positive function of @xmath453 , and so the same property holds for @xmath5 .
this also suffices to conclude , by using ( * ? ? ?
* th.1.3 ) , that @xmath5 has vertical symmetry . in view of lemma [ lemstripposs ] and proposition [ prop : simetrico ] , we have completely classified the isoperimetric candidates for the strip density .
we will now analyse numerically their behaviour .
our computations below lead to the following conjecture [ conj : strip ] .
theorem [ teostrip ] proves everything except the elimination of sets of type ( iv ) .
[ conj : strip ] given @xmath454 with the strip density , @xmath17 in the strip @xmath455 and @xmath349 outside , there exists a value @xmath456 such that the isoperimetric sets are : * balls of type ( i ) for areas less than @xmath457 ; * sets of type ( ii ) for values of the area in @xmath458 $ ] ; * sets of type ( iii ) for areas greater than @xmath459 .
it is not difficult to compare the perimeters of the above candidates ( i ) , ( ii ) , ( iii ) and ( iv ) , for equal volumes . for a ball in the strip enclosing area @xmath460 , its perimeter
is given by @xmath461 ; and a set of type ( ii ) enclosing area @xmath462 has perimeter given by @xmath463 .
on the other hand , by using the snell law and , for instance , computations in ( * ? ? ? * prop .
2.1 ) , sets of type ( iii ) can be parametrized in terms of the corresponding generalised mean curvature @xmath464 .
thus , denoting by @xmath465 it can be checked that the perimeter and the area enclosed by such a set are given by @xmath466 we point out that , when @xmath464 tends to zero , the area enclosed @xmath467 increases , while for @xmath464 close to one , @xmath467 achieves its minimum value . in a similar way , we can express the perimeter and the area enclosed by a set of type ( iv ) in terms of the mean curvature @xmath464 . by denoting by @xmath468 we have that necessarily @xmath469 , and that the perimeter and area enclosed by a set of type ( iv ) with curvature @xmath464 are equal to @xmath470 figure [ fig : perfiles ] shows several graphs where the perimeters of these types of sets are displayed , for different values of @xmath471 .
+ for these and many others values of @xmath471 , we have obtained numerically the behaviour described in conjecture [ conj : strip ] , as those graphs show .
moreover , it seems that sets of type ( iv ) are always beaten by sets of type ( iii ) , which are the solutions for large values of the area ( see theorem [ teostrip ] ) .
thus , sets of type ( iv ) are not expected to be solutions . [ re : re ] it is not difficult to justify that sets of type ( ii ) can not be the solutions for large values of the area . indeed , for any @xmath349 , and for curvature @xmath464 close enough to zero , we have @xmath472 , which means that the set of type ( iv ) is better than the set of type ( ii ) for area @xmath473 ( which shall be a large value of the area ) . in addition
, the reverse inequality holds for values of @xmath464 close enough to one , and then sets of type ( ii ) are better for such ( smaller ) areas .
we also have these same properties when considering type ( iii ) sets instead of type ( iv ) ones .
sets of type ( iii ) and ( iv ) possess an unexpected feature .
they have curvature less than one but , for curvature slightly less , perimeter and area _ decrease_. this can be noticed in figure [ figreversing ] , showing the graph of perimeter in terms of the enclosed area in case @xmath474 : the solid curve ( the lower one ) is associated to sets ( iii ) , the dotted ( upper one ) to type ( iv ) .
curvature is taken from 1 ( corresponding to the `` left '' endpoint of each graph ) to 0.8 .
notice that initially ( i.e. for curvature close to one ) each graph `` goes left '' , until a minimum ( for both area and perimeter ) is obtained in correspondence of the corner point of the graph . even if without a complete result , in theorem [ teostrip ]
we will characterise most of the isoperimetric sets for the strip density .
first we have to consider the following `` line '' density setting , which we will see as a sort of `` blow - down '' of the strip density .
consider the density @xmath80 on @xmath0 given by @xmath475 [ theoline ] for the density in @xmath454 defined by @xmath80 , isoperimetric sets exist for any given volume and are bounded by an arc meeting the horizontal axis with angle @xmath476 , and with the chord on the axis as in figure [ figline ] .
the existence of an isoperimetric set for any prescribed volume , as well as its connectedness , follows from a compactness argument similar to the one in proposition [ propexistn2 ] . as in proposition
[ wsteiner ] , schwarz symmetrisation with respect to the vertical axis ensures that an isoperimetric set symmetric with respect to such axis exists for any volume bound .
we claim that an isoperimetric set with such a symmetry must be the one described in the statement ; the fact that any isoperimetric set is symmetric with respect to the vertical axis easily follows by adapting one of the arguments of proposition [ prop : simetrico ] and remarks [ rem : symm1 ] and [ rem : symm2 ] .
the regularity of the isoperimetric boundary ensures that for @xmath477 one of the following holds : * it consists of a complete circumference ; * it consists of a single arc with chord a segment @xmath478 on the horizontal axis ( possibly reducing to a single point ) ; * it is empty . the same possibilities , which we will call ( @xmath479 ) , ( @xmath480 ) and ( @xmath481 ) , are given also for @xmath482 , with @xmath483 defined analogously .
taking into account the snell law , our thesis is equivalent to proving that the isoperimetric set falls into one of the cases ( @xmath484@xmath481 ) or ( @xmath485@xmath480 ) .
we can discard the cases ( @xmath486@xmath479 ) , ( @xmath486@xmath480 ) and ( @xmath484@xmath479 ) : in fact , one could translate vertically the two components towards the origin until an `` irregular contact '' is obtained .
the case ( @xmath484@xmath480 ) can be discarded too .
first of all , the two segments @xmath478 and @xmath483 must coincide ; otherwise ( by vertical symmetry ) one would be contained in the other , say @xmath487 . the arcs @xmath488 and @xmath489 meet the horizontal axis satisfying snell law , and so one of them forms ( with the same axis ) an angle @xmath490 , and the other one will form an angle @xmath491 .
this easily implies that the chords @xmath478 and @xmath483 have the same length , and so that our set is a round ball .
in particular , perimeter and area are the same as the cases ( @xmath486@xmath481 ) and ( @xmath479@xmath485 ) , which we are going to discard in a while .
the case ( @xmath485@xmath481 ) is of no concern , as it corresponds to null volume . for sets @xmath5 falling into cases ( @xmath486@xmath481 ) or ( @xmath479@xmath485 ) , i.e. when @xmath5 is a ball , the ratio @xmath492 is equal to @xmath493 . on the other hand ,
when @xmath5 falls into ( @xmath484@xmath481 ) ( the case ( @xmath485@xmath480 ) is analogous ) it consists of a circle minus the area enclosed by the chord @xmath478 , the angle at the center determined by @xmath478 being equal to @xmath494 by snell law .
easy computations yield that @xmath495 where we have used that @xmath496 .
this concludes the proof .
we also need to introduce the function @xmath497 defined as the perimeter of a circular arc with unit chord and area @xmath58 .
let us briefly recall some of its properties ( see ( * ? ? ?
* section 15.5 ) ) : * @xmath498 is convex in @xmath499 $ ] and concave in @xmath500 ; * @xmath501 and @xmath502 .
we stress that the boundary case @xmath503 corresponds to a semicircle of radius @xmath115 .
we now collect in the following theorem the partial results we are able to prove for this strip density .
[ teostrip ] the isoperimetric sets for area @xmath357 for the strip density , defined by @xmath17 in the strip @xmath504 and @xmath12 outside , are * balls of type ( i ) if @xmath460 ; * sets of type ( ii ) if @xmath505 ; * sets of type ( iii ) or ( iv ) if @xmath506 ; * sets of type ( iii ) if @xmath507 , for suitable @xmath508 . in particular , for area @xmath459 there are at least two different isoperimetric sets
. moreover , type ( iv ) never occurs for @xmath509 , i.e. in this case one can take @xmath510 .
set @xmath5 an isoperimetric region for area @xmath70 . to verify part ( a ) of the statement
, we just need to show that any set of type ( ii ) , ( iii ) or ( iv ) has area greater than @xmath457 .
this is clear for sets of type ( ii ) . for sets of type ( iii )
we observe that , if the three arcs have radius @xmath511 , then the arc outside the strip is greater than the semicircle with same radius , and so the enclosed area is greater than the area of that semicircle ( times the density @xmath9 ) , whence greater than @xmath512 .
otherwise we must have @xmath513 and it is easy to see that the internal arcs meet @xmath352 in such a way that the angle between the segment @xmath514 and each arc inside the strip is greater than @xmath515 . using the snell law it is not difficult to observe that the sum of the central angles associated with the three arcs is greater than @xmath516 ; moreover , the corresponding sectors of circle do not overlap and are contained in @xmath5 .
this means that the area of @xmath5 is greater than that of a circle of radius @xmath420 , hence greater than @xmath457 . for sets of type ( iv ) we can use a similar argument and obtain that the four central angles have sum greater than @xmath516 .
let us check that the four associated sectors do not overlap ; in this way the area of the candidate is greater than that of the circle with same radius @xmath517 and are done .
it is sufficient that the sectors corresponding to the upper and lower arcs do not overlap , i.e. , that the center of the upper arc has positive @xmath453-coordinate ( and so the center of the lower arc has negative @xmath453-coordinate ) .
since the parameter @xmath518 in denotes half of the central angle corresponding to the upper arc , the @xmath453-coordinate of its center equals @xmath519 .
we claim that the graphs representing the perimeter in terms of the area enclosed , corresponding to sets of type ( ii ) and ( iii ) , intersect only once ( see the graphs of figure [ fig : perfiles ] ) .
indeed , the curvature of the sets of type ( iii ) is always less than one ( see ) , which is the curvature of any set of type ( ii ) .
since curvature coincides with the _ slope _ of those graphs , this implies that those graphs intersect at most once .
moreover , type ( iii ) does not exist for area @xmath457 and , by remark [ re : re ] , beats type ( ii ) for large area .
therefore , there exists some @xmath520 such that sets of type ( ii ) are better than sets of type ( iii ) for areas less than @xmath521 , while sets of type ( iii ) are better for areas greater than @xmath521 .
the same reasoning can be done for sets of type ( iv ) : sets of type ( ii ) are better than type ( iv ) ones for areas less than @xmath522 , the contrary for greater areas . part ( b ) thus follows by noticing that any type-(i ) set has area at most @xmath457 ; the value @xmath459 can be characterised by @xmath523 .
statements ( c ) and ( d ) require more work .
by contradiction , suppose there is a sequence of areas @xmath524 such that the related isoperimetric sets @xmath525 are of type ( iv ) .
let @xmath526 be the radius of the curvilinear components of @xmath527 ; notice that each of the components in @xmath393 and in @xmath394 meets the strip with an angle lying in @xmath528 $ ] .
this forces each of these two arcs to bound a region of area at least @xmath529 ( @xmath530 depending only on @xmath9 ) , while the area of @xmath525 contained in the strip is of order @xmath526 . in particular , for large @xmath531 a `` big '' portion ( at least @xmath532 , @xmath533 ) of the area of @xmath525 lies in the half - plane above the strip , and another `` big '' portion lies below the strip .
let us perform the following `` blow - down '' operation : for any fixed @xmath531 we make a dilation of a factor @xmath534 so that @xmath525 is mapped into a set of area 1 .
more precisely , we consider the `` thin strip '' density @xmath292 on @xmath0 given by @xmath535 clarly @xmath536 pointwise , where @xmath80 is the line density defined in , and the sets @xmath537 are isoperimetric and of unit weighted area .
being connected they satisfy the diameter estimate . up to horizontal translations
, we can therefore suppose that they are contained in a big ball @xmath426 , and a compactness argument ensures that a subsequence converges in @xmath538 to a certain set @xmath94 .
this means that @xmath539 where @xmath80 is the function defined by ; this is because @xmath80 and @xmath292 differ only on the thin strip , whose intersection with the ball @xmath426 has area going to 0 .
semicontinuity of the perimeter and the inequality @xmath540 imply @xmath541 it follows that @xmath94 has to be isoperimetric for the density @xmath80 ; in fact , were another unit area set @xmath542 isoperimetrically better than @xmath94 we would obtain @xmath543 for some positive @xmath544 , and so @xmath545 could not be isoperimetric ( for @xmath292 ) for large @xmath531 . we will reach a contradiction if we show that @xmath94 can not be one of the isoperimetric sets described in theorem [ theoline ] .
in fact , for large @xmath531 the part of @xmath545 lying above the `` thin strip '' has area at least @xmath546 , and the same holds for the part below it .
this must happen also for @xmath94 , and so it can not coincide with one of the sets in theorem [ theoline ] .
it remains to show that for @xmath547 , case ( iv ) can never occur .
consider in fact any such set ; let @xmath447 be the length of the chords of the upper and bottom caps and @xmath424 be the ( euclidean ) area enclosed by each of them .
replace these two caps with a segment of length @xmath447 on @xmath353 and a cap in @xmath393 , enclosing euclidean area @xmath548 and lying upon the same chord as the removed upper cap . to prove that we have decreased perimeter we have to show that @xmath549 setting @xmath550 , this is equivalent to show that @xmath551 since @xmath552 , the convexity / concavity properties of @xmath498 ensure that @xmath553 for @xmath554 and @xmath555 for @xmath556 .
therefore the minimum of @xmath464 is attained at some @xmath557\pi/16,\pi/8[$ ] .
since @xmath498 is convex in @xmath499 $ ] we have @xmath558 while concavity in @xmath500 yields @xmath559 the last two equations imply that @xmath560 as desired .
the bound @xmath561 in the last assertion of theorem [ teostrip ] is not optimal .
in fact , a closer examination of the inequalities we have used leads to @xmath562 , and actually the statement holds for @xmath563 . unfortunately
, since @xmath564 one has @xmath565 : in other words , our argument of `` substituting upper and bottom caps with a unique one '' fails for small values of @xmath9 . in this section
we denote by @xmath26 the `` ball '' density on @xmath0 taking values @xmath566 on the open ball @xmath567 and 1 on @xmath568 .
the value of @xmath26 on @xmath569 is @xmath570 in order to guarantee its lower semicontinuity .
we begin with an existence result for the isoperimetric problem . for the ball density ,
isoperimetric sets exist for any given area .
let @xmath357 be the area bound ; it will be sufficient to prove that there exists @xmath118 and a minimising sequence @xmath571 such that @xmath572 in this way , we will reduce to an isoperimetric problem in a space with finite total mass , where existence of isoperimetric sets is a well known result ( see @xcite ) . to do this , consider any minimising sequence @xmath573 with @xmath574
; it is not restrictive to assume @xmath545 to be smooth open sets , and so one can write @xmath575 , where @xmath576 denote the connected components of @xmath545 . if @xmath147 is such that @xmath577 , then estimate easily provides @xmath578 ( depending only on @xmath70 ) such that @xmath579 . for the other components , we can substitute @xmath580 with a unique ball @xmath581 ( for certain @xmath582 ) enclosing the same volume , thus defining the set @xmath583 . notice that @xmath526 is bounded uniformly in @xmath531 by some @xmath584 depending only on @xmath70 , since @xmath585 , and that @xmath586 . anyway
, we have a lot of freedom in the choice of the center @xmath58 ; i.e. , we can move @xmath581 towards the origin , without touching the other components of @xmath525 , until we obtain ( say ) @xmath587 .
the nature of the isoperimetric problem , as we will see in a moment , is completely different in the two cases @xmath12 and @xmath20 .
therefore , we treat them separately .
we start with a description of all possible candidates for isoperimetric set .
the first one is a round ball , and , as we shall see in theorem [ ballmagg ] , balls are isoperimetric for volumes @xmath588 .
other more complicated candidates arise when we look for isoperimetric sets with area @xmath70 smaller than @xmath589 .
[ ballmino ] if the prescribed area @xmath70 is smaller than @xmath589 , then an isoperimetric set is one of the following ( see figure [ ex_isop ] ) : * a round ball contained in @xmath590 ; * the region enclosed by an arc of @xmath569 and another internal circular arc ( inside @xmath426 ) , the angle between them being equal to @xmath591 ( i.e. , the limit case in the snell law ) . in what follows
, we will denote by @xmath5 an isoperimetric set of area @xmath70 and by @xmath592 its boundary .
_ step 1 : @xmath5 is either a ball outside @xmath426 ( i.e. a set of type @xmath593 ) or the closure of each of its components is simply connected and intersects both @xmath426 and its complement . _
it is not difficult to see that the isoperimetric boundary @xmath594 can be decomposed into countably many rectifiable jordan curves @xmath595 , and that @xmath5 coincides with the union of the connected and simply connected open regions @xmath525 enclosed by each @xmath596 . otherwise , we could write @xmath597 , where @xmath598 is the region enclosed by a certain jordan curve @xmath596 and @xmath599 is a proper subset of @xmath598 .
since our density is rotationally invariant , we could rotate @xmath599 about the origin until it touches @xmath596 : perimeter and area are unaffected , but regularity fails .
since balls outside @xmath426 are better than inside , there is no @xmath525 with @xmath600 , or we could substitute all such regions with a unique ball outside @xmath426 , with same area and less perimeter .
similarly , there is at most one single component outside @xmath426 , which must be a ball .
if this is the case , there is no other component @xmath525 such that @xmath601 intersects both @xmath426 and its exterior , otherwise we could move our external ball until it touches @xmath525 , thus violating regularity . therefore , the isoperimetric set would be a ball like in part @xmath593 of the statement . in the rest of the proof
we will focus our attention on the remaining case that any @xmath601 intersects both @xmath426 and its exterior .
_ step 2 : only a finite number of internal arcs is allowed : as a consequence , @xmath5 has regular trace inside @xmath426 .
_ recall that @xmath602 and @xmath603 consists of circular arcs @xmath604 with endpoints on @xmath569 ( we have already excluded whole circles ) , with the same constant curvature @xmath605 . moreover , @xmath5 must contain all the regions @xmath545 enclosed by a @xmath604 and the corresponding arc on @xmath569 , determined by the endpoints of @xmath604 ( notice that , for internal arcs , there are two possible regions of this type , a bigger and a smaller one ) . otherwise , we could use a cutting technique .
suppose by contradiction that there are internal arcs @xmath604 as in figure [ internalarcs ] of arbitrarily small length and set @xmath606 to be the associated central angle .
due to curvature @xmath605 , the area of the region @xmath545 enclosed by @xmath604 is of order @xmath607 . by cutting @xmath545 from our set
we would gain perimeter @xmath608 , and we could restore the lost area @xmath609 with some ball of radius ( and perimeter ) @xmath610 .
this is a contradiction , since for sufficiently small @xmath606 we would overall gain perimeter . as a straightforward consequence
, only a finite number of internal @xmath604 s is allowed .
this implies also that the number of connected components @xmath525 is finite : in fact , @xmath611 , whence the boundary @xmath596 must meet the open ball @xmath426 , and so contains an internal arc @xmath604 .
_ step 3 : external arcs can not meet @xmath569 with angle strictly greater than @xmath515 . _ this is a technical point : it will allow us to use arguments like `` put external arcs in succession , so that two consecutive ones have a common endpoint '' , which would be precluded if one angle were greater than @xmath515 .
in fact , by doing such an operation it could happen that two consecutive regions overlap ; it is evident that this pathology can not occur if the meeting angles are less than @xmath515 .
suppose now that there exists such an external arc @xmath249 , with endpoints @xmath201 and @xmath413 ; let us prove that the distance between @xmath249 and any other external arc is bounded away from 0 . otherwise , a sequence of external arcs @xmath612 would accumulate at @xmath201 ( or @xmath413 )
, so some arc near @xmath201 would lie `` below '' @xmath249 ( i.e. , there exists a half line from the origin meeting both @xmath249 and @xmath612 ) ; say that the accumulating sequence of arcs stays `` on the left '' of @xmath201 .
since there is only a finite number of internal arcs we just have the following two possibilities : * there is a small arc @xmath613 , with @xmath201 as the right endpoint ( i.e. @xmath614 is staying `` on the left '' of @xmath201 ) which is contained in one of the arcs on @xmath569 determined by some internal arc @xmath604 .
informally , we could say that the sequence @xmath612 `` rests '' on the arc @xmath614 .
we could then use a small segment @xmath615 to join an external arc @xmath612 `` below '' @xmath201 to some point on @xmath249 close to @xmath201 , and add the region enclosed by @xmath616 and the @xmath612 s . in this way we have gained both perimeter and area , and we could use cutting arguments to restore area and gain perimeter again , which is contradictory .
* there is a small arc @xmath614 on @xmath569 , with @xmath201 as the right endpoint , which is not contained in any of the arcs on @xmath569 determined by the internal arcs @xmath604 .
this implies that each external arc @xmath612 , together with the arc it determines on @xmath569 , bounds a connected component contained in the region with density 1 , which could be cut away and replaced by balls to restore area and gain perimeter .
also , it is not possible for @xmath201 and @xmath413 to lie in the _ interior _ of the arc @xmath614 of @xmath569 determined by some internal arc .
otherwise we could use a segment @xmath615 to join some point on @xmath249 , close to @xmath201 , to some point on @xmath614 `` below '' @xmath201 , and add the region enclosed by @xmath249 , @xmath614 and @xmath615 , thus gaining both perimeter and area .
the previous arguments imply that the connected component @xmath525 , whose boundary contains @xmath249 , is `` free '' to move left or right : i.e. , unless @xmath525 is the only component , we can rotate it about the origin until one of the following happens * @xmath249 touches another external arc at a point outside @xmath617 ; * some endpoint of another external arc becomes too close to @xmath201 or @xmath413 ; * some internal arc of @xmath596 touches another internal arc at a point in @xmath426 ; * some internal arc not of @xmath596 becomes too close to @xmath201 or @xmath413 .
each of these possibilities leads to a contradiction ( since in any case we could use cutting arguments ) unless there exists only one connected component @xmath5 .
if this is the case , consider an internal arc @xmath604 : it is not difficult to realise that the arc on @xmath569 determined by @xmath604 s endpoints must be contained in the arc @xmath618 ( again , the one `` internal '' to the region enclosed by @xmath249 ) .
otherwise , it would be contained in the opposite arc @xmath618 ( we have already seen that the endpoints of @xmath604 can not lie one on the arc @xmath618 and one on the other arc @xmath618 , or we could cut at @xmath201 or @xmath413 ) , violating connectedness . since any internal arc `` rests '' on @xmath618 and there is only one connected component , there can not exist another external arc apart from @xmath249 .
therefore , the boundary of @xmath5 is made by @xmath249 plus a finite number of internal arcs @xmath604 , any of them `` resting '' on the arc @xmath618 .
were there more than one , we could rotate them until they touch , contradicting minimality also if the touching point is on @xmath569 ( the snell law ( [ eqfresnel ] ) would not hold ) .
therefore just an internal arc @xmath619 is allowed : by rotating it , we can suppose it to have an endpoint in @xmath201 , where the snell law must hold . by symmetry ,
the other endpoint meets @xmath569 with the same angle , and so must coincide with @xmath413 .
recall that @xmath249 and @xmath619 have the same curvature : so , they can be either symmetric with respect to the segment @xmath618 or form a complete circle . in both cases ,
the snell law would be violated unless the meeting angle is exactly @xmath515 yielding a contradiction .
_ step 4 : only a finite number of external arcs is allowed : in particular @xmath5 has regular traces on @xmath569 . _ by contradiction
: suppose that outer arcs accumulate at a point @xmath620 ; we can then find an arc @xmath621 such that * @xmath201 is an endpoint of @xmath622 ; * there are infinitely many external arcs @xmath623 whose endpoints lie on @xmath622 ; * the trace of @xmath624 on @xmath569 is constantly 0 or 1 on @xmath625 ( in other words : the finitely many arcs @xmath626 on @xmath569 , determined by internal arcs @xmath604 , are such that @xmath627 or @xmath628 )
. we can therefore re - order the @xmath623 s ( and the corresponding enclosed regions @xmath545 ) in sequence , in such a way that they still `` rest '' on @xmath622 and any two consecutive @xmath623 have a common endpoint on @xmath569 . in this way ,
area is clearly not affected , and so does the perimeter , since the third assumption on @xmath622 ensures that we are not creating ( nor destroying ) perimeter on @xmath569 .
regularity clearly fails with this construction , since the number of internal arcs is finite .
_ step 5 : the isoperimetric set is connected , enclosed by a piecewise @xmath96 jordan curve , and the snell law holds . _ as we have seen , the boundary @xmath629 is made by a finite number of arcs
; therefore , the whole @xmath594 is made by a finite number of arcs which can be internal , external or subarcs of @xmath569 . in what follows , we will call `` external '' also an arc of @xmath594 constituted by a subarc of @xmath569 .
the regularity of traces of @xmath5 on @xmath569 implies that the snell law must hold at any endpoint on @xmath569 of any arc of @xmath594 .
more precisely , at any of these endpoints a unique external and a unique internal arc meet , with the snell law satisfied .
this holds true also when the external arc is the limit case , i.e. a subarc of @xmath569 .
also , there is only a finite number of connected components ( which , as we have already seen , are simply connected too ) .
suppose there were more than one : we could rotate one of them about the origin until meeting another , contradicting the snell law .
_ step 6 : an external arc can not meet @xmath569 transversally . _ suppose by contradiction that this happens at a point @xmath620 ; then ( part of ) @xmath594 consists of two circular arcs @xmath630 ( outside @xmath426 ) and @xmath631 ( in @xmath426 ) , of the same radius , which intersect at @xmath201 respecting the snell law : see figure [ ball2 ] . since @xmath12 , the angle @xmath632 at @xmath201 between @xmath630 and @xmath569 is strictly smaller than the angle @xmath633 at @xmath201 between @xmath631 and @xmath569 ,
provided @xmath634 . in some sense ,
the discontinuity of the density causes a rotation , of center @xmath201 and of an angle @xmath635 , of the circle @xmath630 .
we recall that the curvatures of @xmath630 and @xmath631 ( with respect to the inner normal ) , and not just their radii , coincide .
let then @xmath636 be the other points where @xmath569 meets , respectively , @xmath630 and @xmath631 , and suppose @xmath637 : by prolonging @xmath594 after @xmath638 ( thus obtaining an arc @xmath639 , see figure [ ball3 ] ) we will fall into a contradiction .
in fact , the snell law holds again , and so the angle between @xmath639 and @xmath569 is exactly @xmath632 : this means that @xmath639 can be obtained from @xmath630 by means of the rotation centered at the origin which moves @xmath640 to @xmath638 . the arcs @xmath630 and @xmath639 do not intersect ( this happens , e.g. , when @xmath638 lies in the arc @xmath641 ) , or this would lead to a contradiction .
if @xmath638 does not lie on the arc @xmath641 and @xmath642 do not intersect , then @xmath639 meets @xmath569 at another point @xmath643 on @xmath644 .
prolong then @xmath639 from @xmath643 to obtain @xmath645 : this arc will surely meet @xmath631 , since it can be obtained by rotation and @xmath646 .
this is a contradiction . if @xmath647 , then @xmath648 ( circles with same curvature ) are either symmetric with respect to the segment @xmath649 or form a complete circle , and this contradicts the snell law unless @xmath650 , i.e. @xmath630 and @xmath631 form a complete circumference @xmath249 ( figure [ ball4 ] ) .
let @xmath413 be the center of @xmath249 , @xmath254 its radius , and set @xmath651 ; noticing that the radii @xmath652 and @xmath653 are tangent to @xmath569 , we obtain that the density area @xmath654 of the circle is less than @xmath655=r^2[\pi+({\lambda}-1){\beta}/2]\,.\ ] ] since the density perimeter of the circle @xmath249 is @xmath656 , we obtain that @xmath657}= 4[\pi+({\lambda}-1){\beta}/2]>4\pi\,,\ ] ] which is the isoperimetric ratio of any ball outside @xmath426 .
it follows that @xmath249 can not be isoperimetric .
_ step 7 : reduction to case @xmath658 .
_ in step 6 we excluded external arcs which are not subarcs of @xmath569 and have two endpoints on @xmath569 .
in fact they would be tangential to @xmath569 , and then would either enclose the whole @xmath426 ( contradicting assumption @xmath659 ) or a ball external to @xmath426 .
it follows that the isoperimetric boundary @xmath594 is composed by a finite number of internal @xmath604 s and some subarcs of @xmath569 .
were the internal @xmath604 s more than one , we could rotate one of them until it touches another one , thus violating minimality at the touching point .
therefore , the isoperimetric set is the one described in part @xmath658 of the statement .
we can now conclude this section with the main result .
[ ballmagg ] for the plane with density @xmath349 on the unit ball and @xmath17 elsewhere , there exists a @xmath660 such that 1 . if @xmath661 , the isoperimetric set with area @xmath70 is a ball contained in @xmath662 ( type [ ballmino](a ) ) ; 2 .
if @xmath663 the isoperimetric set is of type [ ballmino](b ) ; 3 . if @xmath664 , the isoperimetric set is a round ball containing the ball @xmath426 .
all three subcases of type [ ballmino](b ) of figure [ ex_isop ] , convex , chord and nonconvex , actually occur for some values of @xmath471 .
_ step 1 : if @xmath664 , the isoperimetric set is a round ball containing @xmath426 . _
let @xmath5 be any set of weighted area @xmath70 : since the region with density @xmath9 has euclidean area @xmath457 , the euclidean area of @xmath5 is at least @xmath665 and by classical isoperimetry its euclidean perimeter is at least that of the ball with the same area , i.e. the one with radius @xmath666 .
since @xmath667 , this also gives a lower bound for the @xmath26-perimeter of @xmath5 ; however , any ball as in the statement is such that these inequalities are in fact equalities , and so is isoperimetric .
conversely , for any other set , at least one of the inequalities is strict , and so it can not be isoperimetric .
_ step 2 : if @xmath70 is small , the isoperimetric set is a ball of type @xmath593 .
_ suppose by contradiction this is not true : then the isoperimetric set @xmath5 is a @xmath658-convex type ( see remark [ evolution ] ) , determined by an arc @xmath618 on @xmath569 and another one of center @xmath668 , meeting the arc @xmath618 with angle @xmath669 : as in figure [ smallv ] .
let @xmath421 be the angle @xmath670 and @xmath671 ; we have @xmath672 ( since @xmath673 ) and it is easy to compute the density perimeter of @xmath5 @xmath674 and the density area of @xmath5 @xmath675.\ ] ] we will get a contradiction if we show that @xmath676 . to do this , we express @xmath677 and @xmath9 as functions of @xmath421 and @xmath678 ( remember that @xmath672 and @xmath679 ) and , through a straightforward taylor expansion , we achieve @xmath680 then @xmath681 where we put @xmath682 .
we have to prove that @xmath683 which would allow to conclude . since @xmath684 , it is sufficient to show that @xmath685 in @xmath686 .
indeed it is @xmath687 _ step 3 : for @xmath70 close to @xmath589 the isoperimetric set is of type @xmath658 . _ to prove this , it is sufficient to exhibit a set @xmath94 of area @xmath70 with @xmath688 , which is the isoperimetric ratio of any ball of type @xmath593 .
let @xmath689 , and remove from @xmath426 a little curvilinear polygon of weighted area @xmath544 by a chord @xmath618 as in figure [ ball ] .
clearly , the remaining set @xmath690 is such that @xmath691 and @xmath692 as @xmath693 , and so the isoperimetric ratio converges to @xmath694 .
_ step 4 : there exists @xmath660 such that for @xmath695 the isoperimetric set is of type ( a ) , and for @xmath696 is of type ( b ) .
_ figure [ graph_ballmagg ] plots the perimeter as a function of the volume of type ( a ) configurations ( dashed line ) and type ( b ) configurations ( solid line ) ; the dotted line is the perimeter of balls containing the interior of the unit ball .
the slope of these graph is given by the curvature of the boundary of the isoperimetric set .
.,width=566 ] we have to show that there exists a unique point @xmath697 such that the graph of the type ( a ) touches the graph of type ( b ) .
we only consider convex type ( b ) sets , because otherwise the curvature , and hence the slope of the ( b)-graph , is negative .
we denote by @xmath698 the type ( b ) set with given volume @xmath70 ; it is given by @xmath699 for some @xmath700 and with @xmath701 the euclidean volume of @xmath698 is a given percentage @xmath702 of the euclidean volume of @xmath703 , that is @xmath704 we notice that @xmath705 is a decreasing function ; to see this one can perform a linear transformation in such a way that @xmath703 becomes the unit ball centered in the origin . with this modification , the ball @xmath426 is sent to the ball @xmath706 with @xmath707 meeting @xmath708 at a fixed angle . enlarging @xmath70
, we obtain a smaller ball @xmath706 , so that @xmath709 decreases .
we also denote by @xmath710 a ball of type ( a ) , with @xmath711 .
if @xmath459 is the first value of the volume for which @xmath712 , we have that @xmath713 and , using the slopes of the graphs , @xmath714 this implies that @xmath715 .
then it follows that @xmath716 for @xmath717 , and so @xmath718 , whence @xmath719 and then for @xmath717 the function @xmath720 is strictly smaller than @xmath721 , due to the above inequality of slopes . by numerical computation , all of the cases ( b)convex , ( b)chord , and ( b)nonconvex occur for @xmath471 close to 1 , while for great @xmath471 only ( a ) and ( b)nonconvex are isoperimetric .
[ evolution ] type ( b ) sets evolve from convex to nonconvex when they are isoperimetric solutions .
consider the curvature of the inner arc as a parameter ; since all these inner arcs will meet @xmath569 with constant angle , it turns that the enclosed volume is a decreasing function with respect to the curvature .
so when volume increases , curvature of inner arcs goes from positive to negative , and hence type ( b ) sets go from convex to nonconvex .
we begin with describing the candidate isoperimetric sets ; _ mutatis mutandis _ , most of the arguments are the same as in section [ ballmayor ] for @xmath12
. the candidate isoperimetric sets are ( see figure [ figbm1 ] ) : * balls entirely contained in @xmath426 ; * sets enclosed by two arcs with common endpoints : the first arc lies on @xmath569 , the second one is in @xmath590 and meets @xmath569 with angle @xmath722 ; * balls meeting @xmath426 orthogonally .
we closely follow the proof of proposition [ ballmino ] .
in many cases , it will be sufficient to exchange the roles of internal and external arcs .
we now agree to call _ internal _ an arc lying on @xmath569 .
again , we denote by @xmath5 an isoperimetric set of area @xmath70 and by @xmath594 its boundary .
_ step 1 : @xmath5 is either a ball inside @xmath426 ( i.e. a set of type ( a ) ) or the closure of each of its components intersects both @xmath617 and @xmath568 .
_ this point presents no substantial difficulty with respect to step 1 in proposition [ ballmino ] . in the rest of the proof
we will suppose that @xmath5 is not a type-(a ) ball , i.e. that the closure of any of its component intersects both @xmath617 and @xmath568 ; in particular , it can not be a ball in @xmath568 .
notice also that the boundary of any component must meet @xmath569 : otherwise this component would be a ball enclosing @xmath426 and with boundary out of @xmath617 , a configuration which is easily proved to be worse than a ball outside @xmath426 with same area . _ step 2 : only a finite number of external arcs are allowed . _ this is due to the fact that any external arc must leave @xmath569 with an angle between @xmath723 and @xmath724 ; curvature being constant , this means that each arc must determine on @xmath569 a corresponding arc whose length has a fixed positive lower bound .
so only a finite number of arcs on @xmath569 can be determined , and the claim easily follows .
_ step 3 : internal arcs can not meet @xmath569 with angle strictly greater than @xmath515 . _ _ step 4 : only a finite number of internal arcs is allowed : therefore @xmath5 has regular traces on @xmath569 . _
_ step 5 : the isoperimetric set is connected , enclosed by a piecewise @xmath96 jordan curve , and the snell law holds . _ these proofs are exactly the same as in the corresponding steps of proposition [ ballmino ] : it is sufficient to exchange the words _ internal _ and _
external_. _ step 6 : an internal arc can not meet @xmath569 transversally , unless the contact angle is @xmath515 . _ to prove this , one can follow again the proof of the corresponding step 6 in proposition [ ballmino ] , where however we excluded balls of type ( c ) .
this can not be done in the @xmath20 setting : we will instead show numerically that such a configuration can be isoperimetric ( see the proof of theorem [ conj : ballmino1 ] ) .
_ step 7 : reduction to cases ( b ) and ( c ) . _
we have shown that internal arcs can be only tangential or orthogonal to @xmath569 , and that there are only finitely many of them .
if tangential , they have to lie on @xmath569 ; otherwise they would have curvature more than 1 , thus being complete circles inside @xmath426 .
therefore , a connected component whose boundary possesses one ( or more ) arc on @xmath569 must be bounded by exactly one external arc and one arc on @xmath569 ; otherwise one could rotate the external arcs until forbidden meetings are obtained .
such a component falls into case ( b ) and must enclose the ball @xmath426 ; moreover , the angle between the external arc and @xmath569 must be @xmath722 .
we have shown that the components can be only of type ( b ) or ( c ) .
were there more than one , we could use rotation arguments to obtain a contradiction .
this concludes the proof .
figure [ graph_ballmino2 ] shows the graphs of perimeter as function of the enclosed area for candidates in the case @xmath725 : the dotted , dashed and solid lines refer , respectively , to candidates ( a ) , ( b ) and ( c ) .
then it is then not surprising the following theorem .
.,width=566 ] [ conj : ballmino1 ] for the plane with density @xmath20 on the unit ball and @xmath17 elsewhere , there exist some values @xmath726 such that the isoperimetric set of volume @xmath70 is given by 1 . a ball of type ( a ) if @xmath727 ; 2 . a set of type ( b ) if @xmath728 ; 3
a set of type ( b ) or ( c ) if @xmath729 ; 4 . a ball of type ( c ) if @xmath730
. * remark . *
we believe that @xmath731 in the previous theorem [ conj : ballmino1 ] ; that is , once sets of type ( b ) fail to be isoperimetric , type ( c ) sets are the solutions for larger volumes .
we have observed this fact numerically , although we are not able to give a rigorous proof of it .
since it will be useful in the sequel , we compute explicitly the perimeter and area of candidates ( b ) and ( c ) . for type-(b ) sets , let @xmath732 be the angle determined by the external arc ; by @xmath733 we denote half of the angle at the center of @xmath569 formed by the arc on @xmath569 ( see figure [ fig : alphabeta ] ) .
it is not difficult to show that @xmath734 and that , after setting @xmath735 to be the radius of the external arc , the perimeter is @xmath736 while the enclosed area is @xmath737 regarding balls of type ( c ) , if @xmath738 is the angle at the center determined by the internal arc , then the radius of the ball will be @xmath739 . therefore , the perimeter will be given by @xmath740 and the area enclosed by @xmath741 since type-(b ) sets enclose the ball @xmath426 , for area bounds less than @xmath589 we have only to compare candidates ( a ) and ( c ) . to prove the first part of the statement it will be sufficient to show that the ratio @xmath742 is strictly greater than @xmath743 , which is the same ratio computed for ( a)-balls
let us rewrite @xmath744 as @xmath745 the desired inequality @xmath746 is equivalent to @xmath747 i.e. @xmath748\ ] ] and in turn to @xmath749 remembering that @xmath750 $ ] , setting @xmath751 and discarding the negative term @xmath752 , it will be sufficient to show that @xmath753>0\,.\ ] ] recalling that @xmath754 $ ] it will suffice to prove that @xmath755 is positive in @xmath756 $ ] .
indeed it holds @xmath757 and @xmath758 which allows to conclude because @xmath20 . for area bounds greater than @xmath589 we have only to compare candidates ( b ) and ( c ) . from above
we have that the type-(c ) ball for area @xmath589 has perimeter greater than @xmath759 , with @xmath760 ; this will happen for any ( c)-ball with area greater than @xmath589 .
the second part of the statement follows by noticing that , when the area @xmath761 of the ( b)-candidate decrease to @xmath589 ( equivalently when @xmath762 , i.e. @xmath763 , i.e. @xmath764 ) , its perimeter @xmath765 decreases to @xmath766 . when considering a type ( b)-candidate associated with an external ball of radius @xmath254 , straightforward computations give @xmath767 for large @xmath254 .
analogously , for a type-(c ) ball of big radius @xmath768 we have @xmath769 by imposing the equality @xmath770 of enclosed areas , in order to compute @xmath768 as function of @xmath254 , one obtains @xmath771 . to show that @xmath772 for large areas , we just need to prove that @xmath773 for large @xmath254
. therefore it will be sufficient to show that @xmath774 is strictly positive for any @xmath7750,1[$ ] : this in turn follows since @xmath776 and @xmath7770,1[\,.\ ] ]
this section provides partial results on the isoperimetric problem for certain modifications of gaussian density @xmath778 on the plane .
previously , borell @xcite and rosales et al .
* introduction and section 5 ) proved by symmetrization that for the density @xmath779 isoperimetric regions are balls about the origin .
recently , brock et al .
@xcite proved that for the modified gauss space @xmath780 , @xmath781 , on the half
space @xmath782 , isoperimetric sets are bounded by vertical hyperplanes .
the same argument also shows that the same holds if we consider the density @xmath783 on the whole space @xmath1 , simply by considering the reflection with respect to the hyperplane @xmath784 .
in fact , if @xmath5 is an isoperimetric set , then @xmath785 is isoperimetric with respect to its volume bound , so its boundary in @xmath782 has to be a vertical hyperplane .
it is worth noticing that in the planar case @xmath52 and with @xmath786 , vertical and horizontal lines are the only straight lines with constant generalised curvature ; moreover , the only circles with constant generalised curvature are the one centered at the origin . motivated by these results , we will now consider the density in the plane defined by @xmath787 since the total mass is finite we have existence of isoperimetric regions for any prescribed volume .
we state the following [ conj:1 ] for density ( [ gaussdens2 ] ) on the plane , every isoperimetric region is bounded by a horizontal line ( for prescribed volumes close to half the total volume ) or by a vertical line .
we trivially have the following lemma .
vertical and horizontal lines are the unique straight lines with constant generalised mean curvature .
furthemore , circles do not have constant generalised mean curvature .
a first consequence of the previous lemma is that no arc of a circle will be contained in the boundary of an isoperimetric region .
we now proceed to compare the perimeter of vertical and horizontal lines enclosing the same volume .
we shall see that , for some volumes , vertical lines are better , while horizontal lines beat vertical ones for other volumes .
[ propconj1 ] for volume fractions near @xmath115 , horizontal lines have less perimeter than vertical lines , and for small or large volumes , vertical lines have less perimeter than horizontal ones . for @xmath788 ,
let @xmath789 be the horizontal line @xmath790 .
this line bounds a region with volume equal to @xmath791 where @xmath143 represents the gamma function of two arguments .
on the other hand , for the perimeter of the region bounded by @xmath789 we have @xmath792 for @xmath793 , let @xmath794 be the vertical line @xmath795 .
then , the region bounded by this line has a volume equal to @xmath796 where @xmath797 is the error function .
moreover , the perimeter of the region bounded by @xmath794 is given by @xmath798 given @xmath453 , @xmath793 , it is easy to check that @xmath799 if and only if @xmath800 consider now @xmath453 , @xmath793 satisfying condition .
then @xmath801 hence , if this last condition is satisfied , the corresponding vertical line is isoperimetrically better than the horizontal line .
observe that the left hand side in condition can be expressed as a function of @xmath453 , by using .
numerical computations show that for values of @xmath453 close to zero , e.g. when @xmath802,\ ] ] the left hand side in is greater than @xmath803 , and so , horizontal lines are better . for @xmath804 such a term is less than @xmath803 and so vertical lines are better .
this completes the proof . in this situation ,
domains bounded by two ( symmetric ) lines can be easily discarded , since the perimeters of horizontal and vertical lines are strictly decreasing functions .
proposition [ propconj1 ] indicates that vertical or horizontal lines may appear as minimisers . as in @xcite ,
given a set @xmath805 , we could apply two adapted steiner symmetrisations ( one in the vertical direction and another one in the horizontal direction ) , to obtain another set @xmath806 which would have the same volume and no more perimeter than @xmath34 . in this setting
, it can be checked that horizontal lines far away from the origin ( i.e. for small or large volumes ) are unstable by using the second variation formula ( see ( * ? ? ?
* proposition 3.6 ) ) with a suitable mean zero function .
unfortunately , the study of the stability of the rest of the lines seems a harder question .
the study of manifolds with density , and of the isoperimetric problem in particular , is relatively recent , and many related works have appeared during the last years @xcite . in this section , we will list several open problems in this setting : some of them ( questions 14 ) are closely related to the present paper , while the other ones arose during morgan s lectures on `` manifolds with density '' at the gmtlap meeting in modena , february 2007 .
[ queball ] the case @xmath807 in section [ ss : bmenor1 ] is not completely solved ( see theorem [ conj : ballmino1 ] ) : it is left to prove that , once sets of type ( c ) are isoperimetric , type ( b ) sets cannnot be solutions for any other larger volume .
it would be also interesting to investigate the ball density problem in @xmath1 , @xmath808 . for @xmath12 and volume @xmath809 balls containing @xmath426 are isoperimetric : the proof is the same as the planar case .
one could further imagine densities which are constantly 1 on @xmath0 except for some balls : this situation seems particularly interesting , since it may well provide non connected isoperimetric sets .
consider for example the piecewise constant density @xmath9 on @xmath810 and 1 outside , where @xmath811 and @xmath426 , @xmath706 are two sufficiently distant balls of radius one .
it is not difficult to show that isoperimetric sets , at least for some area bounds , are not connected .
in fact , take @xmath812 as prescribed area : were the isoperimetric set connected , by theorem [ ballmagg ] it would consist of a ball containing @xmath426 or @xmath706 , but not both of them as they are chosen far away ( recall the diameter estimate ) .
therefore , its density perimeter would be at least @xmath813 , while @xmath810 satisfies the area bound with perimeter @xmath814 , which is smaller for large values of @xmath9 .
[ quebobkov ] other modifications of the gaussian density on the plane would be worth investigating .
for instance , we propose the density @xmath815 on @xmath0 .
an interesting approach seems to be related to a work by s.bobkov @xcite , where it is proved that half - spaces are isoperimetric in @xmath1 with density given as a product @xmath816 , under certain assumptions on @xmath817 .
unfortunately , these hypothesis are not quite satisfied in our case @xmath818 .
consider @xmath0 endowed with a density attaining its minimum value in several points .
then it is reasonable to think that for small prescribed areas an isoperimetric region should be a nearly round disc containing one of these points .
the reader may compare this with @xcite , where this result is proven for a particular density with a unique minimum point .
another interesting question consists of finding the isoperimetric regions in euclidean space endowed with a density which is a finite sum of gaussian measures with different barycenters and variances . in the particular case that all the barycenters lie in the same line @xmath622 , orthogonal lines to @xmath622
seem nice candidates for being the solutions .
which regions ( with rectifiable boundary ) are isoperimetric for some density ?
probably any smooth region has this property .
take a density with very low values along the boundary , perhaps a low constant value with normal derivative determined by the constant generalised mean curvature condition .
fairly recent `` quantitative '' isoperimetric inequalities show that a set with perimeter close to the optimal one must approximate a minimiser in shape ( see @xcite ) .
can such results be generalised to manifolds with density ?
suppose the density @xmath26 is `` conic '' , i.e. constant along rays starting from the origin .
give necessary conditions for a cone @xmath822 to be the boundary of a perimeter
minimising set .
the main case is given by the half - space density and the snell refraction law is one necessary condition .
such results could be used more generally .
take a point of an isoperimetric boundary which is also a discontinuity point for the density , and perform a blow - up . in the limit
the above situation is retrieved .
y.d burago and v.a .
zalgaller . .
translated from the russian by a.b .
grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] ( 285 ) . springer series in soviet mathematics .
springer - verlag , berlin , 1988 . | we study the isoperimetric problem in euclidean space endowed with a density .
we first consider piecewise constant densities and examine particular cases related to the characteristic functions of half - planes , strips and balls .
we also consider continuous modification of gauss density in @xmath0 .
finally , we give a list of related open questions . |
the observed scalar particle in 2012 of mass @xmath6 gev @xcite , is consistent with the standard model ( sm ) higgs boson . in the context of the minimal supersymmetric standard model ( mssm ) , this motivates considering models of supersymmetry breaking in which the stop superpartner is heavy ( beyond the reach of the lhc ) or a model in which a large trilinear @xmath0 soft supersymmetry breaking parameter can be generated at low energies .
the first option is disfavored by fine - tuning arguments while the second one allows for lighter stops @xmath2 tev and it is thus preferred by naturalness arguments .
the solution advocated in this work is to resolve the previous issue by introducing an extra ( fifth ) dimension and taking advantage of the power low running @xcite to generate a sizable value of @xmath0 starting from a very small value .
we define 5-dimensional ( 5d ) mssm , the higgs superfields and gauge superfields always live in the bulk . as consequence
these fields will have kaluza - klein modes which contribute to the rges at @xmath7 and additional matter associated to five dimensional @xmath8 super yang - mills .
different possibilities of localising the matter superfields can be studied .
we shall consider the limiting case of superfields with sm matter fields restricted to the brane , and the rges for this scenario can be found in ref .
therefore there will be no additional kaluza - klein contributions of these matter fields to the rges@xcite .
we define 5-dimensional ( 5d ) mssm , the higgs superfields and gauge superfields always live in the bulk . as consequence
these fields will have kaluza - klein modes which contribute to the rges at @xmath7 and additional matter associated to five dimensional @xmath8 super yang - mills @xcite .
different possibilities of localising the matter superfields can be studied .
we shall consider the limiting case of superfields with sm matter fields restricted to the brane , and the rges for this scenario can be found in ref .
therefore there will be no additional kaluza - klein contributions of these matter fields to the rges@xcite . regarding the breaking of supersymmetry , whilst gauge mediation is favoured ( and some recent work on gauge mediated supersymmetry breaking in a five dimensional context may be found in ref .
@xcite ) , ultimately the universality of squark massses in gmsb means that even though the gaugino mediated limit @xcite might allow for light squarks ( and 5d rge evolution allows for a large @xmath0 and the observed higgs mass ) , the collider bounds on first and second generation squarks @xcite , in the supra - tev range would apply also to the @xmath9 generation squarks , i.e. the stops , which as discussed before , is both phenomenologically less interesting and unnatural . therefore we wish for some other description of supersymmetry breaking that may allow for stops to be lighter than their first and second generation counterparts , such as in refs . @xcite . in this work
we will therefore be unspecific about the precise details of how supersymmetry is broken and as a result also our conclusions will apply quite generally .
we do however make some minimal specifications : * we take as inputs the yukawa and gauge couplings at the susy scale , @xmath10 tev .
* we will assume supersymmetry breaking occurs at the unification scale , which is found by finding the scale at which @xmath11 , which is lowered compared to the 4d mssm , by the effects of the compactification .
* we specify the value of the gluino mass , @xmath12 at @xmath10 tev . *
we take the trilinear soft breaking terms , @xmath13 , to vanish at the unification scale .
we solved the combined set of differential equations numerically by using the above conditions , taking the `` third family '' approximation in which we only evolve third generation rges .
this approximation is quite standard and is due to the relative smallness of the other yukawa couplings ( at least one order of magnitude ) compared to those of the third generation and as a result the other a - term values are also very small .
we further specified some parameters such as @xmath14 , @xmath15 and the value of the sfermion masses ( @xmath2 tev ) so as to allow for the rges to be solved , but these do not affect the overall result .
we solved the differential equations between @xmath16 gev and @xmath17 , which was typically only one order larger than the unification scale .
, for two different values of the compactification scales 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath12 of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of the compactification scales 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath12 of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of the compactification scales : 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath18 $ ] of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of the compactification scales : 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath18 $ ] of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of the compactification scales 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath18 $ ] of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of the compactification scales 10 tev ( left panel ) , @xmath1 tev ( right ) , with @xmath18 $ ] of 1.7 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of gluino masses , @xmath12 : 1.7 tev ( left panel ) , 3 tev ( right panel ) , with @xmath19 of 10 tev , as a function of log(e / gev).,title="fig:",width=264 ] , for two different values of gluino masses , @xmath12 : 1.7 tev ( left panel ) , 3 tev ( right panel ) , with @xmath19 of 10 tev , as a function of log(e / gev).,title="fig:",width=264 ] , corresponding to those of the 5d mssm . , scaledwidth=90.0% ] for different values of the stop mass , for @xmath20 of @xmath21 gev ( left panel ) and @xmath22 tev
( right panel ) .
, title="fig:",width=283 ] for different values of the stop mass , for @xmath20 of @xmath21 gev ( left panel ) and @xmath22 tev ( right panel ) . ,
title="fig:",width=283 ] a sufficiently large trilinear @xmath0 soft supersymmetry breaking parameter allows us to reproduce the measured higgs field mass of @xmath6 gev , while keeping a light stop superpartner ( below @xmath2 tev ) as preferred by the fine tuning argument for the higgs mass .
realising such a high @xmath0 is usually difficult ( in supergravity , mgmsb , ... ) @xcite but it is shown here concretely that 5d mssm with compactification scale around 10-@xmath1 tev can achieve large @xmath0 amounts at low scale ( starting from @xmath23 ) , thanks to the power law running and simultaneously with an approximate unification of the gauge couplings , the precision of this gauge unification is illustrated quantitatively ( at one - loop ) for various values of the compactification scale as pictured in fig .
[ fig : alphas5d ] .
we also specify the yukawa coupling rge @xcite boundary conditions at @xmath10 tev , which interestingly appears to vanish when evolved to the unification scale as shown in fig .
[ fig : yuk5d ] .
more precisely , it is found that increasing the compactification radius tends to increase the size of the trilinear @xmath0 parameter at low scale as shown in fig .
[ fig : trilinearcompact ] .
in fact the absolute value of @xmath0 mimics the magnitude of the final value of the gluino mass ( at @xmath24 ) as can be seen in fig .
[ fig : trilineargluino ] . as this interesting result of large @xmath0 generated
radiatively is rather generic , the origin of supersymmetry breaking is left unspecified and the treatment adopted is effective .
this large @xmath0 can be realised in both cases of the three quark generations being on the boundary or only the third generation of matter fields on the boundary ( `` split families '' ) . in the later case
, the stop can be much lighter and possibly fall inside the lhc reach .
all these results are based on the beta function for the gauge couplings , yukawa couplings , trilinear terms , scalar soft masses , higgs masses , @xmath14 and @xmath15 parameters , whose analytical expressions can be found in ref .
@xcite . using the one - loop higgs mass formula in ref .
@xcite and fixing @xmath25 gev , @xmath26 gev , @xmath27 for @xmath28 we can see in fig . [
fig : higgsmassforat ] that for representative values of @xmath0 achievable in the 5d mssm , one may easily accommodate the lightest stop mass in the sub - tev range .
let us pay attention to dependence on the value of @xmath29 as pictured in fig .
[ fig : higgsmassfortandbeta ] . in fact the value of @xmath29 will depend greatly on how @xmath14 and @xmath15 are addressed in the context of supersymmetry breaking and hence the solution of the vacuum tadpole equations , but regardless of this , for values of @xmath30 the functions are approximately flat and we expect the value to fall within this interval .
we expect that the @xmath14 term is naturally of the order of the electroweak scale , where in fig .
[ fig : higgsmassforat ] we took a slightly large @xmath14 value of @xmath31 gev and in fig .
[ fig : higgsmassfortandbeta ] we took @xmath32 gev , leading typically to light higgsinos and winos . in conclusion ,
we have explored how 5d extension of the mssm may generate large @xmath0 to achieve the observed higgs mass and have sub - tev stops , perhaps observable at the lhc .
we computed the full one - loop rges for all supersymmetric and soft breaking parameters .
we find that the magnitude of @xmath0 follows closely that of the magnitude of @xmath12 and increases as the compactification scale decreases .
we would like to thank our collaborators aldo deandrea and moritz mcgarrie for their helpful discussions .
this work is supported by the national research foundation ( south africa ) .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b716 ( 2012 ) 1 [ arxiv:1207.7214 ]
. s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b716 ( 2012 ) 30 [ arxiv:1207.7235 ] .
a. abdalgabar , a. s. cornell , a. deandrea and m. mcgarrie , jhep * 1407 * , 158 ( 2014 ) [ arxiv:1405.1038 [ hep - ph ] ] .
k. r. dienes , e. dudas and t. gherghetta , phys .
b * 436 * , 55 ( 1998 ) [ hep - ph/9803466 ] .
k. r. dienes , e. dudas and t. gherghetta , nucl .
b * 537 * , 47 ( 1999 ) [ hep - ph/9806292 ] .
m. mcgarrie , jhep * 1304 * , 161 ( 2013 ) [ arxiv:1303.4534 [ hep - th ] ] .
a. bharucha , a. goudelis and m. mcgarrie , eur .
j. c * 74 * , 2858 ( 2014 ) [ arxiv:1310.4500 [ hep - ph ] ] .
f. brummer , m. mcgarrie and a. weiler , jhep * 1404 * , 078 ( 2014 ) [ arxiv:1312.0935 [ hep - ph ] ] .
s. abel and m. mcgarrie , jhep * 1407 * , 145 ( 2014 ) [ arxiv:1404.1318 [ hep - ph ] ] .
m. s. carena , j. r. espinosa , m. quiros and c. e. m. wagner , phys .
b * 355 * , 209 ( 1995 ) [ hep - ph/9504316 ] .
h. e. haber , r. hempfling and a. h. hoang , z. phys .
c * 75 * , 539 ( 1997 ) [ hep - ph/9609331 ] .
a. delgado , hep - ph/0409073 . | the possibility of generating a large trilinear @xmath0 soft supersymmetry breaking coupling at low energies through renormalisation group evolution in the 5d mssm is investigated . using the power law running in five dimensions and a compactification scale in the 10-@xmath1 tev range , to show that gluino mass may drive a large enough @xmath0 to reproduce the measured higgs mass and have a light stop superpartner below @xmath2 tev as preferred by the fine tuning argument for the higgs mass .
addtoresetequationsection .9 cm * ammar abdalgabar@xmath3 and a. s. cornell@xmath4 * _ @xmath4 national institute for theoretical physics ; _ _ school of physics and mandelstam institute for theoretical physics ,
_ _ university of the witwatersrand , johannesburg .
wits 2050 , south africa _ _ @xmath5 department of physics , sudan university of science and technology , _ _
khartoum 407 , sudan _ |
in this section , we recall basic definitions and properties of modal dependence logics , which are modal logics defined on the basis of team semantics . though team semantics is intended for the extension of ( classical ) modal logic obtained by adding dependence atoms , for the sake of comparison
, we will start by defining the team semantics and fixing notations for the usual ( classical ) modal logic .
fix a set @xmath6of propositional variables and denote its elements by @xmath7 ( possibly with subscripts ) .
formulas of _ ( classical ) modal logic _ , also called _ classical ( modal ) formulas _ , are defined recursively as:@xmath8 where @xmath9 ( called _ tensor _ ) denotes the disjunction of classical modal logic , and the implication @xmath10 is called _
intuitionistic implication _ for reasons that will become clear in the sequel .
a ( modal ) _ kripke frame _ is a couple @xmath11 consisting of a nonempty set @xmath12 and a binary relation @xmath13 .
elements of @xmath12 are called _ possible worlds _ or _ nodes _ or _
points_. a ( modal ) _ kripke model _ is a triple @xmath14 such that @xmath15 is a kripke frame and @xmath16 is a valuation function .
a set @xmath17 of possible worlds is called a _ team_. for any team @xmath18 , define @xmath19 and write @xmath20 for @xmath21 . a team @xmath22 is called a _ successor team _ of @xmath18 , written @xmath23 , if @xmath24 and @xmath25 for every @xmath26 .
we define inductively the notion of a classical modal formula @xmath27 being _ satisfied _ in a kripke model @xmath28 on a team @xmath17 , denoted @xmath29 , as follows : * @xmath30 iff @xmath31 * @xmath32 iff @xmath33 * @xmath34 iff @xmath35 for all @xmath26 .
in particular , @xmath36 iff @xmath37 * @xmath38 iff @xmath39 and @xmath40 * @xmath41 iff there exist @xmath42 such that @xmath43 , @xmath44 and @xmath45 * @xmath46 iff for all @xmath47 , @xmath48 implies @xmath49 * @xmath50 iff @xmath51 * @xmath52 iff there exists @xmath53 such that @xmath23 and @xmath48 for any kripke models @xmath28 , if @xmath54 holds for all @xmath17 , then we say that @xmath55 is _ true _ on @xmath56 and write @xmath57 . for any kripke frame
@xmath58 , if @xmath59 holds for all valuations @xmath60 on @xmath58 , then we say that @xmath55 is _ valid on _ @xmath58 and write @xmath61 .
if @xmath61 holds for all frames @xmath58 , then we say that @xmath55 is _ valid _ and write @xmath62 .
we write @xmath63 if for all kripke models @xmath56 and all teams @xmath18 , @xmath64 for all @xmath65 implies @xmath66 .
write simply @xmath67 for @xmath68 , and write @xmath69 if @xmath67 and @xmath70 .
it is easy to check that classical modal formulas @xmath27 satisfy the _
flatness property _ : flatness property : : @xmath71{rl } \mathfrak{m},x\models\alpha\!\!\!\!\!\!\!\!\!\!&\iff\mathfrak{m},\{w\}\models\alpha\text { for all } w\in x\\ & \iff \mathfrak{m},w\models\alpha\text { in the usual sense for all } w\in x\end{array}$ ] as a consequence , a few usual equivalences such as the following ones hold for classical formulas : @xmath72 recall that the hilbert - style system of classical modal logic @xmath73consists of the following axioms and rules : axioms : : 1 .
all axioms of classical propositional logic ( @xmath74 ) 2 . @xmath75 3 .
@xmath76 rules : : 1 .
modus ponens : @xmath77 2 .
necessitation : @xmath78 3 .
uniform substitution : @xmath79 for classical formulas , the system of @xmath73is sound and complete with respect to all kripke frames in the sense of team semantics too . to see why , by the completeness theorem of @xmath73with respect to the usual single - world semantics and the flatness property of classical formulas
, we have that @xmath80 in view of the flatness property , lifting the single - possible world semantics to team semantics does not add essentially new features to classical modal formulas .
let us now extend the usual ( classical ) modal logic by adding the dependence atoms @xmath81 and intuitionistic disjunction @xmath82 that violate the flatness property .
we call the resulting logic _
modal downward closed team logic _ ( @xmath83 ) and its language is defined formaly as follows : @xmath84 where @xmath85 are classical modal formulas .
we write @xmath86 for @xmath87 .
the satisfaction relation for @xmath83-formulas is defined as for classical modal formulas and additionally : * @xmath88 iff for any @xmath89 , @xmath90 for all @xmath91 implies @xmath92 . *
@xmath93 iff @xmath39 or @xmath40 immediately from the semantics it follows that @xmath83-formulas have the downward closure property and the empty team property defined below : downward closure property : : @xmath94\longrightarrow \mathfrak{m},y\models\phi$ ] empty team property : : @xmath95 we have discussed that the classical fragment of @xmath83(i.e . , the fragment consisting of classical formals only ) behaves exactly as classical modal logic .
@xmath83also inherits from classical modal logic many other nice properties , such as the preservation property under taking disjoint unions .
for any two kripke models @xmath14 and @xmath96 , their _ disjoint union _
@xmath97 is defined as @xmath98 where @xmath99 takes the disjoint union of two sets .
it can by proved by a routine argument that for any collection @xmath100 of kripke models , for every @xmath101 and every @xmath102 , @xmath103 from this it follows that that @xmath83has the _ disjunction property _ with respect to the intuitionistic disjunction @xmath104 , as shown below .
[ disjunct_prop ] if @xmath105 , then @xmath106 or @xmath107 .
suppose @xmath108 and @xmath109 for some models @xmath110 and @xmath111 and teams @xmath18 and @xmath22 .
let @xmath112 and @xmath113 . by ( [ truth - preserve - operators ] )
, we have @xmath114 hence , by the downward closure property , @xmath115 and @xmath116 , implying @xmath117 . in this paper
, we will also study some interesting fragments of @xmath83(referred to as _ modal dependence logics _ ) defined by restricting the language as follows : * the language of modal dependence logic ( @xmath118 ) : @xmath119 where @xmath27 is an arbitrary classical formula defined recursively as @xmath120 * the language of extended modal dependence logic ( @xmath121 ) : @xmath122 where @xmath85 are classical formulas defined as in the case of @xmath118 . *
the language of modal dependence logic with intuitionistic disjunction ( @xmath123 ) : @xmath124 where @xmath85 are classical formulas defined as in the case of @xmath118 .
* the language of modal intuitionistic dependence logic ( @xmath125 ) : @xmath126 where @xmath85 are classical formulas defined recursively as @xmath127 we leave it for the reader to check that in the presence of intuitionistic connectives dependence atoms are definable , as @xmath128 where @xmath129 , @xmath130 and @xmath131 .
the modality - free fragments of @xmath118 , @xmath123and @xmath125are called _ propositional dependence logic _ , _ propositional dependence logic with intuitionistic disjunction _ and _ propositional intuitionistic dependence logic _ , respectively . these propositional logics were studied in @xcite .
the modality and dependence atom - free fragment of @xmath125is called _ inquisitive logic _ , which was introduced by ciardelli and roelofsen in @xcite and commented also in the context of dependence logic in @xcite .
it is proved in @xcite and @xcite that @xmath118is strictly less expressive than @xmath121 , and the latter has the same expressive power as @xmath123 .
we end this section by pointing out that modal dependence logics are _ not _ closed under _
uniform substitution_. for instance , we have @xmath132 and @xmath133 , whereas @xmath134 and @xmath135 . for this reason , none of the deduction systems of modal dependence logics to be introduced in the next section will have the uniform substitution rule .
in fact , for these logics , _ substitution _ ( being a mapping from the set @xmath6of propositional variables to the set of well - formed formulas ) is not even a well - defined notion , because , for instance , dependence atoms @xmath136 with arbitrary arguments are not necessarily well - formed formulas of the logics . for more details on substitutions of dependence logics ,
we refer the reader to @xcite .
in this section , we introduce deduction systems for the modal dependence logics defined in the previous section and prove the completeness theorems .
we will also introduce a disjunctive normal form for formulas of these logics and show some applications of the normal form ( including deriving a characterization theorem for flat and classical formulas , the interpolation theorem and the finite model property of modal dependence logics ) . in this subsection
, we introduce hilbert - style systems for @xmath125and @xmath83 , and prove the completeness theorems by an argument via the disjunctive normal form of the logics . by expression ( [ dep_def ] ) , dependence atoms are definable in @xmath125 .
let us denote the dependence atom - free fragment of @xmath125by @xmath137 , which turns out to have _ inquisitive logic _ ( @xmath4 ) @xcite as its propositional base .
below we recall the hilbert system of @xmath4defined in @xcite .
we refer the reader to @xcite for further discussions on the connection between inquisitive logic and dependence logics .
the hilbert - style system of inquisitive logic is as follows : axioms : : 1 .
all axioms of intuitionistic propositional logic ( @xmath138 ) , namely 1
. @xmath139 2 . @xmath140 3 .
@xmath141 , @xmath142 4 .
@xmath143 , @xmath144 5 . @xmath145 6 . @xmath146 2
. @xmath147 whenever @xmath27 is a classical formula 3 .
@xmath148 whenever @xmath27 is a classical formula rule : : modus ponens : @xmath149 in @xcite axiom 2 was originally formulated as the kp axiom @xmath150 of kreisel - putnam logic @xmath5@xcite , and axiom 3 is formulated as the atomic double negation law @xmath151 . in other words , the system of @xmath4is @xmath5together with the atomic double negation law @xmath151 ( being not closed under uniform substitution ) .
it can be shown that our presentation above is equivalent to the original presentation in @xcite ( cf .
@xcite ) .
our hilbert - style systems of @xmath83and @xmath125will be extensions of both the system of @xmath4and the system of fischer servi s intuitionistic modal logic @xmath3@xcite .
we refer the reader to @xcite for further discussions on intuitionistic modal logic , and we only remark that @xmath3has intuitionistic propositional logic as its propositional base and adding the law of excluded middle to the logic gives rise to _ classical _ modal logic . in the literature
there are a few ( equivalanet ) variants of the system of fischer servi s intuitionistic modal logic @xmath3 . for the convenience of our argument
, we will use the system defined by plokin and stirling @xcite , which we recall below .
the hilbert - style system of fischer servi s intuitionistic modal logic @xmath3consists of the following axioms and rules : axioms : : 2 + 1 . all axioms of @xmath138 2 . @xmath152 3 .
@xmath153 4 .
@xmath154 5 . @xmath155 6 .
@xmath156 rules : : 1 .
modus ponens : @xmath149 2 .
necessitation : @xmath157 3 .
uniform substitution : @xmath158 now , we present our hilbert - style systems of @xmath83and @xmath125 .
[ hilbert_mt ] * the hilbert - style system of @xmath83is defined as follows : + axioms : : 1 .
all axioms of inquisitive logic @xmath4 2 . all axioms of fischer servi s intuitionistic modal logic @xmath3 3 .
@xmath159 4 .
@xmath160 whenever @xmath27 and @xmath161 are classical modal formulas 5
. @xmath162 6 .
@xmath163 whenever @xmath27 is a classical modal formula 7 .
@xmath164 rules : : 1 .
modus ponens : @xmath149 2 .
necessitation : @xmath157 * the hilbert - style system of @xmath125consists of all of the axioms and rules of the above system of @xmath83except the axioms that involve tensor @xmath165 ( axioms 4 and 5 ) . hereafter within this section , we let @xmath166denote either @xmath83or @xmath125 .
below we derive some interesting clauses in the system of @xmath166that will play a role in the sequel .
[ derivable_thm_mt_mid ] let @xmath167 be arbitrary @xmath166-formulas , and @xmath168 arbitrary classical formulas . 2 1 . [ der_box ] @xmath169 2 .
[ der_dia ] @xmath170 3 .
[ box_dia_inter ] @xmath171 4 .
[ doub_neg ] @xmath172 5 .
[ tensor_comu ] @xmath173 6 .
[ tensor_contr ] @xmath174 7 .
[ tensor_sub ] @xmath175 8 .
[ exm ] @xmath176 the relevant implications in items ( a)-(d ) are either theorems of @xmath3or axioms of @xmath166
. items ( e)-(h ) are derivable from the @xmath166-axiom @xmath177 .
next , we prove the soundness theorem for the systems of @xmath83and @xmath125 . for any @xmath166-formula @xmath55 , @xmath178 @xmath179 @xmath62 .
it suffices to show that all the axioms of @xmath83are valid and all the rules are sound .
we only verify the validity of axiom 2 of @xmath4and axiom 6 .
axiom 2 of @xmath4 : we prove a slightly more general fact that @xmath180 suppose @xmath181
. then @xmath182 and @xmath183 .
thus , there exist @xmath42 such that @xmath184 since @xmath185 is flat and @xmath166has the downward closure property , we have that @xmath186 hence , @xmath187 .
axiom 6 : suppose @xmath188 , where @xmath28 and @xmath27 is a classical formula .
then , for any @xmath26 , we have @xmath189 , i.e. @xmath190 . since @xmath27 is flat , there exists @xmath191 such that @xmath192 .
define @xmath193 clearly , @xmath23 and @xmath194 .
hence , @xmath195 . to prove the completeness theorem for @xmath83and @xmath125
, we will transform every formula into a formula in _ disjunctive normal form _ : @xmath196 where each @xmath197 is a classical formula .
this normal form is a generalization of a similar normal form for @xmath4defined in @xcite , and similar disjunctive normal forms for modal dependence logics without intuitionistic implication were discussed in the literature with a slightly different presentation ( see e.g. , @xcite ) .
let us now introduce our disjunctive normal form for @xmath83as a recursively defined translation @xmath198 for every formula @xmath55 of @xmath83 : base case : : + * @xmath199 when @xmath27 is a classical formula * @xmath200 induction step : : + assume @xmath201 and @xmath202 , where @xmath197 and @xmath203 are classical formulas .
+ * @xmath204 * @xmath205 * @xmath206 * @xmath207 * @xmath208 * @xmath209 the disjunctive normal form for @xmath125is defined the same way as above except that @xmath125does not have the connective @xmath165 in its language . in the next theorem , we show that every formula is provable equivalent to its disjunctive normal form .
[ dn_form_pid ] for any @xmath166-formula @xmath55 , we have @xmath210 .
we only give the proof for @xmath83 , from which the @xmath125case follows .
we proceed by induction on @xmath55 .
the base case is trivial . for the induction step , the cases @xmath211 and @xmath212 follow from the induction hypothesis and @xmath138axioms .
the cases @xmath213 and @xmath214 follow from ( a)(b ) . if @xmath215 , then by the induction hypothesis , we derive in the system @xmath83using axioms 4 and 5 that @xmath216 if @xmath217 , then by the induction hypothesis , we derive in the system @xmath83using axiom 2 of @xmath4and @xmath138axioms that @xmath218 each disjunct @xmath197 of a formula @xmath219 in disjunctive normal form is a classical formula .
we now show that @xmath83and @xmath125prove exactly the same set of classical modal formulas as @xmath73does .
[ class_k_syntax_equiv ] if @xmath27 is a classical formula , then @xmath220 .
it suffices to verify that restricted to classical formulas , @xmath83 , @xmath125and @xmath73have the same axioms and rules .
observe that restricted to classical formulas all the axioms of the systems of @xmath83and @xmath125can be viewed as axiom schemata .
therefore , restricted to classical formulas the systems of @xmath83and @xmath125admit the uniform substitution rule .
as for axioms , we first treat @xmath83 .
it suffices to show that with respect to the disjunction @xmath9 and classical formulas @xmath138axioms , the law of excluded middle @xmath221 and @xmath73axioms are derivable in @xmath83 .
the @xmath138axioms that do not involve disjunction are included in the system of @xmath83 .
it then suffices to derive those axioms that involve disjunction . for @xmath222
, we have @xmath223 , which by axiom 4 implies @xmath224 . applying ( e )
we derive also the axiom @xmath225 . for the axiom @xmath226 , by ( e)-(g ) we derive @xmath227 , which yields the required axiom .
the law of excluded middle is checked in ( h ) , and the @xmath73axiom @xmath228 is included in the system of @xmath83 .
we now derive the remaining axiom @xmath229 of @xmath73 . by ( c )
, we have @xmath230 , which by ( d ) implies @xmath231 .
finally , we show that all classical axioms are derivable also in the system of @xmath125 .
note that in the classical fragment of @xmath125we apply the usual definition @xmath232 for disjunction . since @xmath160 was assumed as an axiom in the system of @xmath83
, one can easily adapt the above proof for @xmath83to a proof for @xmath125 .
we are now ready to prove the completeness theorem for the two systems .
[ completeness_mt ] for any @xmath166-formula @xmath55 , @xmath233 .
it suffices to show the direction `` @xmath179 '' .
suppose @xmath62 .
by lemma [ dn_form_pid ] , @xmath234 , where @xmath235 and each @xmath197 is a classical formula . by the soundness theorem and the disjunction property ( ) , we must have @xmath236 for some @xmath237 . by expression ( [ team2k ] ) , it follows that @xmath238 , which yields @xmath239 by theorem [ class_k_syntax_equiv ] . hence , @xmath178 . in this section ,
we introduce a sound and complete natural deduction system of @xmath240 , a fragments of @xmath83that does not have intuitionistic implication in its language .
our system is an extension of the system of propositional dependence logics defined in @xcite .
[ natrual_deduct_mdor]the natural deduction system of @xmath123is defined as follows : axiom : : rules : : all rules of propositional dependence logic with intuitionistic disjunction @xmath241for @xmath242 as defined in @xcite ( or see table [ tab : pdor ] ) together with the rules for negation , modalities and dependence atoms listed in , where @xmath243 stand for arbitrary formulas and @xmath244 are classical formulas . & & + & & + & + & & + + [ tab : pdor ] + + + + ( @xmath245 ) where the derivation @xmath246 in both rules @xmath247and @xmath248has no undischarged assumptions + + [ tab : mdor_rule ] it can be easily checked that all of the axiom and the rules of the above system are sound , and the clauses listed in the proposition below are derivable .
[ mdbor_derivable_rules ] let @xmath249 be @xmath123-formulas and @xmath27 a classical formula . 2 1 . [ dnegrule ] @xmath250 2 .
[ dstrxvx ] @xmath251 3 .
[ dstrdiav ] @xmath252 4 .
[ distrboxv ] @xmath253 next , we show that every @xmath123-formula is provably equivalent to a formula in disjunctive normal form .
[ nf_mdor ] for any @xmath123-formula @xmath55 , @xmath254 for some set @xmath255 of classical formulas .
we prove the lemma by induction on @xmath55 .
if @xmath55 is a classical formula , then the lemma holds trivially .
if @xmath256 , then we derive @xmath257 the cases @xmath212 and @xmath211 follow easily from the induction hypothesis . the case @xmath258 follows from the induction hypothesis and ( b )
the cases @xmath213 and @xmath214 follow from the induction hypothesis and ( c)(d ) .
[ class_k_equiv_mdor ] if @xmath27 is a classical formula , then @xmath259 .
it is easy to verify that restricted to classical formulas , the axiom and rules of @xmath123are sound in the sense of the usual single - world semantics of classical modal logic .
the direction `` @xmath260 '' then follows from the completeness theorem of the usual classical modal logic @xmath73 . for the direction `` @xmath179 '' , it suffices to show that restricted to classical formulas , all the axioms and rules of the hilbert - style system of @xmath73are derivable in the natural deduction system of @xmath123 . for the rules
, necessitation is a special case of @xmath247 , and restricted to classical formulas , the system of @xmath123admits uniform substitution . for modus ponens
, we derive @xmath261 as follows : for the axioms , restricted to classical formulas , the natural deduction system of @xmath123has the same rules as the natural deduction system of classical propositional logic @xmath74 .
thus all @xmath74axioms are derivable . for axiom 3 of @xmath73
, we derive @xmath262 by ( a ) , @xmath248and @xmath263 . axiom 2 of @xmath73is equivalent to @xmath264 , and we derive @xmath265 by @xmath247 .
[ completeness_mdor ] for any @xmath123-formula @xmath55 , @xmath266 . by a similar argument to that of the proof of , where we apply instead .
we devote this section to three interesting applications of the disjunctive normal form ( [ dnf_mt ] ) of modal dependence logics .
we first prove a characterization theorem for flat and classical formulas .
items ( c ) and ( d ) in the proposition below apply only to those logics that have the relevant formulas in their languages .
[ flat_char ] the following are equivalent .
@xmath55 is flat @xmath267 for some classical formula @xmath27
@xmath268 @xmath269 we only give the detailed proof for ( a)@xmath270(b ) .
assume ( a ) .
then @xmath271 for some set @xmath255 of classical formulas . since @xmath55 is flat , that @xmath272 implies by expression ( [ kp_valid ] ) and the disjunction property ( ) that there exists @xmath273 such that @xmath274
on the other hand , @xmath275 .
hence , @xmath276 .
we write @xmath277 to indicate that the propositional variables occurring in @xmath55 are among @xmath278 . next , we prove craig s interpolation theorem for modal dependence logics that have intuitionistic disjunction in their languages .
let @xmath166be a modal dependence logic that has intuitionistic disjunction in its language .
for any @xmath166-formulas @xmath279 and @xmath280 , if @xmath281 , then there exists an @xmath166-formula @xmath282 such that @xmath283 and @xmath284 .
suppose @xmath285 .
then @xmath286 for some sets @xmath287 and @xmath288 of classical formulas .
then , for each @xmath101 , there exists @xmath289 such that @xmath290 . since @xmath197 and @xmath291 are classical formulas , @xmath292 .
now , by the interpolation theorem of @xmath73 , there exists a classical formula @xmath293 such that @xmath294 and @xmath295 .
thus , @xmath296 and @xmath297 .
the formula @xmath298 is in the language of @xmath166 , and clearly @xmath299 and @xmath300 .
lastly , we prove that modal dependence logics have the finite model property
. [ fmp ] if @xmath301 , then there exists a finite kripke model @xmath56 and finite team @xmath18 such that @xmath302 . for any formula @xmath55
, we have @xmath271 for some finite set @xmath255 of classical formulas . if @xmath301 , then @xmath303 for all @xmath101 .
since each @xmath197 is a classical formula , by expression ( [ team2k ] ) , @xmath304 for each @xmath101 . by the finite model property of @xmath73 , for each @xmath101 , there exists a finite kripke model @xmath305 and @xmath306 such that @xmath307 .
it follows that @xmath308 in the sense of team semantics .
consider the finite model @xmath309 and the finite team @xmath310 . by expression ( [ truth - preserve - operators ] )
, we obtain that for each @xmath101 , @xmath311 , which implies @xmath312 by the downward closure property .
hence , we conclude that @xmath313 , thereby @xmath302 . in this section ,
we define natural deduction systems for @xmath118and @xmath121 .
these systems are generalizations of the system of propositional dependence logic defined in @xcite . to introduce the systems , we follow @xcite and first introduce some notations . a formula in the language of @xmath121or @xmath118is a finite string of symbols
we number the symbols in a formula with positive integers starting from the left , as in the example : [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] if the @xmath314th symbol of a formula @xmath55 starts a string @xmath315 that is a subformula of @xmath55 , we denote the subformula by @xmath316_\phi$ ] , or simply @xmath316 $ ] .
when referring to an occurrence of a formula @xmath317 inside a subformula @xmath315 of @xmath55 , we will be sloppy about the notations and use the same counting also for the subformula @xmath315 .
we write @xmath318)$ ] for the formula obtained from @xmath55 by replacing the occurrence of the subformula @xmath319 $ ] by @xmath161 .
for example , for the formula @xmath320 , we denote the second occurrence of the dependence atom @xmath321 by @xmath322 $ ] , and the same notation also designates the occurrence of @xmath321 inside the subformula @xmath323 .
the notation @xmath324)$ ] designates the formula @xmath325 .
+ + [ tab : dep_rule ] [ natrual_deduct_md ] * the natural deduction system of @xmath121consists of the axioms and rules of @xmath123 ( ) that do not involve @xmath82 , together with the rules for dependence atoms in that are generalized from those in the system of propositional dependence logic @xmath326defined in @xcite . *
the natural deduction system of @xmath118is the same as that of @xmath121except that the dependence atoms can only have propositional variables as arguments .
it is easy to check that the above system is sound . for simplicity
, we will only give the proof of the completeness theorem for the system @xmath121 ; the completeness theorem for the system @xmath118can be proved similarly .
we follow the argument in @xcite for propositional dependence logic , and first define _ realizations _ of formulas , a crucial notion of this argument .
let @xmath327 be a dependence atom .
we call a function @xmath328 a _ realizing function _ for @xmath329 , where we stipulate @xmath330 , and the formula @xmath331 is called a _ realization _ of the dependence atom @xmath329 over @xmath332 .
let @xmath333,\,\dots,\,[d_c , m_c]\rangle$ ] be the sequence of all occurrences of dependence atoms in @xmath55 .
a _ realizing sequence of @xmath55 _ is a sequence @xmath334 such that each @xmath335 is a realizing function for @xmath336 .
we call the classical formula @xmath337 defined as follows a _ realization of @xmath55 _ : @xmath338,\dots,(d_c)_{f_c}^\ast/[d_c , m_c]).\ ] ] for example , consider the formula @xmath339 that we discussed earlier .
the functions @xmath340 are two realizing functions for @xmath321 , and both @xmath341 and @xmath342 are realizing sequences of @xmath55 giving rise to two realizations @xmath343 and @xmath344 of @xmath55 .
the next lemma states the crucial properties of realizations that will be applied in the proof of the completeness theorem .
[ realization_lm ] let @xmath55 be a formula , and @xmath345 the set of all realizing sequences of @xmath55 .
1 . @xmath346 2 .
@xmath347 to prove the above lemma , we first prove a technical lemma .
[ replacement_lm ] let @xmath315 be a subformula of @xmath55 that is not inside the scope of a dependence atom or a negation .
if @xmath348 , then @xmath349)\vdash_{{\ensuremath{\mathbf{md}^+}\xspace}}\phi(\theta/[\psi , m])$ ] .
we prove the lemma by induction on the subformulas @xmath317 of @xmath55 .
the case when @xmath317 is an atom is trivial .
if @xmath350 and without loss of generality we assume that the formula @xmath316 $ ] occurs in the subformula @xmath351 . by the induction hypothesis , @xmath352)\vdash\chi_0(\theta/[\psi , m])$ ] , which , by @xmath353 , implies @xmath352){\ensuremath{\otimes}\xspace}\chi_1\vdash\chi_0(\theta/[\psi , m]){\ensuremath{\otimes}\xspace}\chi_1 $ ]
the case @xmath354 is proved analogously .
the case @xmath355 follows from the induction hypothesis and @xmath248 , and the case @xmath356 follows from the induction hypothesis and @xmath247 .
\(a ) we prove the item by induction on the complexity of @xmath55 .
if @xmath55 does not contain any occurrences of dependence atoms , then the property holds trivially .
if @xmath357 is a dependence atom and @xmath328 a realizing function of @xmath358 , by @xmath359 , to show @xmath360 it suffices to derive @xmath361 .
this is proved by a similar argument to that of the proof of lemma 4.15 in @xcite . if @xmath55 is a complex formula with @xmath362 occurrences of dependence atoms , and @xmath363,\dots,(d_c)_{f_c}^\ast/[d_c , m_c])$ ] , where @xmath334 .
then , since @xmath364 for each @xmath365 , we derive @xmath366,\dots,(d_c)_{f_c}^\ast/[d_c , m_c])\vdash\phi(d_1/[d_1,m_1],\dots,(d_c/[d_c , m_c])\ ] ] by applying repeatedly .
\(b ) the direction @xmath367 follows from item ( a ) and the soundness theorem .
we now prove the other direction @xmath368 by induction on @xmath55 .
the case when @xmath55 is a dependence atom can be easily checked using expression ( [ dep_def ] ) .
the other propositional cases are proved by the same argument as in lemma 4.16 in @xcite .
the case when @xmath213 or @xmath214 follows from the fact that @xmath369 and @xmath370 ( see also ( c)(d ) ) . for any @xmath121-formula @xmath55 , @xmath371 .
we only show the direction @xmath179 " .
suppose @xmath62 . by ( b ) , we have @xmath372 , where @xmath345 is the ( nonempty ) set of all realizing sequences of @xmath55 . by a similar argument to that in the proof of , we obtain @xmath373 for some @xmath374 .
hence , we conclude @xmath375 by ( a ) .
we have seen in the previous sections that team - based modal dependence logic inherit many nice properties from the single - world based usual modal logic .
it is then natural to think that the team semantics of modal dependence logics actually has a single - world semantics interpretation . in this section
, we make this connection precise . in particular
, we will prove that ( the dependence atom - free fragment of ) modal intuitionistic dependence logic can be viewed as a type of ( single - world based ) intermediate modal logic over a certain class of bi - relation intuitionistic kripke models , and the tensor ( disjunction ) @xmath376 of team semantics can be interpreted naturally as a binary diamond modality .
our approach is inspired by the similar interpretation for inquisitive logic given in @xcite . in this section ,
we study the connection between the dependence atom - free fragment of modal intuitionistic dependence logic @xmath137and ( the single - world based ) intermediate modal logics ( i.e. , modal logics between intuitionistic modal logic and classical modal logic ) .
we have defined the hilbert system of @xmath125(and therefore the system of @xmath137 ) as an extension of both fischer servi s intuitionistic modal logic @xmath3and kreisel - putnam logic @xmath5(see ) .
recall that @xmath3is complete with respect to bi - relation intuitionistic kripke frames ( see e.g. @xcite ) and @xmath5is complete with respect to @xmath5-frames ( see e.g. @xcite ) .
we will show in this section that @xmath137is complete ( in the single - world semantics sense ) with respect to a class of kripke models whose frames are both bi - relation intuitionistic kripke frames and @xmath5-frames .
let us first recall relevant definitions for @xmath3 .
\(1 ) at ( 0 , 0 ) @xmath377 ; ( 3 ) at + ( 0,2 ) @xmath378 ; ( 4 ) at + ( 2,2 ) @xmath379 ; ( 2 ) at + ( 2,0 ) @xmath380 ; ( 2 ) ( 4 ) ; ( 1 ) ( 3 ) ; ( 3 ) ( 4 ) ; ( 1 ) ( 2 ) ; at ( 1,-0.5 ) ( f1 ) ; ( 11 ) at ( 5 , 0 ) @xmath377 ; ( 13 ) at + ( 5,2 ) @xmath378 ; ( 14 ) at + ( 7,2 ) @xmath379 ; ( 12 ) at + ( 7,0 ) @xmath380 ; ( 12 ) ( 14 ) ; ( 11 )
( 13 ) ; ( 13 )
( 14 ) ; ( 11 )
( 12 ) ; at ( 6,-0.5 ) ( f2 ) ; * @xmath12 is a nonempty set * @xmath382 is a partial ordering and @xmath383 is a binary relation on @xmath12 * @xmath383 and @xmath382 satisfy the following two conditions ( f1 ) and ( f2 ) ( see figure [ fig_f1f2 ] ) : + f1 : : if @xmath384 and @xmath385 , then there exists @xmath386 such that @xmath387 and @xmath388 .
f2 : : if @xmath389 and @xmath387 , then there exists @xmath390 such that @xmath384 and @xmath391 .
a _ bi - relation intuitionistic kripke model _ is a quadruple @xmath392 such that @xmath393 is a bi - relation intuitionistic kripke frame and @xmath16 is a _ valuation _ satisfying _ monotonicity _ with respect to @xmath382 , that is , @xmath394 and @xmath395 imply @xmath396 .
[ sat_df_bir ] the _ satisfaction _
relation @xmath397 between a bi - relation intuitionistic kripke model @xmath398 , a node @xmath399 and a formula @xmath55 in the language of @xmath3is defined inductively as follows : * @xmath400 iff @xmath394 * @xmath401 * @xmath402 iff @xmath397 and @xmath403 * @xmath404 iff @xmath397 or @xmath403 * @xmath405 iff for all @xmath406 such that @xmath395 , if @xmath407 , then @xmath408 * @xmath409 iff there exists @xmath406 such that @xmath385 and @xmath407 * @xmath410 iff for all @xmath411 such that @xmath412 and @xmath413 , it holds that @xmath407 we will now interpret the team semantics of @xmath137over classical modal kripke models as the single - world semantics over certain bi - relation intuitionistic kripke models .
a crucial observation is that every classical modal kripke model induces a bi - relation intuitionistic kripke model which we shall call _ powerset model_. in a sense , the powerset model @xmath417 carries the information of teams of its associated modal kripke model @xmath56 .
the relation @xmath424 resembles the successor team relation @xmath383 on @xmath12 , but @xmath425 as the successor team relation @xmath383 is a relation between teams ( sets of worlds ) , whereas @xmath424 is a relation between single worlds of @xmath420 .
( 0 , 1 ) ellipse ( 0.5 and 1 ) ; ( 0 , 1 ) ellipse ( 0.2 and 0.6 ) ; ( 3 , 1 ) ellipse ( 0.5 and 1.2 ) ; ( 3 , 1 ) ellipse ( 0.2 and 0.8 ) ; ( 0.2,0 ) ( 2.8,-0.2 ) ; ( 0.2,0.4 ) ( 2.8,0.2 ) ; ( 0.2,2 )
( 2.8,2.2 ) ; ( 0.2,1.6 ) ( 2.8,1.8 ) ; at ( 1.5,-0.5 ) ( f1 ) ; ( 6 , 1 ) ellipse ( 0.5 and 1 ) ; ( 6 , 1 ) ellipse ( 0.2 and 0.6 ) ; ( 9 , 1 ) ellipse ( 0.5 and 1.2 ) ; ( 9 , 1 ) ellipse ( 0.2 and 0.8 ) ; ( 6.2,0 ) ( 8.8,-0.2 ) ; ( 6.2,0.4 ) ( 8.8,0.2 ) ; ( 6.2,2 ) ( 8.8,2.2 ) ; ( 6.2,1.6 ) ( 8.8,1.8 ) ; at ( 7.5,-0.5 ) ( f2 ) ; to see that @xmath417 is indeed a bi - relation intuitionistic kripke model , note that the superset relation @xmath421 is a partial ordering , and the monotonicity of @xmath426 follows from the downward closure property of team semantics . to verify the condition ( f1 ) , for any @xmath427 such that @xmath428 and @xmath422 , letting @xmath429 , it is easy to show that @xmath430 and @xmath431 ( see also figure [ fig_f1f2_mid ] ) .
similarly , to verify the condition ( f2 ) , for any @xmath432 such that @xmath422 and @xmath430 , letting @xmath433 , it is easy to show that @xmath428 and @xmath431 ( see also figure [ fig_f1f2_mid ] ) .
next , we show that the team - based satisfaction relation with respect to classical modal kripke models is equivalent to the single - world - based satisfaction relation with respect to the associated powerset models .
we prove the lemma by induction on @xmath55 .
the only interesting case is when @xmath213 .
if @xmath435 , then @xmath436 , since @xmath437 and @xmath438 .
the induction hypothesis implies that @xmath439 .
hence @xmath440 .
conversely , if @xmath440 , then @xmath439 . for all @xmath441
such that @xmath442 and @xmath443 , since @xmath444 , the downward closure property implies that @xmath445 yielding @xmath446 by the induction hypothesis . hence @xmath435 .
inquisitive logic ( being the propositional fragment of @xmath137 ) is shown in @xcite to be complete with respect to negative saturated ( single - relation ) intuitionistic kripke models , which are also negative @xmath5-models .
let us now give the corresponding definitions in the context of bi - relation intuitionistic kripke models .
a point @xmath377 in a bi - relation intuitionistic kripke model @xmath398 is called an _ @xmath382-endpoint _
iff there is no point @xmath406 such that @xmath395 .
denote by @xmath447 the set of all @xmath382-endpoints seen from @xmath377 , i.e. , @xmath448 a bi - relation intuitionistic kripke frame @xmath381 is said to be _
@xmath382-saturated _ if for every @xmath399 , @xmath449 , and for every nonempty subset @xmath450 , there exists a point @xmath406 such that @xmath395 and @xmath451 .
a model @xmath452 is called _ negative _ if @xmath453 .
we leave it for the reader to verify that a powerset model @xmath418 is a negative @xmath382-saturated model , and in particular , @xmath421-endpoints in @xmath417 are singletons @xmath454 of elements @xmath377 in @xmath12 .
the system of @xmath137(see ) extends the systems of @xmath3and @xmath4with two extra axioms : @xmath455 and @xmath456 , where @xmath27 is any classical formula . by the characterization theorem of classical formulas given in ,
a classical formula @xmath27 is equivalent to a ( double ) negation @xmath457 .
thus , we can rewrite the axiom @xmath456 as @xmath458 . in what follows
we show that the axioms @xmath459 and @xmath458 both characterize certain frame condition .
we write @xmath460 for the _ composition _ of the two binary relations @xmath461 and @xmath462 on a set @xmath12 , defined as @xmath463 \(1 ) at ( -7 , -0.5 ) @xmath377 ; ( 3 ) at + ( -7,1 ) @xmath464 ; ( 4 ) at + ( -7,2.5 ) @xmath465 ; ( 2 ) at + ( -8,1.2 ) @xmath464 ; ( 5 ) at + ( -9,3 ) @xmath466 ; ( 6 ) at + ( -6,1.2 ) @xmath464 ; ( 7 ) at + ( -5,3 ) @xmath380 ; ( 1 ) ( 2 ) ; ( 1 ) ( 6 ) ; ( 4 )
( 5 ) ; ( 4 ) ( 7 ) ; ( 1 ) ( 3 ) ; ( 3 ) ( 4 ) ; ( 2 ) ( 5 ) ; ( 6 ) ( 7 ) ; ( 0 , 0 ) ellipse ( 2.3 and 0.7 ) ; ( -1 , 0 ) ellipse ( 0.55 and 0.3 ) ; ( 1 , 0 ) ellipse ( 0.55 and 0.3 ) ; ( -1.7 , -0.4 ) rectangle ( 1.7 , 0.4 ) ; ( -1 , 2.5 ) ellipse ( 0.8 and 0.3 ) ; ( 1 , 2.5 ) ellipse ( 0.8 and 0.3 ) ; ( -1.9 , 2 ) rectangle ( 1.9 , 2.9 ) ; before we give the proof of the lemma , let us first check that the underlying frames of powerset models satisfy ( g1 ) . indeed , for any three points @xmath474 in a powerset model @xmath418 such that @xmath475 , clearly , @xmath476 is a nonempty subset of @xmath12 such that @xmath477 , @xmath478 and @xmath479 ( see ( b ) ) . as a powerset model
@xmath417 carries the information of teams in the model @xmath452 , condition ( g1 ) can be viewed as a property that is abstracted from the corresponding property of teams of the usual classical modal kripke frames .
suppose @xmath484 satisfies ( g1 ) and @xmath485 for some valuation @xmath60 and some @xmath399 .
then there exists @xmath411 such that @xmath486 , @xmath487 , @xmath488 let @xmath470 be the point given by ( g1 ) . then by the @xmath382-monotonicity , we have @xmath489 , which implies that @xmath490 .
conversely , suppose that @xmath484 does not satisfy ( g1 ) .
then there exists @xmath468 such that @xmath486 , @xmath487 and for all @xmath470 such that @xmath491 , either @xmath492 or @xmath493 .
clearly , we can find a @xmath382-monotone valuation @xmath60 such that @xmath494 for each @xmath470 such that @xmath471 , either @xmath495 or @xmath496 .
thus @xmath497 , thereby @xmath498 . on the other hand , @xmath499 and @xmath500 .
hence @xmath501 . ( 0 , 0)[fill = gray , color = gray ] ellipse ( 2 and 0.6 ) ; ( 0 , 2.5 ) ellipse ( 2.5 and 0.6 ) ; ( 0.4 , 2.5 ) ellipse ( 1.6 and 0.45 ) ; ( 1 ) at ( -1 , 0 ) @xmath464 ; ( 2 ) at ( 0 , 0 ) @xmath464 ; ( 3 ) at ( 1 , 0 ) @xmath464 ; ( 4 ) at ( -1 , 2.5 ) @xmath464 ; ( 5 ) at ( -1.5 , 2.5 ) @xmath464 ; ( 6 ) at ( -0.3 , 2.5 ) @xmath464 ; ( 7 ) at ( 0.4 , 2.5 ) @xmath464 ; ( 8) at ( 1.5 , 2.5 ) @xmath464 ; ( 1 ) ( 4 ) ; ( 1 ) ( 5 ) ; ( 1 ) ( 6 ) ; ( 2 ) ( 6 ) ; ( 2 ) ( 7 ) ; ( 3 ) ( 8) ; ( l4 ) at (-8 , 2.5 ) @xmath464 ; ( l5 ) at ( -8.5 , 2.5 ) @xmath464 ; ( l6 ) at ( -7.3 , 2.5 ) @xmath464 ; ( l7 ) at ( -6.6 , 2.5 ) @xmath464 ; ( l8 ) at ( -5.5 , 2.5 ) @xmath464 ; ( l1 ) ( l4 ) ; ( l1 ) ( l5 ) ; ( l1 ) ( l6 ) ; ( l2 ) ( l6 ) ; ( l2 ) ( l7 ) ; ( l3 ) ( l8 ) ; g2 : : let @xmath399 be an arbitrary point and @xmath504 a set of @xmath382-endpoints such that @xmath507 and for each @xmath508 , there exists @xmath509 with @xmath510
. then , there exists @xmath470 such that @xmath511 ( see ( a ) ) before we give the proof of the lemma , let us first check that the underlying frames of powerset models satisfy ( g2 ) . indeed , for any point @xmath377 and any set @xmath504 of @xmath421-endpoints ( i.e. , a set of singletons of elements in @xmath12 ) in a powerset model @xmath418 such that @xmath512 and for each @xmath508 , there exists @xmath509 with @xmath513 , clearly , @xmath514 is a nonempty subset of @xmath12 such that @xmath515 and @xmath516 ( see ( b ) ) .
suppose @xmath484 satisfies ( g2 ) and @xmath517 for some valuation @xmath60 and some @xmath399 .
then , for each @xmath508 , @xmath518 .
since @xmath380 is an @xmath382-endpoint and @xmath484 is @xmath382-saturated , there exists an @xmath382-endpoint @xmath519 such that @xmath520 and @xmath521 . by ( f2 )
, there exists @xmath386 such that @xmath387 and @xmath522 .
but as @xmath380 is a @xmath382-endpoint , we must have that @xmath523 and @xmath524 .
thus , the set @xmath525 satisfies the condition in ( g2 ) . by ( g2 ) , there exists a point @xmath470 such that @xmath511 hence , @xmath526 .
conversely , suppose @xmath484 does not satisfy ( g2 ) .
then there exists @xmath399 and a set @xmath504 of @xmath382-endpoints satisfying @xmath507 and for each @xmath508 , there exists @xmath509 with @xmath510 , and for all @xmath470 , @xmath527 implies @xmath528 . clearly , we can find a @xmath382-monotone valuation @xmath60 such that @xmath529 . for each @xmath508 , since @xmath530
, we have @xmath518 .
thus @xmath531 . on the other hand
, for each @xmath532 , by the assumption , there exists @xmath533 such that @xmath534 meaning @xmath535 .
hence @xmath536 .
let @xmath537 be the class of all finite negative @xmath382-saturated bi - relation intuitionistic kripke models satisfying ( g1 ) and ( g2 ) .
in the remaining part of this section , we show that the system of @xmath137is complete with respect to @xmath537 , that is , we will prove the following theorem .
the idea of the proof is inspired by that of theorem 3.2.18 in @xcite . note that since @xmath137is not closed under uniform substitution , one can only obtain the completeness theorem in the sense of the theorem below for a class @xmath537 of _ models _ ( with restricted valuations ) instead of a class of _ frames _ ( with arbitrary valuations ) . to prove the other direction `` @xmath179 '' of the above theorem , we first show that every model in @xmath537 can be mapped p - morphically into a finite powerset kripke model .
as p - morphisms are truth - preserving , the required result will then follow .
now , we recall the definition of p - morphisms of bi - relation intuitionistic kripke models given by wolter and zakharyaschev in @xcite .
p1 : : @xmath543 for all propositional variables @xmath544 p2 : : @xmath545 @xmath179 @xmath546 p3 : : @xmath547 @xmath179 @xmath548 p4 : : @xmath549 @xmath179 @xmath550 s.t . @xmath551 and @xmath545 p5
: : @xmath552 @xmath179 @xmath550 s.t . @xmath553 and @xmath547 p6 : : @xmath554 @xmath179 @xmath550 s.t .
@xmath555 and @xmath556 [ bi_mod_p_map_comod ] for every finite bi - relation intuitionistic kripke model @xmath398 in @xmath537 , there exists a finite classical modal kripke model @xmath561 such that there exists a p - morphism @xmath332 of @xmath56into the powerset model @xmath562 induced by @xmath561 .
now , consider the powerset kripke model @xmath567 associated with @xmath561 .
define a function @xmath568 by taking @xmath569 since @xmath56is saturated , @xmath570 for all @xmath399 .
thus @xmath571 and @xmath332 is well - defined .
note that an @xmath382-endpoint @xmath572 of @xmath56is mapped through @xmath332 to the singleton @xmath573 .
intuitively , @xmath382-endpoints of @xmath56are simulated in our argument by singletons of @xmath562 , and it may be helpful for the reader to think of a node @xmath377 of @xmath56as the team formed by all @xmath382-endpoints seen from @xmath377 , namely the set @xmath447 .
it suffices to show that @xmath574 .
the direction `` @xmath179 '' follows from the @xmath382-monotonicity of @xmath60 . for the direction `` @xmath260 '' , if @xmath575 , then since @xmath60 is negative , @xmath576 .
thus , there exists @xmath508 such that @xmath577 , which implies that @xmath578 , thereby @xmath579 .
assume @xmath385 , we show that @xmath582 , namely @xmath583 . for any @xmath584 , by ( f1 ) of @xmath56
, there exists @xmath470 such that @xmath585 for each @xmath586 such that @xmath587 , by ( f2 ) , there exists @xmath588 such that @xmath589 as @xmath590 is an @xmath382-endpoint , we must have @xmath591 and @xmath592 . on the other hand , for any @xmath593 , consecutively applying ( f2 ) and ( f1 ) of @xmath56 , by a similar argument to the above , we can find an @xmath594 such that @xmath595 .
hence , we conclude that @xmath583 .
( p5 ) . if @xmath598 , then @xmath599 .
clearly , @xmath379 is a set of @xmath382-endpoints such that @xmath600 and for each @xmath584 , there exists @xmath601 such that @xmath602 .
thus , by ( g2 ) of @xmath56 , there exists @xmath406 such that @xmath603 as required .
( p6 ) suppose @xmath604 and @xmath605
. then @xmath606 by ( g1 ) of the finite model @xmath56 , there exists @xmath406 such that @xmath607 since @xmath379 is a set of @xmath382-endpoints , the latter of the above implies that @xmath608 .
suppose @xmath609 .
for each @xmath610 , by lemma [ bi_mod_p_map_comod ] , there is a classical modal kripke model @xmath561 and a p - morphism @xmath611 . by the assumption , we have @xmath612 , which implies @xmath613 by . finally , by
, we conclude that @xmath614 , as required .
in the previous section , we have interpreted the intuitionistic connectives and modalities of team semantics as the corresponding logical constants in the single - world semantics setting over special bi - relation intuitionistic kripke models . in this section
, we will also incorporate the tensor @xmath376 of team semantics into the setting discussed in the previous section , and prove a completeness theorem for the system of the dependence atom - free fragment of the modal downward closed team logic @xmath83(i.e .
, the system of @xmath83as defined in without the axioms for dependence atoms , denoted by @xmath615 ) in the single - world semantics sense . the team semantics of tensor is derived from the semantics of the usual disjunction of classical logic . however , in the team semantics the tensor @xmath376 being a disjunction has quite a few odd behaviors , such as , it does not admit the usual elimination rule and distributive rules ; in particular , none of @xmath616 , @xmath617 and @xmath618 is in general true .
indeed , although the tensor inherits properties of disjunction and it behaves truely as a disjunction over classical formulas , it is observed by abramsky and vnnen in @xcite that the tensor @xmath376 of team semantics should rather be understood as a _ multiplicative conjunction _ ( hence the notation @xmath376 ) as in linear logic ( or , in fact , in bunched implication logic @xcite ) .
multiplicative conjunction can often be read as a _ binary diamond _ modality . in what follows
, we will follow this approach and interpret the tensor of team semantics as the usual binary diamond of the single - world semantics over a modified version of powerset models . to define the semantics of tensor as a binary modality
, we will add a ternary relation into the powerset models .
moreover , to characterize the peculiar behavior of tensor as also a type of disjunction , especially to characterize the fact that the constant @xmath619 is the neutral element of the tensor disjunction , i.e. , @xmath620 for all formulas @xmath55 , we will need to include also the empty team in the powerset models .
let us now define this stronger notion of powerset model .
* @xmath623 i.e. , @xmath12 consists of all teams @xmath17 including the empty team @xmath624 * @xmath625 is a ternary relation defined as @xmath626 iff @xmath43 * and the other components are defined as in .
a _ tri - relation intuitionistic kripke frame _ is a quardruple @xmath627 , where @xmath393 is a bi - relation intuitionistic kripke frame and
@xmath628 is a binary relation on @xmath12 satisfying condition ( h1 ) defined below : * @xmath636 iff @xmath377 is an @xmath382-endpoint * @xmath637 iff there exist @xmath638 such that @xmath639 , @xmath640 and @xmath641 * the other cases are defined the same way as the @xmath642 relation in . note that under the above definition the constant @xmath619 does not any more behave as the falsum of the logic ( which is satisfied nowhere in any model ) and the semantics of the negation @xmath643 is changed accordingly . the real falsum will be denoted by @xmath644 and we will not consider this falsum @xmath644 in the present paper . if @xmath258 , then @xmath649 implies that there exist @xmath650 such that @xmath651 , @xmath652 and @xmath653 . since @xmath412 , by ( h1 ) , there exist @xmath654 such that @xmath655 , @xmath656 and @xmath587 . by the induction hypothesis
, we have @xmath657 and @xmath658 .
thus , @xmath659 .
we now show that the team - based satisfaction relation with respect to classical modal kripke models is equivalent to the single - world - based satisfaction relation with respect to full powerset models .
the reader may compare the lemma below with .
we remarked in the previous section that singletons in a powerset model correspond to @xmath382-endpoints in its associated bi - relation intuitionistic kripke models . in a tri - relation intuitionistic kripke model @xmath634
, singletons are simulated by the @xmath382-second least points instead , i.e. , points @xmath377 in @xmath12 such that @xmath395 implies that @xmath664 or @xmath380 is an @xmath382-endpoint .
we denote by @xmath665 the set of all @xmath382-second least points seen from @xmath377 , i.e. , @xmath666 in particular , if @xmath377 is an @xmath382-endpoint , then @xmath667 ; and if @xmath377 is itself a @xmath382-second least point , then @xmath668 .
we say that a tri - relation intuitionistic kripke frame @xmath627 is _
weakly @xmath382-saturated _ if @xmath669 for every non-@xmath382-endpoint @xmath399 , and for every subset @xmath670 , there exists a point @xmath406 such that @xmath395 and @xmath671 .
a model @xmath452 is called _ weakly negative _ if @xmath672 , and for all @xmath382-endpoints @xmath572 , @xmath673 .
let @xmath674 be the class of all finite weakly negative and weakly @xmath382-saturated tri - relation intuitionistic kripke models satisfying ( g1 ) , ( g2 ) with `` @xmath382-endpoints '' and `` @xmath447 '' in the definition replaced by `` @xmath382-second least points '' and @xmath665 " , respectively , and ( h2 ) , ( h3 ) and ( h4 ) defined as follows : a full powerset model @xmath621 has a unique @xmath421-endpoint , namely the empty set @xmath624 .
since @xmath677 for all @xmath544 , by we know that @xmath678 as well
. we leave it for the reader to verify that all the other conditions of @xmath674 are satisfied by the powerset model @xmath621 of any finite modal kripke model @xmath452 , i.e. , @xmath679 .
the proof of the above theorem goes through a similar argument to that of . the direction
@xmath260 " follows from and the finite model property of @xmath615 ( ) .
the other direction @xmath179 " will follow from and to be stated and proved in the remainder of this section .
@xmath260 " : suppose @xmath684 for some @xmath685 . for any point @xmath572 such that @xmath686 and @xmath687 , @xmath572 is an @xmath382-endpoint implying @xmath688 .
it follows that @xmath689 . since @xmath690 , we conclude @xmath691 .
`` @xmath179 '' : suppose @xmath692
. then , there exists @xmath693 such that @xmath694 and @xmath695 .
the latter implies that @xmath466 is not an @xmath382-endpoint .
since @xmath452 is weakly @xmath382-saturated , @xmath696 . pick an element @xmath697 . by the monotonicity of @xmath382 , we have @xmath689 implying @xmath698 , as required .
we prove the theorem by induction on @xmath55 .
we only check the interesting cases ; the other cases are proved as in or see @xcite .
the case @xmath646 follows from ( q3 ) .
the case @xmath712 is proved by applying ( q1 ) and ( q2 ) . [ tri_p_map_bullet ] for every finite tri - relation intuitionistic kripke model @xmath452 in @xmath674 , there exists a finite classical modal kripke model @xmath561 such that there exists an endpoint preserving p - morphism @xmath332 of @xmath56into the full powerset model @xmath713 induced by @xmath561 .
we show that @xmath332 is an endpoint - preserving p - morphism . the conditions ( p2 ) and
( p4 ) are verified by a similar argument to that in the proof of taking into account the fact that @xmath447 is replaced by @xmath665 in this proof .
we now give the proof of the other conditions .
we show that @xmath717 . if @xmath572 is an @xmath382-endpoint , since @xmath452 is weakly negative , @xmath673 .
we also have @xmath718 , and @xmath719 by the empty team property and .
if @xmath377 is not an @xmath382-endpoint , applying , the condition is proved by a similar argument to that in the proof of .
assume @xmath385 , we show that @xmath720 , namely @xmath721 .
if one of @xmath377 and @xmath380 is an @xmath382-endpoint , then by ( h3 ) , both @xmath377 and @xmath380 are @xmath382-endpoints , which implies that @xmath722 . by definition , @xmath723 ,
i.e. , @xmath721 .
now , assume that both @xmath377 and @xmath380 are not @xmath382-endpoints . by a similar argument to that in , for any @xmath724 , by ( f1 ) of @xmath56 , there exists @xmath470 such that @xmath585 for each @xmath725 such that @xmath587 , by ( f2 ) , there exists @xmath588 such that @xmath589 but since @xmath590 is a @xmath382-second least point , either @xmath726 or @xmath727 is an @xmath382-endpoint . in the latter case
, we conclude from ( h3 ) that @xmath728 is also an @xmath382-endpoint , which is a contradiction .
thus , the former is the case , and @xmath592 . on the other hand , for any @xmath729 , by a similar argument
, we can find an @xmath730 such that @xmath595 .
hence we conclude that @xmath721 .
suppose @xmath731 . if one of @xmath665 and @xmath379 is the empty set , then @xmath732 by the definition of @xmath733 .
since @xmath452 is weakly @xmath382-saturated , @xmath377 must be an @xmath382-endpoint , which by ( h4 ) implies that @xmath734 .
also , clearly , @xmath735 , as required . if @xmath736 , then the condition follows from a similar argument to that in the proof of .
suppose @xmath737 .
if @xmath738 , then let @xmath380 be any @xmath382-endpoint such that @xmath395 .
clearly , @xmath739 . by ( h4 ) ,
@xmath740 , and thus @xmath487 .
if @xmath741 , then the condition follows from a similar argument to that in the proof of .
suppose @xmath744 .
then @xmath745 .
since @xmath452 is weakly @xmath382-saturated , there exist @xmath411 such that @xmath746 , @xmath747 and @xmath748 .
it follows that @xmath705 , @xmath551 . since @xmath742 , by ( h2 )
, we obtain @xmath629 .
if @xmath572 is an @xmath382-endpoint , then @xmath749 , which is the unique @xmath421-endpoint of @xmath713 .
conversely , if @xmath377 is not an @xmath382-endpoint , then since @xmath452 is weakly @xmath382-saturated , we have @xmath750 , i.e. , @xmath751 is not an @xmath421-endpoint .
ciardelli , i. , `` dependency as question entailment '' , pp . 129182 , in _ dependence logic : theory and application _ , edited by samsom abramsky and juha kontinen and jouko vnnen and heribert vollmer , progress in computer science and applied logic , birkhauser , 2016 .
ebbing , j. , l. hella , a. meier , j .- s .
mller , j. virtema , and h. vollmer , `` extended modal dependence logic '' , pp . 126137 , in _
20th international workshop , wollic 2013 , proceedings _ , volume 8071 of _ lecture notes in computer science _ , springer berlin heidelberg , 2013
. ebbing , j. , and p. lohmann , `` complexity of model checking for modal dependence logic '' , pp . 226237 , in _ sofsem _ , edited by mria bielikov and gerhard friedrich and georg gottlob and stefan katzenbeisser and gyrgy turn , volume 7147 of _ lecture notes in computer science _ ,
springer , 2012 .
ebbing , j. , p. lohmann , and f. yang , `` model checking for modal intuitionistic dependence logic '' , in _ post - proceedings of the ninth tbilisi symposium on language , logic and computation _ , volume 7758 of _ lecture notes in computer science _ , pp . 231256 , 2013 .
hodges , w. , `` some strange quantifiers '' , pp . 5165 , in _ structures in logic and computer science : a selection of essays in honor of a. ehrenfeucht _ , edited by j. mycielski and g. rozenberg and a. salomaa , volume 1261 of _ lecture notes in computer science
_ , london : springer , 1997 .
kontinen , j. , j .- s .
mller , h. schnoor , and h. vollmer , `` a van benthem theorem for modal team semantics '' , in _
24th eacsl annual conference on computer science logic ( csl 2015 ) _ , edited by stephan kreutzer , pp . 277291 , 2015 .
mller , j .- s . , and h. vollmer , `` model checking for modal dependence logic : an approach through post s lattice '' , in _
20th international workshop ( wollic 2013 ) proceedings _ , edited by leonid libkin and ulrich kohlenbach and ruy de queiroz , volume 8071 of _ lncs_. springer berlin heidelberg , 2013 . vnnen , j. , `` modal dependence logic '' , pp . 237254 , in _
new perspectives on games and interaction _ , edited by krzysztof r. apt and robert van rooij , volume 4 of _ texts in logic and games _ , amsterdam university press , 2008 .
wolter , f. , and m. zakharyaschev , `` intuitionistic modal logic '' , pp .
227238 , in _ logic and foundations of mathematics _ , edited by a. cantini , e. casari and p. minari , synthese library , kluwer , 1999 . | in this paper , we introduce sound and complete deduction systems for various modal dependence logics
. we also study model - theoretic properties of these team - based modal logics by interpreting them as variants of single - world - based intuitionistic modal logics .
_ dependence logic _ is a logical formalism , introduced by vnnen @xcite , that captures the notion of _ dependence _ in social and natural sciences . the modal version of the logic is called _ modal dependence logic _ and was introduced in @xcite .
modal dependence logic extends the usual modal logic by adding a new type of atomic formulas @xmath0 , called _ dependence atoms _ , to express dependencies between propositions , and by lifting the usual single - world semantics to the so - called _ team semantics _ , introduced by hodges @xcite
. formulas of modal dependence logic are evaluated on sets of possible worlds of kripke models , called _
teams_. intuitively , a dependence atom @xmath0 is true if within a team the truth value of the proposition @xmath1 is _ functionally determined _ by the truth values of the propositions @xmath2 .
research on modal dependence logic and its variants has been active in recent years .
basic model - theoretic properties of the logics were studied in e.g. , @xcite , and a van benthem theorem for the logics was proved in @xcite .
the expressive power and the relevant computation complexity problems of the logics were investigated extensively in e.g. , @xcite . in this paper
, we study a few problems that have not yet been well addressed in the literature , particularly the axiomatization problem and the comparison of team semantics and the single - world semantics .
sano and virtema gave in @xcite sound and complete hilbert systems and label tableau calculi for modal dependence logic and its extended version . however , their axiomatization did not cover the modal dependence logics with intuitionistic connectives , especially with intuitionistic implication . in this paper
, we will define hilbert systems for these intuitionistic variants of modal dependence logic .
these systems are extensions of the systems of fisher servi s intuitionistic modal logic @xmath3@xcite and inquisitive logic @xcite .
we also introduce a natural deduction system for ( extended ) modal dependence logic by extending the system of propositional dependence logic given in @xcite .
moreover , we will discuss the disjunctive form of modal dependence logics , which is essentially known in the literature , and apply the normal form to prove the interpolation theorem and the finite model property of these logics .
team semantics is a natural generalization of the usual single - world semantics . in this paper
, we will also interpret the team semantics of modal dependence logics as the usual single - world semantics over certain kripke models .
it was proved in @xcite that inquisitive logic @xmath4(which is the propositional base of modal dependence logic with intuitionistic implication ) can be viewed as a weak kreisel - putnam s intermediate logic @xmath5@xcite , and @xmath4is complete ( in the usual single - world semantics sense ) with respect to the class of negative intuitionistic kripke models of @xmath5 .
building on this result we will prove in this paper that modal dependence logics can be viewed as certain intermediate modal logics , and they are complete ( in the usual single - world semantics sense ) with respect to a class of bi - relation or tri - relation intuitionistic kripke models .
these bi - relation models are special bi - relation intuitionistic kripke models of fisher servi s intuitionistic modal logic @xmath3 , and the tri - relation intuitionistic kripke models are endowed with an extra ternary relation interpreting a binary diamond . in particular , we will interpret the tensor ( disjunction ) connective of team semantics as a binary diamond modality under the single - world semantics .
this idea is inspired by @xcite , where tensor is read as a multiplicative conjunction .
this paper is organized as follows . in section 1
we recall basics of modal dependence logics . in section 2
we study the axiomatization problem for modal dependence logics , and also prove a few fundamental properties of the logics , including the interpolation theorem and the finite model property .
section 3 provides single - world semantics interpretation of modal dependence logics .
preliminary results of this paper were included in the author s dissertation @xcite . |
we consider the numerical solution of initial value problems for large systems of ordinary differential equations ( odes ) , @xmath1 with given vector - valued function @xmath2 , given initial vector @xmath3 , and unknown vectors @xmath4 ( for @xmath5 ) .
our interest in this paper lies in systems ( [ ode ] ) that arise from semi - discretization of initial - boundary value problems for two - dimensional convection - diffusion equations possessing a mixed spatial - derivative term , @xmath6 here @xmath7 and @xmath8 denote a given real vector and a given positive semi - definite real matrix , respectively .
a main application area of equations of the kind ( [ pde ] ) is financial option pricing theory , where mixed derivative terms @xmath9 arise naturally since the underlying brownian motions are usually correlated to each other .
extensive details and examples of financial applications are given in , for example , the references @xcite .
for the numerical solution of semi - discrete problems ( [ ode ] ) , splitting schemes form an effective and popular means , cf .
e.g. @xcite .
this paper is devoted to the analysis of a recent splitting scheme of the alternating direction implicit ( adi ) type that has been tailored so as to deal with equations possessing a mixed derivative term .
let @xmath10 be a given fixed parameter .
assume the right - hand side function @xmath2 is decomposed into a sum @xmath11 where @xmath12 represents the contribution to @xmath2 stemming from the mixed derivative term , and @xmath13 ( for @xmath14 ) represents the contribution to @xmath2 stemming from all spatial derivative terms in the @xmath15-th spatial direction .
let @xmath16 be a given time step and define temporal grid points by @xmath17 ( @xmath18 ) .
we consider the following splitting scheme for ( [ ode ] ) , generating in a one - step fashion successive approximations @xmath19 to @xmath20 : 0.5@xmath21 method ( [ mcs ] ) is called the _ modified craig sneyd ( mcs ) scheme_. it has recently been introduced , in a slightly more general form , by in t hout & welfert @xcite .
taylor expansion yields that the mcs scheme has classical order of consistency equal to two for any value @xmath0 .
the mcs scheme can be viewed as an extension of the second - order craig
sneyd ( cs ) scheme proposed in @xcite .
the latter scheme , called `` iterated scheme '' in loc .
cit . , is equivalent to ( [ mcs ] ) with parameter value @xmath22 .
a perusal of ( [ mcs ] ) shows that the @xmath12 term is always treated _ explicitly _ , whereas the @xmath23 and @xmath24 terms are treated _
implicitly_. more precisely , the mcs scheme starts with an explicit euler step applied to the full system ( [ ode ] ) which is succeeded by two implicit corrections corresponding to each of the two spatial directions .
subsequently , an explicit update is performed , which is followed again by two implicit , unidirectional corrections .
accordingly , the mcs scheme retains the well - known key advantage of adi schemes over standard implicit methods , such as the crank nicolson scheme , that the ( linear or nonlinear ) systems to be solved in each time step are much easier to handle .
the adaptation of adi schemes to convection - diffusion equations with mixed derivative terms has been studied by a number of authors .
several stability results , in the sense of von neumann , have been obtained .
mckee et al .
@xcite considered a simpler version of ( [ mcs ] ) , which is equivalent to the first two lines with @xmath25 .
this basic scheme , also known as the douglas scheme , is of order one for any value @xmath0 in the presence of a mixed derivative term .
mckee et al
. showed that if @xmath22 , then it is unconditionally stable when applied to a standard finite difference ( fd ) discretization of ( [ pde ] ) .
next , craig & sneyd @xcite formulated the second - order cs scheme and proved that this scheme is unconditionally stable in the case of ( [ pde ] ) with @xmath26 . recently in t hout & welfert @xcite extended the above stability results in various ways .
we state here the main results pertinent to the situation at hand .
firstly , for the cs scheme unconditional stability was proved @xcite in the general case of ( [ pde ] ) . secondly
, it was shown @xcite that in the case of ( [ pde ] ) with @xmath26 the mcs scheme ( [ mcs ] ) is unconditionally stable whenever @xmath27 .
up to now it is an important open question when the mcs scheme , with @xmath28 , is unconditionally stable in the application to general equations ( [ pde ] ) , i.e. , with arbitrary @xmath29 and positive semi - definite @xmath30 . as it turns out , an analysis of this is not straightforward , related to the fact that the eigenvalues of the semi - discrete linear operators move from the real line in the pure diffusion case to the complex plane in the general , convection - diffusion case . in the present paper
we shall arrive at positive results on the above question . for the stability analysis
we consider the linear scalar test equation @xmath31 with complex constants @xmath32 ( @xmath33 ) .
when applied to ( [ test ] ) , the mcs scheme ( [ mcs ] ) reduces to the scalar iteration @xmath34 with @xmath35 ( @xmath33 ) and @xmath36 where we use the notation @xmath37 the iteration ( [ uns ] ) is stable if latexmath:[\[\label{stabconds } in the von neumann framework , the @xmath32 represent eigenvalues of the linear operators @xmath13 that are obtained after semi - discretization , on a uniform spatial grid , of the convection - diffusion equation ( [ pde ] ) with constant coefficients and periodic boundary condition .
corresponding to the positive semi - definiteness of the diffusion matrix @xmath30 , it was shown in @xcite ( cf .
also sect .
[ conclusion ] ) that for standard fd discretizations the following condition on the scaled eigenvalues @xmath39 is fulfilled , @xmath40 ) when applied to equations ( [ pde ] ) with mixed derivative terms is that @xmath41 holds whenever @xmath42 is satisfied .
an outline of the rest of this paper is as follows . in sect .
[ stabresults ] we study for which parameter values @xmath0 the implication @xmath43 is fulfilled .
four cases are investigated , depending on whether @xmath44 is real or complex valued and whether @xmath45 are ( both ) real or complex valued . in sect .
[ conclusion ] the results of sect . [ stabresults ] are applied and discussed relevant to an application of the mcs scheme ( [ mcs ] ) to ( [ pde ] ) .
let @xmath46 denote the imaginary unit .
in this section we study the stability requirement @xmath43 .
the following introductory result gives a criterion on @xmath0 for the case @xmath47 .
this is pertinent to the situation where no mixed derivative term is present in ( [ pde ] ) .
[ thm1 ] there holds @xmath48 for all @xmath49 with @xmath50 , @xmath51 if and only if @xmath52 .
the rational function @xmath53 has no poles in the set @xmath54 and therefore attains its maximum on the boundary of this set .
thus assume @xmath55 , @xmath56 with @xmath57 .
we have @xmath58 and @xmath59
. then @xmath60 and it follows after some algebraic manipulations that @xmath61 v^4.\ ] ] hence , @xmath62 if and only if @xmath63 which is equivalent to @xmath52 .
0.3 cm in @xcite the stability of adi schemes for pure diffusion equations with mixed derivatives was analyzed .
this concerns the case where all @xmath39 are real - valued . for the mcs scheme and two spatial dimensions ,
the following criterion on @xmath0 was obtained .
[ thm2 ] there holds @xmath64 whenever @xmath65 satisfy @xmath42 if and only if @xmath66 .
see ( * ? ? ?
0.3 cm in most applications , also a convection term is present .
accordingly , one is led to considering complex - valued @xmath45 .
the next theorem gives a necessary condition on @xmath0 for this situation .
[ thm3 ] suppose @xmath64 for all @xmath67 and @xmath68 satisfying @xmath42
. then @xmath69 .
the result is obtained by a taylor expansion at the point @xmath70 .
we take @xmath71 and @xmath72 where @xmath73 and @xmath74 with @xmath75 .
this choice was found to be convenient after numerical experimentation . inserting into ( [ s ] ) and using @xmath76 ( @xmath77 )
, it follows that @xmath78 this yields @xmath79 the right - hand side is bounded by 1 for @xmath75 only if @xmath80 .
hence , it must hold that @xmath69
. 0.3cmbased on strong numerical evidence ( see sect .
[ conclusion ] ) we conjecture that the condition on @xmath0 in theorem [ thm3 ] is also sufficient , but a proof is currently lacking . the above results dealt with real - valued @xmath44 .
the two subsequent theorems concern arbitrary , complex - valued @xmath44 . a preliminary result is [ le1 ] let @xmath81 be given . if @xmath82 and @xmath83 whenever @xmath84 with @xmath85 , then @xmath86 .
consider the function @xmath87 defined by @xmath88 there holds @xmath89 ^ 2 + [ a\sin(2\varphi)+b\sin(\varphi)]^2\\ & = a^2+b^2+c^2 + 2ab \cos(\varphi)+2bc\cos(\varphi)+2ac\cos(2\varphi ) .
\end{split}\ ] ] one readily verifies that @xmath90 , @xmath91 , @xmath92 and hence @xmath93 using that @xmath94 whenever @xmath95 , proves the assertion . 0.3cmfor the case where @xmath45 are real - valued , we obtain the following necessary lower bound on @xmath0 .
numerical experiments indicate that this bound is sufficient as well .
[ thm4 ] suppose @xmath96 for all @xmath97 and @xmath98 satisfying @xmath42
. then @xmath99 . setting @xmath100 and @xmath101 , we can write @xmath102 let @xmath103 . since for all @xmath104 with @xmath105 we have @xmath106
assume @xmath107 .
then @xmath108 and it is easily seen that @xmath109 . therefore lemma @xmath110 can be applied and , using @xmath111 , this leads to the necessary condition @xmath112 denote @xmath113 .
then @xmath114 next , after some computations , there follows @xmath115 by ( [ nec_con ] ) , we arrive at @xmath116 the right - hand side is a rational function of @xmath117 , which is readily seen to have a global maximum at @xmath118 . inserting this value yields the lower bound @xmath119 .
the final result in this section concerns the most general case , where all @xmath39 are complex - valued . to derive this result
we employ a lemma from @xcite pertinent to the condition @xmath42 . for completeness
, its concise proof is included here .
[ le2 ] if @xmath68 with @xmath50 , @xmath51 , then @xmath120 define the vectors @xmath121 \left|1+\theta z_j\right|/\sqrt{2\theta}\\ \end{array } \right ) , \quad j=1 , 2.\ ] ] their euclidean norms are @xmath122 next , their standard inner product is @xmath123 applying the cauchy
schwarz inequality gives @xmath124 which concludes the proof .
0.3cmfor the most general case , we have the following positive result : [ thm5 ] if @xmath125 , then @xmath126 whenever @xmath127 satisfy @xmath42 .
the expression ( [ sform ] ) for @xmath128 yields @xmath129 by invoking lemma [ le2 ] it follows that @xmath130 is bounded from above by @xmath131 we can write @xmath132 define @xmath133 then it follows that latexmath:[\[\label{ups } let @xmath125 .
we prove that the right - hand side of ( [ ups ] ) is bounded by @xmath135 for all @xmath136 , @xmath137 .
first note that @xmath138 ( @xmath139 ) and therefore it suffices to consider @xmath140 .
let @xmath141 $ ] be fixed but arbitrary and define @xmath142 ( @xmath139 ) . for the function
@xmath143 it is readily verified that @xmath144 this directly implies that @xmath143 , and hence @xmath145 , is nonincreasing on @xmath146 $ ] . for the function @xmath147 a more elaborate computation shows @xmath148 r \sin\varphi.\ ] ] in view of @xmath149 we find that also @xmath147 , and hence @xmath150 , is nonincreasing on @xmath146 $ ] .
consequently , it is sufficient to prove that the right - hand side of ( [ ups ] ) is bounded by @xmath135 whenever @xmath136 , @xmath151 .
write @xmath152 $ ] .
one easily verifies that @xmath153 inserting this and rearranging terms , it follows that the upper bound ( [ ups ] ) is ( in fact ) equal to @xmath135 whenever @xmath136 , @xmath151 .
0.3cmnumerical evidence leads to the conjecture that the conclusion of theorem [ thm5 ] is valid for all @xmath99 , i.e. , under the ( necessary ) lower bound of theorem [ thm4 ] .
a proof of this does not appear to be straightforward .
we note that in the proof above the assumption @xmath154 is used in an essential manner .
in this section we discuss an application to convection - diffusion equations ( [ pde ] ) .
we semi - discretize on the unit square @xmath155\times [ 0,1]$ ] by using central second - order fd schemes on a cartesian grid with mesh widths @xmath156 and @xmath157 in the @xmath158 and @xmath159 directions , respectively : [ xy ] ( u_x)_i , j & & + ( u_y)_i , j & & + ( u_xx)_i , j & & + ( u_yy)_i , j & & + ( u_xy)_i , j & & + & & + .
here @xmath160 denotes a real parameter with @xmath161 and we use the notation @xmath162 .
we note that the right - hand side of ( [ xy]e ) is the most general form of a second - order fd approximation of the mixed derivative @xmath9 based on a centered 9-point stencil .
when @xmath163 , it reduces to the well - known 4-point formula @xmath164 assuming constant coefficients and a periodic boundary condition for ( [ pde ] ) , the above fd discretization yields a splitted , semi - discrete system ( [ ode ] ) , ( [ splitting ] ) where @xmath165 for @xmath166 with constant matrices @xmath167 .
the matrix @xmath168 represents the cross derivative term in ( [ pde ] ) and @xmath169 , @xmath170 represent the spatial derivatives in the @xmath158 and @xmath159 directions , respectively .
the periodicity condition implies that the @xmath167 are kronecker products of circulant ( thus normal ) matrices that commute with each other , and are therefore simultaneously diagonalizable by a unitary matrix .
hence , stability can be rigorously analyzed by considering the scalar test equation ( [ test ] ) with @xmath32 eigenvalues of @xmath167 ( @xmath33 ) .
this is equivalent to a von neumann stability analysis . by inserting discrete fourier modes
, it follows that the scaled eigenvalues @xmath39 are given by [ zzz ] z_0&=&(d_12+d_21)b[-_1_2+(1-_1)(1-_2 ) ] , + z_1&=&-2d_11a_1(1-_1)+ic_1q_1_1 , + z_2&=&-2d_22a_2(1-_2)+ic_2q_2_2 , where @xmath171 the angles @xmath172 are integer multiples of @xmath173 ( @xmath174 ) where @xmath175 , @xmath176 are the dimensions of the grid in the @xmath158 and @xmath159 directions , respectively . using the positive semi - definiteness of the diffusion matrix @xmath30
, an elementary calculation shows @xcite that @xmath177 fulfill the condition ( [ cond ] ) , independently of @xmath178 , @xmath156 , @xmath157 . upon invoking theorem [ thm5 ]
the following neat stability result is obtained for the mcs scheme applied to ( [ pde ] ) .
[ th3.1 ] consider equation ( [ pde ] ) with positive semi - definite matrix @xmath30 and periodic boundary condition .
let the semi - discrete system ( [ ode ] ) , ( [ splitting ] ) be obtained after fd discretization and splitting as described in this section .
then the mcs scheme ( [ mcs ] ) is unconditionally stable when applied to ( [ ode ] ) , ( [ splitting ] ) whenever @xmath179 .
moreover , this conclusion remains valid when any other stable fd discretizations for @xmath180 , @xmath181 are used in place of ( [ xy]a ) , ( [ xy]b ) .
the last part of theorem [ th3.1 ] follows directly from the fact that the real parts of the new eigenvalues @xmath182 , @xmath183 are always smaller than those of ( [ zzz]b ) , ( [ zzz]c ) , respectively , and hence , ( [ cond ] ) remains true .
an inspection of ( [ zzz]a ) yields that the eigenvalues @xmath44 have the property that their imaginary part is identically equal to zero .
accordingly , it is of particular interest to know all parameter values @xmath0 such that the stability requirement @xmath43 holds for just real - valued @xmath44 .
theorem [ thm5 ] provides the sufficient condition @xmath179 , whereas theorem [ thm3 ] yields the necessary condition @xmath69 .
next , we remark that the mcs scheme has recently been applied successfully in @xcite to actual convection - diffusion equations ( [ pde ] ) with mixed derivative terms using the parameter value @xmath184 .
this seems to be surprising , as this value was determined @xcite for pure diffusion equations ( [ pde ] ) and it clearly does not satisfy the necessary condition @xmath69 for equations with convection .
we note that reasons for choosing a smaller @xmath0 in the mcs scheme are a reduced error constant and better damping properties compared to the original cs scheme , see @xcite . under ( [ cond ] ) with @xmath67.,title="fig:",scaledwidth=95.0% ]
+ theoretical results on the latter two issues are not known at this moment . to gain insight ,
we have performed a numerical experiment .
let @xmath185 and @xmath186 for @xmath187 denote independent , uniformly distributed random numbers in @xmath155 $ ] and consider random triplets @xmath188 given by @xmath189 then ( [ cond ] ) holds and @xmath67 .
for each @xmath190 with @xmath191 we computed the maximum value of @xmath192 over two million points @xmath188 above .
the outcome is displayed in figure [ maxabs ] .
figure [ maxabs ] reveals the intriguing result that the estimated maximum value of @xmath193 is very close to @xmath135 whenever @xmath27 .
for @xmath194 we arrive at a maximum value of @xmath195 .
additional experiments in this case suggest that @xmath193 is larger than 1 for a limited set of points @xmath188 , and at most 1 under only a slightly stronger condition on @xmath44 than in ( [ cond ] ) . because of these observations , it is very plausible that the mcs scheme performs well in actual applications to ( [ pde ] ) , also with convection , already when @xmath184 .
subsequently , an examination of the obtained numerical results indicates that @xmath196 for all @xmath69 .
this supports our conjecture formulated below theorem [ thm3 ] . in view of the above ,
it is likely that the condition on @xmath0 in theorem [ th3.1 ] can be relaxed to @xmath69 , and next , that a slightly modified version of theorem [ th3.1 ] holds under the ( weaker ) assumption @xmath27 . in future research
we intend to study these issues theoretically .
this work has been supported financially by the research foundation flanders , fwo contract no .
g.0125.08 .
k. j. in t hout & b. d. welfert , _ unconditional stability of second - order adi schemes applied to multi - dimensional diffusion equations with mixed derivative terms _ , appl .
num . math .
* 59 * ( 2009 ) 677692 . | the modified craig sneyd ( mcs ) scheme is a promising splitting scheme of the adi type introduced by in t hout & welfert [ _ appl .
num .
math .
* 59 * ( 2009 ) _ ] for multi - dimensional pure diffusion equations having mixed spatial - derivative terms . in this paper
we investigate the extension of the mcs scheme to two - dimensional convection - diffusion equations with a mixed derivative . both necessary and sufficient conditions on the parameter @xmath0 of the scheme
are derived concerning unconditional stability in the von neumann sense .
* keywords : * initial - boundary value problems , convection - diffusion equations , method - of - lines , adi splitting schemes , von neumann stability analysis , fourier transformation . *
ams subject classifications : * 65l04 , 65l05 , 65l20 , 65m12 , 65m20 , 91g60 . |
topological solitons play an important role in many areas of physics .
they can be beneficial , like the abrikosov vortices in type ii superconductors or unwanted like the dislocations in nematic liquid crystals .
likewise , the creation and observation of topological solitons in bose - einstein condensates is nowadays routine in laboratories around the world .
recent experimental observations have renewed interest in the types of models that can support topological solitons , and it is important to understand the basic properties and phenomenology of topological solitons in these models . theoretically , the possibility of the presence of skyrmion structures in certain materials has been known for long@xcite ( for a more recent work and references , see @xcite ) , and this has recently been experimentally verified at least in mnsi@xcite and fe@xmath0co@xmath1si @xcite .
a newer theoretical discovery is the family of materials collectively known as topological insulators ( for a review , see @xcite ) .
the experimental observation of topological insulator phase in bi@xmath0sb@xmath2 @xcite , bi@xmath3se@xmath4@xcite and bi@xmath3te@xmath4@xcite confirmed these theoretical predictions and served to further add relevance to the basic research involving topological solitons .
for the present work , perhaps the most interesting possibility is the discovery that frustrated magnetic materials may support topological insulator phases , where the wave - function in momentum - space is classified by the hopf invariant@xcite . as will be discussed below , when the system is described by the hopf charge
, it will support knotted structures .
knotted structures have a long history in physics .
they were first considered by lord kelvin , who proposed in 1867 @xcite that atoms could be knotted tubes of ether .
this idea did not yield a satisfactory atomic theory , but subsequently more realistic models have been proposed with potential for knotted structures .
this has been done , for example , in the context of ferromagnets@xcite , bose - einstein condensates@xcite , and optics@xcite .
a unifying feature of all these , in addition to the knotted structures , is that all these phenomena and their knots can be described by classical field theory . in a field
theoretical description the knot is not made of a rope in empty space but through the twisting of a globally defined vector field . here
we will only consider the case in which the knottedness is characterized by the hopf charge .
the minimal model with stable knotted structures seems to be the faddeev - skyrme model @xcite and it is believed that this model is contained within various models with immediate physical application , such as the two - component ginzburg - landau model modified with an additional ward like term . @xcite . the physical carrier field in the faddeev - skyrme model is a smooth 3d unit vector field @xmath5 with @xmath6 .
unit vector fields with such properties have been proposed to exist , e.g. , in super - fluid @xmath7he in its a - phase ( continuous unlocked vortex)@xcite and in ferromagnets@xcite . in order to be able to define the hopf charge it is also necessary that the vector field @xmath8 approaches the same value at all asymptotic directions : @xmath9 when @xmath10 ( asymptotic triviality ) .
asymptotic triviality means that from the point of view of @xmath8 the euclidean space @xmath11 is topologically like the sphere @xmath12 and this allows one to define the hopf charge as an element of the homotopy class @xmath13 .
a concrete hopfion vortex ring with hopf charge @xmath14 is given by @xmath15 where @xmath16 , however , this configuration is not a solution for the equation of motion studied here .
it is easy to see that @xmath17 at infinity in any direction and also on the @xmath18-axis , while the preimage of @xmath19 , which is defined as the vortex core , is the ring @xmath20 .
( vortex core ) ; blue line : the preimage of @xmath21 .
the lines have been assigned compatible directions and their linking number can then be computed as explained in text .
[ f : apu],width=226 ] since the core is just a simple loop it does not explain the hopf charge . for that purpose
we need to consider also the preimage - curve of some other value of @xmath8 and the linking of these two curves .
let us for example consider the tubular iso - surface defined by @xmath22 and on it the curve that is the preimage of @xmath23 , see figure [ f : apu ] .
the linking number of this curve with the core - line can now be computed using the standard rules of knot theory : if the top arrow can be aligned with the bottom arrow using a clockwise rotation , the crossing is assigned @xmath24 , for a counterclockwise rotation + 1 .
the linking number is then @xmath25 of the sum of signed crossings . from figure
[ f : apu ] we can see that there are two crossings of signature @xmath24 and therefore the hopf charge is @xmath14 .
the hopf charge can also be computed by integrating a charge density : for the antisymmetric field @xmath26 one constructs the potential @xmath27 such that @xmath28 and then the hopf charge is given by @xmath29 a proof that these two methods give the same result is given in @xcite . for the dynamics we use a variation of the skyrme model
as proposed by faddeev in 1975@xcite , it is defined by the lagrangian @xmath30 d^3x,\quad f_{\mu\nu}:={\phi\cdot\partial_\mu \phi\times\partial_\nu \phi}.\ ] ] note that in the static case we have the following scaling property : under the scaling @xmath31 the integrated kinetic term scales as @xmath32 while the integrated @xmath33 term scales as @xmath34 . thus according to the virial theorem nontrivial configurations
will attain some fixed size determined by the _ dimensional _ coupling constant @xmath35 . in the computations reported below we used @xmath36 .
a very important property of this model is that the energy is bounded from below@xcite and from above@xcite @xmath37 where @xmath38 are some constants and @xmath39 the hopf charge .
since the exponent of the charge is less than 1 it follows that configurations with higher hopf charge tend to form bound states rather than split into several charge 1 states .
an intriguing question is the form of the minimum energy states with higher hopf charge .
after some tentative works@xcite detailed taxonomy has been obtained by battye and sutcliffe@xcite and hietarinta and salo@xcite .
the results follow very well@xcite the bound ( [ vk ] ) .
it turns out that a twisted torus - like configuration ( similar to ( [ q1n ] ) ) is a minimum energy state only for @xmath40 @xcite , for higher charges linked or knotted configurations are found .
the trefoil knot is persistently found for charge @xmath41 from a wide variety of initial states , even for long vortices.@xcite we will next give the computational method and then describe the results for two dynamical studies : scattering head on collisions , and attractive and repulsive channel simulations inspired by the work of ward in @xcite .
for scattering dynamics we used the lagrangian ( [ lag ] ) in the full 3 + 1-dimensional minkowski space - time ( with @xmath42 ) and the resulting euler - lagrange equations , derived with a lagrange multiplier to guarantee @xmath43 during evolution . the second order ( in time )
equation was split into two first order equations for numerical computations .
the spatial derivatives are discretized on @xmath44 points .
the formulae are quite similar to those in @xcite .
time evolution was computed using 4th order runge - kutta ( five stage low memory version@xcite ) .
for scattering studies the computational lattice varied between @xmath45 and @xmath46 , but all the simulations used the same lattice constant @xmath47 . in the three - hopfion collisions the physical size of the grid was @xmath48 and in the double - collisions @xmath49 .
the constant of courant - friedrichs - lewy like condition linking time and space discretization is @xmath50 .
ward s channel simulations were computed in a lattice of @xmath46 unless otherwise noted ; the grid constant was still @xmath47 but @xmath51 .
the courant instability was controlled by smoothing . for computational platform
we used samrai@xcite .
as our work progressed the significance of boundary condition became more and more clear and thus we have tried a variety of different methods and parameter configurations to minimize boundary errors .
the presented two- and three - hopfions scatterings are calculated with dirichlet boundary conditions @xmath52 , and the ward s channel simulations with neumann boundary condition @xmath53 .
furthermore , in the two- and three - hopfion scatterings we had a thin and mild absorption layer near the boundary in order to prevent radiation from reflecting and re - entering the scattering region .
visualization of a vector field is problematic . since we can not plot the vector field at all points of the 3d space we chose to plot iso - surfaces determined by the value of the third component @xmath54 , often with @xmath55 , which is a narrow tube around the core .
this tube is painted@xcite with color determined by the first two components of @xmath8 . for visualization platform we used visit@xcite .
before analyzing specific scattering processes it is important to recall the possible elementary deformation processes . it was already noted above that it is not enough to consider only the core - line but also how the unit - vector field twists around it .
in fact , the proper knot theoretical setting is to use _ framed links _ which can be realized as _ directed ribbons _
( pre - images of line segments@xcite ) .
this adds local information near the curve , e.g. , twisting around it .
knot deformations correspond to ribbon deformations , which allow certain types of crossing and breaking , in which the hopf charge will nevertheless be conserved .
examples of ribbon deformations can be seen in@xcite .
we will now consider the scattering of low - charge `` un - knot '' hopfions in two different situations : three - body scattering on a fixed target with projectiles from left and right , and two - body scattering with impact parameter .
the the individual hopfions ( minimum energy states ) were first created using the steepest descent method ( corresponding to 1st order dissipative dynamics with a fictitious time ) @xcite .
these hopfions were then rotated , lorentz boosted and combined to form the initial configurations for scattering .
this was done using interpolation routines provided by the scipy python library . in the first scattering process , for which snapshots are given in figure [ f1 ] , with full details in http://www.youtube.com/watch?v=zkhctrp6s_y[movie s1 ]
, we have two charge @xmath56 hopfions approaching a stationary @xmath57 target head on , from left and right .
the moving hopfions were lorentz boosted numerically to speeds @xmath58 and @xmath59 , respectively , and then immersed into the computational box . since the total charge is @xmath60 we expect the result to be a trefoil knot .
this is indeed the case , after several intermediate steps .
the resulting trefoil is not stationary but rather oscillates and rotates , and eventually the hopfion settles down drifting slowly .
since @xmath61 there is theoretically 24% excess energy , some of which will radiate as low amplitude , close to vacuum ( @xmath62 ) , waves from the scattering region .
these are not visible at the iso - surface @xmath55 level .
the resulting drift is probably due to momentum conservation in an asymmetric radiation background .
hopfions with velocities @xmath63 on a stationary @xmath57 target . for more details see http://www.youtube.com/watch?v=zkhctrp6s_y[movie s1].[f1],width=332 ] in figures [ f4],[f5 ] ( for details see http://www.youtube.com/watch?v=3x2ns8akble[movies s2 ] and http://www.youtube.com/watch?v=ywwla0sdaqo[s3 ] ) we have a situation similar to the one in figure [ f1 ] , except that now the central target has charge @xmath64 .
the total charge is then @xmath14 and indeed the final state has one hopfion of @xmath14 .
as the hopfions approach each other they deform by the rule `` like colors attract '' , thereby preparing for ribbon crossings . since @xmath65 most of the energy has to radiate away .
this can be clearly seen in figure [ f5 ] ( http://www.youtube.com/watch?v=ywwla0sdaqo[movie s3 ] ) which describes in further detail the later stages of the process . in this figure
we have indicated the distribution of energy density by a white cloud . up to @xmath66
the energy is localized near the vortex tube but when the deformation towards @xmath14 continues more radiation is released and can be seen to radiate away . and velocities @xmath67 , respectively , resulting with a hopfion of charge @xmath24 .
for details see http://www.youtube.com/watch?v=3x2ns8akble[movie s2 ] .
[ f4],width=332 ] but with further intermediate states for the later stage of the deformation process .
here we have added the energy density plot which appears as a grayish halo .
see also http://www.youtube.com/watch?v=ywwla0sdaqo[movie s3 ] .
[ f5],width=332 ] for the two - body scattering with impact parameter we present the scattering of two @xmath57 hopfions with two different impact parameters . in figure [ f2 ]
s4 ) ] we have the impact parameter @xmath68 and the hopfions have speed @xmath69 .
although the hopfions touch briefly at @xmath70 the speed and distance prevent a bound state from forming .
the individual hopfions just continue tumbling along a slightly bent trajectory . for the same speeds but with slightly smaller impact parameter value @xmath71 the result
is entirely different ( figure [ f3 ] , http://www.youtube.com/watch?v=jqidqwaqjow[movie s5 ] ) . for a moment
it seems that the hopfions would again continue along their individual trajectories but there is just enough time to form an elongated loop .
after this the evolution is typical for the total charge @xmath72 case : the loop behaves like an over - twisted rubber band and proceeds to make one ribbon crossing deformation to reach the linked loop configuration that is standard for this hopf charge .
the linked configuration continues to rotate and oscillate . .
the initial velocities were @xmath73 and the impact parameter @xmath68 . for more details
see http://www.youtube.com/watch?v=nrw1z81i4hq[movie s4 ] .
[ f2],width=332 ] except that impact parameter is @xmath71 . for more details
see http://www.youtube.com/watch?v=jqidqwaqjow[movie s5 ] .
[ f3],width=332 ] the above results illustrate the difficulty associated with trying to make any generalizations about the behavior of hopfions .
however , some progress has been made by modeling hopfions with elastic rods@xcite .
it would be interesting to see if the elastic rod model could be extended to approximate dynamical features as well .
we will next consider the case of two unit charge hopfion unknots placed on static positions not too far from each other .
this case was studied by ward @xcite , who observed ( using energy minimization ) that under symmetry preserving relaxation an axisymmetric two hopfion state reaches a local minimum ( @xmath74 ) where the soliton cores around the symmetry axis are separated by a nonzero but finite distance .
he also noticed that this separated configuration has @xmath75 higher total energy than the twice wound multisoliton state ( @xmath76 ) both of which are in the same homotopy class .
the nature of the axisymmetric state is not yet clear and , as ward stated , more study is needed to investigate whether the state is actually a local minimum or just a saddle - point .
for these cases we prepared the initial states using wards proposal , that in the stereographic coordinates @xmath77 , a good ansatz for a @xmath78 hopfion is given by @xmath79 , where @xmath80 and @xmath81 is some specific function of @xmath82 ( see equation ( 4 ) in @xcite , ward uses @xmath83 ) .
actually this can be generalized by adding some signs in the expression : @xmath84 here the sign of @xmath85 determines whether this is a hopfion or anti - hopfion , while the signs of the @xmath86 s can be changed by @xmath87 rotations around one of the coordinate axes .
thus for a single hopfion the signs @xmath88 are irrelevant but they do play a role when we construct an initial state with two hopfions . ] one can not extract @xmath85 in order to say whether it describes a hopfion or an anti - hopfion . ]
it was noted by ward @xcite that a good two hopfion ansatz is obtained just by adding the corresponding @xmath89 functions .
thus for the case with two hopfions located at @xmath90 and sharing the same axis ( ward s channel a ) we can take @xmath91 as the starting configuration . here
@xmath92 .
ward uses a dipole pair to describe the configuration , in our notation @xmath93 and @xmath94 , so that @xmath95 , etc
. furthermore the angle between @xmath96 and @xmath97 is 0 if @xmath98 and @xmath87 if @xmath99 .
the signs in can be changed by various @xmath87 rotations and this way the 16 possible sign combinations @xmath100 can be reduced to a canonical 6 .
they are 1 ) @xmath101 ( @xmath102 s to same direction , angle @xmath103 ) , 2 ) @xmath104 ( @xmath102 s to same direction , angle @xmath87 ) , 3 ) @xmath105 ( @xmath102 s away from each other , angle @xmath103 ) , 4 ) @xmath106 ( @xmath102 s away from each other , angle @xmath87 ) , 5 ) @xmath107 ( @xmath102 s towards each other , angle @xmath103 ) , 6 ) @xmath108 ( @xmath102 s towards each other , angle @xmath87 ) . the time evolution for these cases is as follows . for cases 1 and 2
the system is in the attractive channel ( figure [ f7 ] , http://www.youtube.com/watch?v=erzeu3v0wak[movie w1 ] ) .
the hopfions approach each other , bounce a few times , during which they develop some asymmetry and eventually turn over , join and deform into a @xmath109 single ring hopfion .
the four remaining cases 3 , 4 , 5 and 6 are repulsive : as the hopfions recede they slowly turn around their axis and it is plausible that they eventually turn enough to attract each other and return to form a single ring state .
that is , we know that due to the fact @xmath110 the asymptotically separated configuration is a higher energy state than the single - ring configuration but it is unclear whether the initial configurations for these four repulsive systems have large enough repulsive ( escape ) energy to form the higher energy state . as the hopfions quickly reach the border of any finite computational space ,
this is a hard question to answer using numerical methods alone .
produces an attractive interaction between the two @xmath111 hopfions .
the pair oscillates multiple times before a slight asymmetry brings the system to the more stable @xmath112 single - ring configuration . for details
see http://www.youtube.com/watch?v=erzeu3v0wak[movie w1 ] , and the discussion in the end of section [ stacked hopfions ] .
[ f7],width=453 ] in order to better understand the effect of boundaries , we considered the @xmath104 case further and calculated its evolution in physical grids of different size , while keeping the grid constant and boundary conditions unchanged .
our results are given in http://www.youtube.com/watch?v=tgc4_tca1x0[movie w2 ] , which combines the result from cubes of physical size @xmath113 , @xmath114 , @xmath115 , and @xmath116 having @xmath45 , @xmath46 , @xmath117 , and @xmath118 grid points , they are located at top left , top right , bottom left and bottom right , respectively . in all cases
the simulation starts with a slight repulsion during which the colors of the isosurfaces rotate towards the same phase .
then the hopfions approach each other and bounce several times and eventually form a @xmath56 un - knot hopfion .
the difference in the dynamics can be explained by the existence of two competing effects : the fixing of the boundary value to @xmath119 tends to stabilize the configuration , while the radiation reflecting from the boundaries tends to un - stabilize it .
the radiation effects are illustrated in figure [ f : err ] , where we have plotted the values of the usually very revealing quantity @xmath120 on a slice of the box .
the box in upper cases is relatively small and the closeness of the fixed boundary stabilizes the situation well and many bounces are required to un - stabilize it , even though there is lots of radiation bouncing around . the fastest un - stabilization occurs with box size @xmath117 ( lower left ) , in that case radiation effects overwhelm the boundary stabilization . in the largest box ( lower right )
the radiation takes a longer time to reach the interaction region and therefore more bounces are again possible before deformations .
clearly the major difference between these simulation is the time at which the radiation wave of the initial relaxation travels to the boundary and is reflected back to center of the grid . in smaller grids
the boundary is so close that the first reflection happens before the hopfions even collide the first time .
after a few more reflections an interference pattern is formed in the background but it is reasonably symmetric . in the larger grids
the first reflected wave passes over the hopfion pair during the first couple of oscillations and this asymmetric perturbation causes the state to unwind faster .
nevertheless the final result is always the same slices of the canonical derivative field @xmath121 between a fixed interval @xmath122 $ ] at the same moment ( @xmath123 , @xmath124 ) in different physical size grids .
we can see that the boundary effects are strongly present at the area around the hopfion s core in the first case , just reaching it in the second and third case and still outside the observed area in the fourth case .
[ f : err],width=453 ] another case considered by ward is that of hopfion rings side by side on the @xmath125-axis .
the ansatz is now @xmath126 here
in this case also the 16 possible sign combinations can be divided into 6 essentially different ones , to which others are related by rotations : 1 ) @xmath101 , polarization vectors @xmath102 parallel , 2 ) @xmath106 , antiparallel , 3 ) @xmath104 , parallel , 4 ) @xmath128 , parallel , 5 ) @xmath105 , antiparallel , 6 ) @xmath107 .
cases 3 and 4 ( and cases 5 and 6 ) are related by an overall sign change , which can not be generated by a rotation .
considering hopfions asymptotically as dipole pairs and studying the forces between them led ward to introduce the second attractive channel which corresponds to our case 3 above .
ward s energy minimization gave an axially - symmetric ring for @xmath109 .
this is believed to be the true global static minimum .
@xcite our numerical simulations show that the attractive and repulsive channels yield the same qualitative behavior as in the stacked case .
however , now the attractive channel comprises of initial configurations 3 , 4 , 5 , and 6 while the repulsive channel contains 1 and 2 .
as an example of the attractive channel , we give case 5 in figure [ f10 ] , http://www.youtube.com/watch?v=fudsiyc8o0c[movie w3 ] )
. seems to be attractive without delay and the pair end up in the familiar @xmath112 single ring configuration . at the end of the simulation the pair oscillates relatively softly having already emitted most of its excess energy . for more detail
see http://www.youtube.com/watch?v=fudsiyc8o0c[movie w3 ] .
[ f10],width=453 ] is repulsive from the start of the simulation and the system end up in the unattached state where the initial hopfions advance in opposite directions .
[ f11],width=453 ] in order to study the repulsive case further we chose to change somewhat the rotation angle of the initial state . after rotating the second hopfion by angle @xmath129
the ansatz can be written as @xmath130 as can be seen the dependent variable @xmath131 is invariant under rotation .
recall that the configurations @xmath105 and @xmath106 were respectively in the attractive and repulsive channels .
now let us fix @xmath86 configuration to repulsive @xmath106 with @xmath132 , then the attractive @xmath105 is obtained from with @xmath133 .
it has already been observed that even in the repulsive case the hopfions turn around and could return if the computational box would be big enough . in order to study a case with only slight repulsion we chose the configuration @xmath106 for @xmath134 and found it to be attractive ( figure [ f12 ] and http://www.youtube.com/watch?v=icjajn21pzc[movie w4 ] ) while @xmath135 was still repulsive ( http://www.youtube.com/watch?v=ud_gaiuijhg[movie w5 ] ) .
similarly , for states @xmath104 and @xmath101 : @xmath101 is still repulsive at @xmath136 but changes to attractive at @xmath137 .
it seems the attractive channels yield qualitatively the same behavior regardless of whether the configuration is stacked or side - by - side and likewise for the repulsive case .
the study of the dependence on the initial rotation angle , however , shows that there are quantitative differences , of which we have uncovered one : the repulsive channel contains a different range of initial relative orientations depending on the parameters @xmath86 and @xmath138 .
this further illustrates the extremely rich set of phenomena associated with hopfions .
changes after a rotation of the second hopfion by an angle of @xmath139 .
now the hopfions move further away from each others under a color rotation until the phase difference between them becomes attractive and the single ring configuration @xmath112 is formed . for more detail
see http://www.youtube.com/watch?v=icjajn21pzc[movie w4 ] ) .
[ f12],width=415 ]
now that we are close to seeing knotted structures in experiments it is important to understand their dynamics .
the static knots in the faddeev - skyrme model are already well understood and in this paper we have shown that these objects have also dynamically many particle - like properties .
we have here studied the scattering collisions of low charge hopfions . in general the results
show remarkable particle - like behavior .
we have verified again their tendency to form bound states , due to the charge - energy dependence @xmath140 ; we have also seen the associated radiation of the released energy . with suitable speeds and impact parameters we can also have grazing scattering , without bound state formation .
since the internal structure of hopfions is quite complicated the results of the scattering process depend strongly on the initial configuration , and not only on the initial positions and velocities but also on the internal specifics ( such as orientation , spin , and relative phase ) of the scattering objects@xcite .
we have also investigated the dynamical stability of static configuration of two stacked or side - by - side unknots discussed by ward @xcite . as expected their evolution depends sensitively on computational aspects due to the mere length of the calculation ( due to the initial state being static ) .
the errors arise both from discretizations and in particular from boundary effects of the finite computational box .
it seems that the stacked and side - by - side configurations made of two charge 1 hopfions are all unstable and eventually relax into a charge 2 hopfion .
this work is supported by academy of finland grants 47188 and 123311 , and by the uk engineering and physical sciences research council .
the computations were done using the resources provided by the _ csc
it center for science ltd _ , www.csc.fi and by the _ m - grid project _ , supported by the academy of finland , and by _ csc it center for science ltd_. | several materials , such as ferromagnets , spinor bose - einstein condensates , and some topological insulators , are now believed to support knotted structures .
one of the most successful base - models having stable knots is the faddeev - skyrme model and it is expected to be contained in some of these experimentally relevant models .
the taxonomy of knotted topological solitons ( hopfions ) of this model is known . in this paper
we describe some aspects of the dynamics of hopfions and show that they do indeed behave like particles : during scattering the hopf charge is conserved and bound states are formed when the dynamics allows it .
we have also investigated the dynamical stability of a pair of hopfions in stacked or side - by - side configurations , whose theoretical stability has been recently discussed by ward . |
x - ray plasma ejections ( xpes ) are sudden expulsions of hot magnetized plasma in the solar corona seen in x - rays . they establish a wide range of macroscopic motions showing different morphology , kinematics and physical conditions .
xpes occur usually during the impulsive phase of flares , but their connection with other solar - activity phenomena like : coronal mass ejections ( cmes ) , prominences , radio bursts , coronal dimmings , global waves is also known .
there are some restrictions in calling any motions in the corona around the flare times as xpes .
the restrictions regard the size , duration , brightness , speed , etc . and
are introduced mainly by spatial , temporal and spectral resolutions of imaging instruments and their operational schemes .
xpes have been systematically observed since 1991 when _ yohkoh _ satellite began to operate .
they became commonly known since the paper written by @xcite was published .
however , we note earlier articles on essentially the same phenomena from the _ solar maximum mission _ @xcite and from _ yohkoh _ @xcite . until now images recorded by the
_ yohkoh _ soft x - ray telescope , sxt @xcite are the largest database of xpes , even though the newer solar x - ray imaging instruments operate , e.g. , _ goes _ solar x - ray imager , _ reuven ramaty high - energy solar spectroscopic imager _
( _ rhessi _ ) , _ hinode _ x - ray telescope .
detailed analyses of individual xpes were performed first by @xcite and @xcite . in these papers
the authors determined values of physical parameters describing an xpe using temperature and emission measure maps obtained from sxt images .
the maps allowed them to investigate overall magnetic configuration including a flare loop and a reconnection region .
they also used hard x - ray light curves , derived by the _ yohkoh _
hard x - ray telescope , hxt @xcite , for a detailed description of reconnection timing .
@xcite made the first attempt to correlate xpes and cmes . for 17 well - observed limb flares they found that flares associated with cmes
show xpes and opposite flares not associated with cmes also lack xpes
. a more extensive investigation of association between xpes and flares was performed by @xcite . for 57 well - observed limb flares they found that almost 70% show xpes .
they also reported dependence on x - ray class , namely the association is larger for stronger flares , but it could be caused by observational biases . to investigate interesting examples of xpes other _
yohkoh _ instruments also have been used , namely the hxt @xcite and the bragg crystal spectrometer , bcs @xcite . in both papers
a special location of investigated events has been chosen .
these xpes occurred far behind the solar limb and due to their fast expansion they came into the view of an instrument before brighter flares , which expand slower .
it is virtually the only way for using full - sun instruments like the bcs to resolve faint soft x - ray emission of xpes .
the behind - the - limb location also protects against strong emission of footpoint hard x - ray sources of flares , which usually dominate fainter coronal emission .
the obtained results proved that an xpe can conatin energetic non - thermal electrons @xcite and superhot thermal plasma @xcite .
an important progress in investigation of xpes gave a trilogy made by @xcite .
they investigated systematically sxt observations obtained during a two - year interval and found 137 xpes .
the events were a subject of multipurpose analysis
the authors introduced a morphological classification of xpes , investigated their kinematics , specified the association with flares and cmes .
the present name of xpes also comes from these papers .
we recapitulate the results of kim et al . in details in further sections , where we compare them to our results .
more recently , an association between xpes and radio events and prominences has been investigated . in statistical surveys @xcite studied type ii radio bursts , whereas @xcite studied drifting pulsating structures ( dps ) .
both surveys suggest a kind of connection between xpes and radio events but further examinations are needed to establish the connection .
the relationship between hot ( xpes ) and cold ( prominences ) ejections was discussed by @xcite for single events , which was followed up with statistical studies @xcite . finally , in our short review illustrating a research progress we would like to recall the two following papers .
firstly , the results of a quantitative analysis of sxt images describing time evolution of basic physical parameters for 12 xpes were given by @xcite .
secondly , after extensive analysis of a complex xpe that consisted of several recurrent episodes , @xcite reported a close connection between sequential ejections and successive hard x - ray bursts .
the most commonly accepted physical explanation of xpes connects these phenomena directly with flare magnetic reconnection .
@xcite regarded xpes as a proof of the presence of plasmoids driven by magnetic reconnection occurring above a soft x - ray loop in short - duration , compact - loop flares similar to the _ canonical _ 2d cshkp model ( * ? ? ? * and references therein ) , which was proposed for long - term , two - ribbon flares .
in this way , @xcite postulated a unification of two observationally distinct classes of flares , i.e. two - ribbon flares and compact - loop flares , by a single mechanism of magnetic reconnection called the _ plasmoid - induced - reconnection model_. the first qualitative studies of individual events @xcite reported that the measured velocities of xpes are much smaller than the velocity of reconnection outflow expected from the model to be about the alfv@xmath0n speed . to reconcile this discrepancy
the authors suggested : ( 1 ) the high density of the xpes , ( 2 ) the time evolution effect ( i.e. , the plasmoid should be accelerated as it propagates , thus the investigated xpes have not yet reached the maximum velocity ) , or ( 3 ) an interaction with coronal magnetic fields overlying the xpes .
although more recently , 2d resistive - mhd numerical simulations of the reconnection explain kinematical properties of various observational features attributed to the current - sheet plasmoids @xcite , it has been expected that 3d reconnection renders a more realistic description of eruptive phenomena .
for example , @xcite suggested 3d quadrupolar reconnection of two loop systems that appear to exchange their footpoints as a result of loop - loop interaction @xcite . on the other hand , in some cases the xpes seem to play the same role as phenomena called precursors of cmes @xcite .
this opinion is supported observationally by common kinematical evolution of xpes and cmes @xcite as well as their morphological resemblance @xcite .
if so , loss of equilibrium or mhd instability , commonly accepted as one of the cme triggering mechanism @xcite , also should be taken seriously into consideration as a cause of xpes .
reports concerning xpe observations in the sxt database were scattered until now across many different sources : refereed articles , conference communications , electronic bulletins , etc .
an exception was the survey given by @xcite , which includes almost all the xpes associated with limb flares for a two - years interval .
our motivation was to ingest all available reports in one catalogue and organize them in a uniform way for an easy usage .
we have examined sxt images in those time intervals , in which any systematic searches of xpes did not perform .
knowledge about xpes has so far been shaped by a limited number of events that have repeatedly appeared in the literature .
our catalogue is meant to serve as a convenient tool for every scientist who wants to better understand the nature of xpes .
the catalogue contains the all xpes we know that were observed by the sxt during the entire _ yohkoh _ operations , i.e. between 1991 october 1 and 2001 december 14 .
there are three main surveys of events that we used in our catalogue : 1 .
@xcite , which contains 137 limb events , observed between 1999 april and 2001 march .
2 . ohyama ( 2009 , private communication ) with 53 limb events that occurred between 1991 october and 1998 august .
the survey was prepared for the aim of statistical research @xcite , but it was not published .
3 . @xcite , which reports 113 events , observed basically within two time intervals : 1998 september 1999 march and 2001 april 2001 december that were not systematically searched before .
we also incorporated 65 xpes reported in other scientific papers as well as in the electronic bulletin _ yohkoh sxt science nuggets_. keeping in mind the examination of sxt images made by different authors , we can conclude that the list of xpes associated with limb flares ( defined as @xmath1 , where @xmath2 is heliographic longitude ) is almost complete . on the other hand ,
the list of xpes associated with disk flares is largely incomplete , with the exception of time intervals examined by @xcite .
occasional reports of xpes not associated with any flares @xcite teach us that the sxt images made without any flares should be also examined and this work still awaits to be done . in summary ,
our catalogue contains 368 events .
time frequency of xpes occurrence during the _ yohkoh _ mission is given in fig .
[ occ ] , where sizes of bins are 6 or 3 months for years 1991 - 1997 and 1998 - 2001 , respectively .
this traces variability of general solar activity .
a larger occurrence rate in cycle 23 in comparison with cycle 22 may be attributable to the revision of the sxt flare mode observing sequences .
indeed , the ratio of the number of xpes and that of flares taken from the _ solar - geophysical data _ ( sgd ) is 3.6 times greater for solar cycle 23 than for cycle 22 . before 1997
, a routine scheme of observations during the flare mode was dominated by images in which the exposure time and the position of the field of view ( typically [email protected] arcmin@xmath4 ) were automatically adjusted by the signals and locations of the brightest pixels .
xpes are distinctly fainter and located higher in the corona than flares .
thus , they had usually too poor statistics during short exposure times and due to a fast expansion they left immediately the narrow field of view . under these circumstances , xpes were rarely well - observed . in 1997 ,
the frequency of images with sufficiently long and constant exposures and broader field - of - view ( 5.2 @xmath3 5.2 arcmin@xmath4 and 10.5 @xmath3 10.5 arcmin@xmath4 ) was increased to every 10 - 20 s @xcite .
this observational scheme worked more favorably for the xpes identification , however flare structures seen in those images often suffer from heavy saturation that manifests itself as vertical spikes disturbing a picture of xpes .
we registered events to the catalogue on the basis of the sxt observations exclusively .
for this reason , we omitted some x - ray ejections from years 1991 - 2001 identified using observations made with other instruments alone , like the hxt , e.g. @xcite .
the online catalogue resides at http://www.astro.uni.wroc.pl/xpe/catalogue.html since 2010 october 22 .
it is also linked from the _ yohkoh _ legacy data archive @xcite .
the general arrangement of the catalogue as a matrix of years and months of observation is presented in fig .
[ mentry]a . after clicking the month ,
each xpe is identified by a chronological catalogue number , date , and time of occurrence .
the letter ( a ) added to a start time means that the xpe began earlier than shown in the available movies . the letter ( b )
added to an end time means that the xpe finished later than shown in the available movies .
the letter ( c ) added to an end time means a time interval of available movies in which we can not identify the xpe reported earlier by other authors .
each events has links to five entries that provide detailed information on the xpe , flare ( sxr and hxr ) , cme , and references on the xpe ( see an example in fig . [
mentry]b ) .
this entry contains 8 columns labeled as follows : ( 1 ) event i d , ( 2 ) date , ( 3 ) time , ( 4 ) quality , ( 5 ) classification , ( 6 ) movies , ( 7 ) results of analysis , and ( 8) references ( see an example in ( fig . [
mentry]c ) .
the first three columns are replicated from the higher entry . in col .
( 4 ) , we indicate the quality of available sxt observations by assigning one letter between ( a ) and ( d ) .
the letter ( a ) means the highest quality : an xpe is clearly seen and only slightly disturbed by flare saturation , observations have almost full spatial and time coverage , images are made by at least two different filters . in conclusion ,
events with this letter are a good source for any kind of quantitative analysis including plasma diagnostics on the basis of the filter - ratio method @xcite .
the letter ( b ) also means quite good quality of observations , but the usage of only a one filter in some cases makes a plasma diagnostics unavailable .
nevertheless , xpes marked with this letter are always good for kinematical studies .
the letter ( c ) means poor quality for some of the following reasons : the brightness of the xpe only marginally above the background , short observation window , inadequate field - of - view , or strong effect from flare saturation . for events with this letter
only limited analyses are usually possible , e.g. a description with our 3-parameters classification . the letter ( d )
is designed for xpes , which were mentioned by other authors but whose presence is not confirmed in the movies that we made . in col .
( 5 ) we characterize general observational features of xpes using a new classification scheme that we have developed in this catalogue . in our classification
we define three criteria considering : ( a ) morphology of an xpe , ( b ) its kinematics , and ( c ) recurrence . examining each criterion
we distinguish two subclasses of events only : ( a ) 1 collimated , 2 loop - like ; ( b ) 1 confined , 2 eruptive ; ( c ) 1 single , 2 recurrent . in consequence , our classification can resolve @xmath5 subclasses .
our motivation should be commented in the context of the earlier classification made by @xcite , who proposed 5 morphological groups of xpes : a loop - type , spray - type , jet - type , confined , and other . in our opinion ,
the classification that is too ` hair - splitting ' may be uncomfortable in practical usage because it is easy to make a wrong assignment in case of poor quality of the observational data or their limited coverage .
we may recall attempts of organizing properties of cmes as observed by different coronagraphs @xcite .
they have never worked out any commonly accepted classification scheme for cmes on the basis of morphological features only .
our morphological criterion resolves only a direction of soft x - ray plasma movement in comparison with the direction of local magnetic field . roughly speaking , in the case of the subclass 1
the direction is parallel i.e. along the already existing magnetic field lines , in the case of the subclass 2 perpendicular i.e. across the already existing lines ( or strictly speaking together with them ) .
xpes from the first morphological subclass usually take a form of a blob or a column of matter propagating within a bundle of magnetic lines without any serious modification of their structure
. therefore , these events are more collimated ( hence its name ) and less energetic .
a direction of their motions depends on the configuration of guiding lines .
our subclass 1 comprises the majority of events classified by @xcite as the spray - type and jet - type events .
xpes from the second morphological subclass take a form of a rising loop or a system of loops .
our subclass 2 is very similar to the loop - type events proposed by @xcite .
some events showed features of both morphological subclasses , 1 and 2 , in this case we classified them according to a more evident feature .
for the xpe assignment into one of the kinematical subclasses we have chosen height increase rate above the chromosphere , @xmath6 .
a negative value , @xmath7 , means the subclass 1 , the opposite case , @xmath8 means the subclass 2 .
there are several papers presenting plots @xmath9 for events belonging to both subclasses @xcite .
xpes from the first kinematical subclass can be connected with plasma motion within closed magnetic structures as well as with some changes in a plasma situation or in the local magnetic field structure which do not evacuate any mass from the sun . in summary ,
xpes from the first kinematical subclass suggest the presence of kind of magnetic or gravitational confinement of x - ray plasma . for xpes from the second kinematical subclass ,
an increasing velocity in the radial direction in the field of view of the sxt allows us to anticipate further expansion leading to irreversible changes ( eruption ) of the local magnetic field . in consequence , at least a part of the plasma escapes from the sun . for many weak xpes
a construction of the diagram @xmath10 vs. @xmath11 was impossible .
therefore , we estimated @xmath12 qualitatively by watching the expansion rate of xpes in movies made with images uniformly spaced in time . in some cases
the classification was problematic because of a limited coverage of available observations , hence we added a question mark to the digit for this criterion .
according to our third criterion we separate disposable , unique xpes that occurred once in a time ( subclass 1 ) from recurrent events for which following expanding structures can be seen with time ( subclass 2 ) .
the majority of xpes described in the literature belongs to the subclass 1 , however samples of the subclass 2 were already presented @xcite .
a partial time coverage of the available observations probably introduces a bias toward single xpes , because a narrow observational window allows us to resolve only a single feature even for recurrent xpes .
especially important is col .
( 6 ) in which all available movies illustrating evolution of the xpe are collected .
the movies consist of images obtained by the sxt and are written in the mpeg format .
images made by using particular filters and spatial resolutions are collected in separate movies .
we label the movies to indicate their contents , e.g. , almg / hn marks images obtained with the almg filter of half resolution .
we use the following standard annotations of filters and spatial resolutions applied in the _ yohkoh _ software @xcite : the filter al.1 the wavelength range 2.5 - 36 , almg 2.4 - 32 , mg3 2.4 - 23 , al12 2.4 - 13 , be119 2.3 - 10 ; the full resolution , fn , 2.45 arcsec , half resolution , hn ,
4.9 arcsec , quarter resolution , qn , 9.8 arcsec .
a particular resolution means a specific field of view : 2.6 @xmath3 2.6 arcmin@xmath4 , 5.2 @xmath3 5.2 arcmin@xmath4 , 10.5 @xmath3 10.5 arcmin@xmath4 , for the fn , hn , and qn resolution , respectively .
sometimes we divided images made with the same filter and the same spatial resolution onto separate movies consisting of images made with the same time exposition . in that case , the labels contain additionally successive roman digits .
the movies consist of images that we previously processed using the standard _ yohkoh _
routine sxt_prep , allowing us to reduce an influence of typical instrumental biases , e.g. telemetry compression , electronic offset , dark current , straylight , de - jittering . in the images
the heliospheric coordinates are overwritten by using the solarsoft routine plot_map . for better identification of
a faint features slightly above the background , we represented a signal distribution with non - linear color tables nos .
16 ( `` haze '' ) , 33 ( `` blue - red '' ) , or 3 ( `` red temperature '' ) available in the interactive data language ( idl ) . images that form movies in the catalogue are sometimes non - uniformly spaced in time , therefore it is strongly recommended to watch a time print that is present in each image .
the xpes for which a more detailed analysis have been already performed show in col .
( 7 ) an entry with a concise report concerning results . inside the report ,
the obtained values of investigated parameters like velocity , acceleration , temperature , emission measure , electron density , pressure , and secondaries , as well as references , are given .
for 14 events ( nos . 29 , 30 , 34 , 53 , 62 , 67 , 72 , 126 , 144 , 169 , 252 , 293 , 303 , and 330 ) a more complete set of results is presented in form of plots and tables illustrating the whole evolution @xcite . finally , in col .
( 8) references to all the reports ( also in the electronic form ) in the chronological order are given .
the `` sxr flare '' entry contains a basic info about a flare that was associated with the given xpe . finding the associated flare for the majority of xpes in the catalogue was very easy .
a flare is seen usually in movies illustrating evolution of an xpe as heavy saturation due to its much stronger soft x - ray radiation . in several cases
a flare occurred simultaneously with an xpe but in another active region .
we considered this flare as the associated event only when some distinct magnetic loops connecting both active regions were seen .
finally , there are several xpes that occurred when no flare was observed on the sun .
each record describes the following attributes : date , time of start , maximum , and end defined on the basis of the _ geostationary operational environmental satellites ( goes ) _ 18 light curve , _ goes _ class , location in heliographic coordinates , noaa active region number . by clicking on
the _ goes _ class , one can view the _ goes _ light curves in two wavelength ranges : 18 ( upper ) and 0.54 ( lower ) .
time span of plots is always two hours and includes the occurrence of an xpe , which is marked by vertical lines . the hatched area on the plot represents _ yohkoh _ nights .
records presented in this entry are generally adopted from the sgd , however some clarifications and supplements were necessary .
for example , the lacking locations were completed on the basis of sxt images as a place of flare bright loop - top kernels .
the values obtained in this way are given in parenthesis .
coordinates of events that occurred behind the solar limb are taken basically from @xcite . if no flare was associated with an xpe , the tags devoted to flare characteristics are empty .
exceptions are _ goes _ light curves , heliographic coordinates , and noaa active region number .
the last two tags describe then an xpe .
the `` hxr flare '' entry presents some attributes of hard x - rays emitted by a flare that was associated with a given xpe .
we used data from the hxt onboard _
yohkoh_. this telescope measured the hard x - ray flux in four energy bands : 14 - 23 ( l ) , 23 - 33 ( m1 ) , 33 - 53 ( m2 ) , and 53 - 93 kev ( h ) .
each record contains peak time and peak count rate ( together with the background ) , inferred for the energy band m1 . by clicking on the peak count rate , one can view the hxt light curves in all energy bands .
time span of plots usually includes the maximum of hard x - ray flux and the occurrence of an xpe , which is marked by vertical lines .
wherever available , we also include the event i d from the _ yohkoh _ flare catalogue ( hxt / sxt / sxs / hxs ) if no flare was associated with an xpe , the tags devoted to flare characteristics are empty . in case the hard x - ray flux in the energy band m1 was below the doubled value of the background we left the tags describing peak time and peak time rate empty .
the `` cme '' entry contains some attributes of a cme that was associated with a given xpe .
the observations are derived by the large angle and spectrometric coronagraph ( lasco ) onboard the _ solar and heliospheric observatory ( soho)_. from the _ soho _ lasco cme catalog @xcite values of the following parameters are given : date and time of the first appearance in the c2 coronagraph field of view , central position angle , angular width , speed from linear fit to the @xmath9 measurements , acceleration inferred from the quadratic fit .
the first appearance time is the link to the beginning of the list of events in the _ soho _ lasco cme catalog for a given year and month . by clicking on the entry `` related links
'' one can view a javascipt movie of the cmes within the c2 field of view for a given day .
movies reside at the the homepage of the _ soho _ lasco cme catalog . according to @xcite
, we consider a pair xpe - cme as physically connected if the xpe occurred within position angles defined by the cme angular width increased by 10@xmath13 from both sides .
moreover , time of the xpe occurrence had to fall within 3-hours - interval centered around extrapolated time of the cme start for @xmath14 . for extrapolation we used time of the first appearance in the lasco / c2 field of view and linear velocity taken from the _ soho _ lasco cme catalog .
the `` cme '' entry is empty when the xpe occurred during a lasco gap .
xpes not associated with a cme are labeled : no related event. in `` references '' entry the references to all reports ( also in the electronic form ) , known for us , that mentioned a particular xpe are given in the chronological order .
some useful characteristics of xpes included in the catalogue are extracted in table [ tb1 ] . in this subsection
we discuss general statistics of the catalogue content . in table[tb2 ] the quality of xpe observations from the catalogue
is summarized .
the most frequent are events that we categorized as ( b ) and ( c ) .
contribution of remaining categories is marginal . in sections[ass - fl ] and [ ass - cme ] we present results of a statistical analysis that was performed for two different populations of events : from ( a ) to ( c ) and from ( a ) to ( b ) .
the first population is more frequent , which offers some advantages in statistical approach , however for events categorized as ( c ) observations are often not complete enough to give confidence in our classification choices . in consequence , in the first population an additional bias can be introduced which is inadvisable .
we expect that this problem is overcome for less frequent , second population of xpes which was better observed . in table[tb3 ]
we summarize heliographic longitudes of xpes in the catalogue . a distinct concentration of the xpes around the solar limb is seen .
this is an artificial effect caused by observational constrains .
xpes are easier for detection when we observe them against the dark background sky than when we observe them between plenty of different features seen on the solar disk . moreover , in two from three main surveys that we used in our catalogue ( * ? ? ?
* ohyama 2009 , private communication ) , only flares that occurred close to the solar limb ( @xmath1 ) were systematically reviewed .
is it possible to estimate the actual number of xpes , which occurred on the sun during the _ yohkoh _ years ? assuming their uniform distribution with heliographic longitude and taking the number for the interval @xmath15 as the most representative , we obtain a value @xmath16 . including a duty time of _ yohkoh _ to be about 0.65 ( ratio of satellite day to the total orbital period ) we obtain a number @xmath17 .
however , even this huge number could be significantly lower than actual due to several reasons .
firstly , we do not include an influence of worse detection conditions before 1997 .
secondly , for strong flares , especially in 2001 , the conditions for detecting xpes were quite bad because of too long exposures .
thirdly , the estimated number is roughly representative for flares stronger than the _ goes _ class c5-c6 . only for those events
the flare mode was initiated in the _ yohkoh _ operation @xcite and this mode guarantees a sufficient time resolution of an image cadence for a successful detection of xpes .
xpes associated with weaker flares are only known accidentally , since no systematic examination of images recorded during the quiet mode of the _ yohkoh _ satellite has been performed yet . in conclusion
, xpes should be considered as very frequent events occurring in the solar corona .
xpes described in the catalogue are only a minor representation of a countless population of events , which are typical for the hot solar corona . in table[tb4 ]
we present a time coverage of the observed xpes .
evolution of an important fraction of events ( 42.6% ) is illustrated only partially .
this limits its detailed investigation .
even relatively simple activities like classification can be meaningless .
for example , xpes classified as single and observed only partially can be actually recurrent . in table[tb5
] we present a number of xpes that were classified onto one of eight subclasses defined under following three criteria : morphological , kinematical , and recurrence .
we organize the results twofold : for the total population and for carefully selected events . in the second case we omitted events categorized as ( c ) , untrustworthy assignments of kinematical criterion as shown in the catalogue with the question mark , and examples classified as single in case of partial time - coverage of observations .
it reduces the whole population almost three times and for particular subclasses even more , but we believe that numbers less affected by observational limits , seen in the last column of table[tb5 ] , are more representative for real conditions . what do these numbers tell us about xpes ? at first sight , loop - like xpes seem to be more frequent than collimated xpes of a factor 2.1 or 3.2 for total population and special selection , respectively .
however , it can be caused by the effect of observational selection .
on average , loop - like xpes are more massive than collimated ones @xcite , thus they are easier for detection above the background .
indeed , the exclusion of faint events categorized as ( c ) increases relative contribution of loop - like xpes .
it is interesting that in @xcite the number of loop - type xpes ( 60 ) only slightly outnumbers the sum of spray - type and jet - type events ( 51 ) . for other relations ,
the carefully selected events seem to be less affected by observational constraints than those of the entire populations .
therefore we conclude that the former events adequately characterize intrinsic features of particular subclasses .
for example , for loop - like xpes the relation between eruptive and confined or between recurrent and single events is distinctly different from that for collimated xpes .
loop - like xpes are dominantly eruptive ( 82 to 14 ) and recurrent ( 59 to 37 ) , whereas collimated xpes are more frequently confined ( 19 to 11 ) and single ( 20 to 10 ) .
we would like to stress that subclasses of xpes defined by us resemble some types of classical prominences observed in h@xmath18 line @xcite .
namely , a surge is a prominence that is collimated and confined , a spray is a prominence that is collimated and eruptive , a loop - like and confined event we call as an activation of a prominence , and a loop - like and eruptive event is an eruptive prominence or disparision brisque. classifications of prominences do not distinguish the recurrence criterion , nevertheless there are known observations , in which aforementioned types of prominences were observed as single or recurrent events ( rompolt 2011 , private communication ) .
this similarity between xpes and prominences suggests a close association between hot and cold components of active regions .
this relation was not investigated in details so far , except for @xcite .
using the classification we were able to separate several subclasses of xpes looking more homogeneous than the full population .
unfortunately , we are not sure if particular subclasses of xpes refer to events that are physically different . a quantitative analysis of soft x - ray images would give closer confirmation , however for the majority of xpes in the catalogue this kind of analysis is practically unreliable , due to minor signal and other observational limits .
therefore , the main motivation of this section is to justify the presence of physically different subclasses of xpes by a comparison of properties of other solar - activity phenomena associated with particular xpes .
basic characteristics of flares and cmes are known well . in this section
we present the association of xpes with flares , in section [ ass - cme ] we present the association of xpes with cmes .
it has been commonly agreed that an xpe is a consequence of a flare occurrence . as a matter of fact , there are 5 xpes in our catalogue , for which we could not find any associated flare . however
, this sample is too small to justify the existence of flareless xpes .
moreover , three of the five xpes occurred close enough to the solar limb that they might have come from flares from the backside .
are they just the tip of an iceberg ?
the answer may depend on extensive and careful examination of sxt images made in the quiet mode . for better insight in time coincidence between xpes and flares
we mark the time of an xpe on the _ goes _ light curve of an associated flare . in fig .
[ flb - xpeb ] we present a histogram of time differences between start times of flares and xpes for 330 pairs of events . in 311 cases from 330 ( 94.2% ) ,
an increase of soft x - ray emission occurred earlier than the xpe .
the time difference is very often several minutes only ( see maximum and median of the histogram ) , however higher values also occur .
similar conclusions can be given regarding better observed xpes of quality a and b ( compare the gray bins in fig .
[ flb - xpeb ] ) .
similar histograms made for particular subclasses of the xpes introduced in our classification do not show any important differences . in fig .
[ flm - xpee ] we present a histogram of time differences between the end of xpes and the peak of the associated flares as determined from the _ goes _ light curves . in
about 20% of investigated samples ( 41 from 198 ) any xpes were completed before the flare peak , i.e. within the rising phase of a flare . for almost 80% of events
the final evolution of xpes is seen after the maximum of soft x - ray emission , very often no longer than 10 minutes ( 108 examples from 198 , 54.5% ) .
similar conclusions can be given regarding better observed xpes of quality a and b ( compare the gray bins in fig .
[ flm - xpee ] ) .
similar histograms made for particular subclasses of xpes introduced in our classification do not show any important differences .
the above results should be normalized to the time scales of flares .
therefore , we have prepared counterparts of figs .
[ flb - xpeb ] and [ flm - xpee ] in which we normalize time differences with the flare rising - phase duration . in fig .
[ flbn - xpebn ] we illustrate occurrences of the xpe start : negative values mean that a xpe preceded its flare , the value 0 simultaneous start , the value 1 start of a xpe at the maximum of its flare , values grater than 1 later xpe start . as we see , the majority of xpes ( 282 from 330 , 85.5% ) starts within the rising phase of flares .
this rule is fulfilled even stronger for better observed xpes ( gray bins ) 176 from 198 , 88.8% .
interesting results are revealed by further versions of fig .
[ flbn - xpebn ] , in which particular subclasses of xpes are separated : collimated and loop - like , confined and eruptive , single and recurrent , in figs .
[ flbn - xpebn - cl]-[flbn - xpebn - sr ] , respectively . in these figures ,
we show side - by - side the distributions of the xpes that have contrasting properties .
for example , in fig .
[ flbn - xpebn - cl ] loop - like xpes show tendency to start earlier in the rising phase of flares than collimated ones : the difference for medians is more than 0.2 of the rising - phase duration .
a similar tendency is seen in fig .
[ flbn - xpebn - ce ] where eruptive xpes start earlier in the rising phase of associated flares than confined ones and in fig .
[ flbn - xpebn - sr ] where recurrent xpes precede , on average , single ones . in fig .
[ flmn - xpeen ] we have normalized time differences between the xpe end and the associated flare start with the rising - phase duration of a flare . in this scale
the value 1 means that the xpe end occurred exactly at the maximum of the associated flare .
the histogram is rather gradual with two maxima between 1.2 - 1.4 and 1.6 - 1.8 .
the number of xpes lower than the first maximum and greater than the second one decreases systematically with a marginal contribution of those that are lower than 0.4 and greater than 3 .
it means that all soft x - ray plasma motions are limited within a relatively narrow part of total duration of associated flares .
the variants of fig .
[ flmn - xpeen ] , in which particular subclasses of xpes are separated : collimated and loop - like , confined and eruptive , single and recurrent , are presented in figs .
[ flmn - xpeen - cl]-[flmn - xpeen - sr ] , respectively . in these figures , as in figs .
[ flbn - xpebn - cl]-[flbn - xpebn - sr ] , we show side - by - side the distributions of the xpes that have contrasting properties . in fig .
[ flmn - xpeen - cl ] the collimated xpes seems to last longer , on average , than the loop - like ones : medians of both distribution differ by 0.4 of the rising - phase duration .
similarly , the confined xpes seems to last longer than the eruptive ones ( fig .
[ flmn - xpeen - ce ] ) medians differ about 0.6 of the rising - phase duration , in case of better observed xpes . in fig .
[ flmn - xpeen - sr ] the single xpes last longer , on average , than the recurrent ones medians differ about 0.35 of the rising - phase duration , in case of better observed xpes . for each associated flare we determined x - ray class and
total duration based on light curves recorded by _ goes _ , in the wavelength range of 1 - 8 .
we defined the total duration as the interval between a constant level of the solar soft x - ray flux before and after a flare , therefore our values of this parameter are larger than intervals between a start time and end time that are routinely reported in the sgd . in some cases
we could not estimate the total duration properly .
this is the reason why a number of considered events in this paragraph is slightly lower than the number of xpes associated with flares . in fig .
[ xpefl1 ] we present scatter plot of x - ray class versus total duration for flares associated with morphological subclasses of xpes , i.e. , collimated and loop - like xpes .
all points are marked with dots .
additionally we emphasized well - observed xpes ( quality a or b ) and flares that are non - occulted by the solar disk .
these flares associated with well - observed collimated and loop - like xpes are marked with boxes and stars , respectively .
both groups of flares are mixed in the plot , however some shifts toward higher x - ray class and longer duration can be seen for flares associated with loop - like xpes .
similar scatter plots of x - ray class versus total duration for flares associated with kinematical and recurrence subclasses of xpes are given in figs .
[ xpefl2]-[xpefl3 ] .
all points are marked with dots .
again , additionally we emphasized well - observed xpes ( quality a or b ) and flares that are non - occulted by the solar disk .
moreover , we excluded events for which the assignment of kinematical subclasses for xpes was uncertain ( fig .
[ xpefl2 ] ) and events associated with xpes that were classified as single in case of partial time - coverage of observations ( fig .
[ xpefl3 ] ) . in both figures , flares
associated with well - observed xpes classified as subclass 1 ( confined and single , respectively ) are marked with boxes , whereas flares associated with xpes classified as subclass 2 ( eruptive and recurrent , respectively ) are marked with stars . similarly to fig .
[ xpefl1 ] , both groups of flares are mixed in the plots and some shifts toward higher x - ray class and longer duration is seen for flares associated with xpes of subclasses 2 .
the shifts seen in figs .
[ xpefl1]-[xpefl3 ] are confirmed by medians calculated separately for both groups of flares for each classification criterion . as it is seen in table [ flare ] ( bold - faced columns ) , medians for flares associated with xpes of subclass 2 are 1.53.8 times and 2.12.7 times greater than medians for flares associated with xpes of subclass 1 for flare x - ray class and flare duration , respectively .
higher x - ray class and longer duration mean a more energetic flare , thus we can conclude that more energetic xpes are , on average , associated with more energetic flares and less - energetic xpes rather prefer less - energetic flares .
one can expect that a difference between characteristics describing associated flares should be even higher for two subclasses of xpes defined by combining our three criteria simultaneously .
indeed , medians in table [ flare ] for flares associated with the subclass ( 1,1,1 ) collimated , confined , single xpes and the subclass ( 2,2,2 ) loop - like , eruptive , recurrent xpes show extreme differences ( a factor 6.0 and 3.7 for x - ray class and duration , respectively ) . as it is seen in fig .
[ xpefl4 ] , in the diagram x - ray class versus duration , flares associated with the subclasses ( 1,1,1 ) and ( 2,2,2 ) of well - observed xpes are almost separated . in unbold - faced columns in table [ flare ] we present medians for flares associated with different subclasses of xpes that were defined less strictly , i.e. by including quality c events and without excluding any doubtful examples .
ratios of medians for subclasses 2 and subclasses 1 that were constituted more liberally are usually lower in comparison with the more strictly defined bold - faced values .
it shows how some physical differences can be masked by observational constrains .
we included in the catalogue hard x - ray light curves of associated flares , recorded by _
hxt , for investigating the relation between xpes and non - thermal electron signatures .
we considered light curves in energy band m1 ( 23 - 33 kev ) and interpreted a signal above the doubled value of the background as the proof that in a particular flare an acceleration of an appropriate number of non - thermal electrons occurred . for 353 flares which were associated with xpes we found that 235 events , i.e. 66.6% showed this signature . in the second and the third columns of table[flare ] we present detailed results of this relation for both populations of events and for particular subclasses of xpes .
a percentage of associated flares showing non - thermal electrons depends on how energetic is the subclass , with higher values ( 75%-82% ) for subclasses 2 and lower values ( 57%-75% ) for subclasses 1 .
the difference is the highest for the kinematical criterion , but for the recurrence criterion percentages for both subclasses are almost the same . after applying all the three criteria we found that the difference between the least energetic subclass ( 1,1,1 ) collimated , confined , single and the most energetic ( 2,2,2 ) loop - like , eruptive , recurrent is maximal : 57% and 86% of associated flares indicating non - thermal electrons , respectively .
we also investigated time coincidence between macroscopic x - ray plasma motions ( xpes ) and non - thermal electron signatures ( hxrs ) in detail . in this aim , we measured time differences between the xpe start and the hxt / m1 flare peak .
the results are presented as a histogram in fig .
[ hxr - xpe ] . in 185 from 227 cases ( 81.5% ) xpes started before the hxr flare peak , in 18.5% of cases the chronology was opposite .
however , the most frequent bin : 02 minutes in about 45% of cases , suggests that both considered processes , i.e. soft x - ray plasma motion and non - thermal electron acceleration , are strongly coupled .
similar investigation was performed by @xcite . at first glance our fig .
[ hxr - xpe ] and their fig .
5 are different .
however , it is needed to know that the values in our histogram have opposite sign and we used the hxr peak time for higher energy band m1 , 23 - 33 kev , than kim et al . who used energy band l ( 14 - 23 kev ) .
in flares with a strong contribution of the non - thermal component , the peak time in those energy bands are close , but in flares with a stronger contribution of the superhot component in l band , the m1 peaks tend to occur earlier than the l peaks .
keeping in mind the above mentioned differences in data organization we can conclude that our results are consistent .
we also prepared variants of fig .
[ hxr - xpe ] for particular subclasses of xpes .
however , we did not find any evident differences between the considered distributions , namely , each of them shares the common peak bin .
in order to associate out xpes with cmes , we used the _ soho _
lasco cme catalog @xcite .
only 275 xpes occurred when the lasco coronagraphs were operational .
we found that 182 xpes ( 66.2% ) were associated with cmes .
this is slightly less than 69% ( 95 from 137 events ) obtained by @xcite .
for particular subclasses of xpes the association was between 44% and 88% ( see table [ cmes ] ) . according to @xcite
, we consider a xpe - cme pair as physically connected if the xpe occurred within the position angle range defined by the cme angular width increased by 10@xmath13 from either side .
moreover , the time of the xpe had to fall within 3-hours - interval centered around the extrapolated time of the cme front start at @xmath14 .
for the extrapolation we used the time of the first appearance in the lasco / c2 field of view and the linear velocity taken from the cme catalog .
the histogram of the time differences between the extrapolated cme front onset and the xpe start is presented in fig .
[ cme - xpe ] .
as we can see , there are more events with negative values , i.e. , those in which the cme starts before the xpe , than those with the opposite chronology .
the frequencies are 73.7% ( 87 from 118 ) and 26.3% ( 31 from 118 ) , respectively .
the carefully selected subgroup ( well - observed xpes of quality a or b that occurred close to the solar limb , @xmath1 ) shows slightly different proportions : 66,1% ( 41 from 62 ) and 33.9% ( 21 from 62 ) , respectively .
both distributions : all the xpes and the selected xpes , are quite gradual with slightly different medians : -17.3 min . and -10.6 min .
. @xcite performed similar analysis for xpes from the two - years interval 1999 - 2001 .
their fig .
6 containing 43 events was made under slightly different assumptions : ( 1 ) the cme - front times were extrapolated at individual locations of xpes in the _ yohkoh _ field of view , ( 2 ) the cme speed was determined from the first two observing times and heights . despite these differences , our histogram looks quite similar to those of kim et al .
therefore we conclude that , at least in a statistical sense , different ways of extrapolating the cme onset time do not seriously affect the temporal relation between xpes and cmes . as in the analysis in section [ fl - tc ] , we give further versions of fig .
[ cme - xpe ] , in which particular subclasses of xpes : collimated and loop - like , confined and eruptive , single and recurrent , are separated in figs .
[ cme - xpe - cl ] , [ cme - xpe - ce ] , and [ cme - xpe - sr ] , respectively . in fig .
[ cme - xpe - cl ] , the histogram for loop - like xpes shows a relatively narrow maximum located close to the zero point .
it means that a large fraction of xpes ( @xmath1950% ) starts almost simultaneously with the cme onset .
collimated xpes are shifted towards negative values in this figure and their maximum is distinctly broader .
the histogram made for the selected subgroup of better observed events ( gray and black bins in fig .
[ cme - xpe - cl ] ) show a similar trend .
a similar pattern can be seen in fig .
[ cme - xpe - ce ] . in this figure
the histogram made for eruptive xpes is narrower and centered closer to the zero point than the histogram made for confined xpes .
the distribution for confined xpes is much broader , especially in the plot for the selected subgroup of better observed events ( black bins ) , which makes an impression of a random occurrence within almost the whole time window of the cme onset . in fig .
[ cme - xpe - sr ] both histograms made for single and recurrent xpes show a similar width , but recurrent xpes tend to start earlier ( almost simultaneously with the cme ) than single ones .
the difference between medians is about 15 minutes for all events as well as for the selected subgroup of better observed events . in fig .
[ xpecme1 ] we present a scatter plot of an angular width versus a linear velocity for cmes associated with morphological subclasses of xpes ,
i.e. , collimated and loop - like xpes .
all points are marked with dots .
additionally we emphasized well - observed xpes ( quality a or b ) that occurred close to the solar limb ( @xmath20 ) .
cmes associated with well - observed collimated and loop - like xpes are marked with boxes and stars , respectively .
both groups of cmes are mixed in the plot , however some shifts toward wider and faster events is seen for cmes associated with loop - like xpes .
similar scatter plots of an angular width versus a linear velocity for cmes associated with kinematical and recurrence subclasses of xpes are given in figs .
[ xpecme2]-[xpecme3 ] .
all points are marked with dots .
again , additionally we emphasized well - observed xpes ( quality a or b ) that occurred close to the solar limb ( @xmath1 ) .
moreover , we excluded events for which the assignment of kinematical subclasses for xpes was uncertain ( fig .
[ xpecme2 ] ) and events associated with xpes that were classified as single in case of partial time - coverage of observations ( fig .
[ xpecme3 ] ) . in both figures , cmes
associated with well - observed xpes classified as the subclass 1 ( confined and single , respectively ) are marked with boxes , whereas cmes associated with well - observed xpes classified as the subclass 2 ( eruptive and recurrent , respectively ) are marked with stars . similarly to fig .
[ xpecme1 ] , both groups of cmes are mixed in the plots and some shifts toward wider and faster events is seen for cmes associated with xpes of the subclasses 2 .
the shifts seen in figs . [ xpecme1]-[xpecme3 ] are confirmed by medians calculated separately for both groups of cmes for each classification criterion . as it is seen in table [ cmes ] ( bold - faced columns for well - observed events ) ,
the medians for cmes associated with xpes of the subclass 2 are 1.22.0 times and 1.21.4 times greater than those for cmes associated with xpes of the subclass 1 for cme angular width and linear velocity , respectively . a higher angular width and
velocity mean a more energetic cme , thus we can conclude that more energetic xpes are , on average , associated with more energetic cmes and less - energetic xpes rather prefer less - energetic cmes .
one can expect that a difference between characteristics describing associated cmes should be even higher for two subclasses of xpes that we define by applying our three criteria simultaneously .
indeed , medians in table [ cmes ] for cmes associated with the subclass ( 1,1,1 ) collimated , confined , single xpes and the subclass ( 2,2,2 ) loop - like , eruptive , recurrent xpes show extreme differences ( factors 2.1 and 1.7 for angular width and linear velocity , respectively ) . the difference between characteristics describing associated cmes
is also seen in fig .
[ xpecme4 ] . in unbold - faced columns in table [ cmes ] we present medians for cmes associated with different subclasses of xpes that were defined less strictly
, i.e. by including quality c , for @xmath21 and without excluding any doubtful examples . ratios of medians for the subclasses 2 to medians for the subclasses 1 that were constituted more liberally are often comparable to the more strictly defined bold - faced values .
it is opposite to flares for which bigger differences between unbold - faced and bold - faced values are evident ( see table [ flare ] ) .
we suggest that the main reason for this is the condition @xmath1 , that we constituted for specially selected ( bold - faced ) events .
it excludes the majority of halo cmes being systematically wider and faster than ordinary cmes @xcite . in other words ,
more strict selection criteria undoubtedly limit a scatter of values in two physically different groups , however it is compensated with the bias introduced by halo cmes .
the xpes collected in the catalogue confirm the strong association with flares .
starts of xpes observed since their very beginning fall usually within the rising phase of associated flares ( figs.[flb - xpeb ] and [ flbn - xpebn ] ) and well coincide with the hxr peaks ( fig .
[ hxr - xpe ] ) .
it means that symptoms of sxr plasma motions occur when magnetic energy conversion in flares via reconnection is most vigorous @xcite .
a small number of exceptions is connected mainly with complex events in which x - ray enhancements , recorded by _ goes _ and _ yohkoh_/hxt , are accumulated from at least two different positions on the sun .
we would like to stress a lower correlation between xpes and hxr flares than between xpes and sxr ones . as approximately one third of flares associated with xpes
did not show any clear signatures of non - thermal electrons , we can conclude that some macroscopic motion of sxr plasma is more obvious characteristics of reconnection than acceleration of non - thermal electrons .
it can be caused by the the limited sensitivity of the hxt . as we can see in table [ flare ] , the more energetic subclasses 2 in our classification scheme of xpes show a stronger correlation with hxr flares .
thus , under the assumption that the more energetic xpes are associated with more energetic flares , we can expect a larger fraction of them to be able to produce the hxr emission above the threshold of the hxt .
another proof of close association between processes responsible for xpes and flares is similarity between their durations . a comparison between the medians in figs .
[ flbn - xpebn ] and [ flmn - xpeen ] shows that an xpe lasts , on average , as long as the rising phase of an associated flare .
histograms presented for particular subclasses of xpes ( figs . [ flbn - xpebn - cl]-[flbn - xpebn - sr ] and [ flmn - xpeen - cl]-[flmn - xpeen - sr ] ) show that the more energetic subclasses 2 occur earlier and last shorter than the less energetic subclasses 1 .
this difference probably reflects some differences in reconnection processes occurring in both subclasses . at first sight , this result is in contradiction with figs .
[ xpefl1]-[xpefl3 ] , in which xpes from subclasses 2 seem to prefer flares of longer duration , however we should remember that in figs .
[ flbn - xpebn - cl]-[flbn - xpebn - sr ] and [ flmn - xpeen - cl]-[flmn - xpeen - sr ] the time is normalized with the flare rising - phase duration .
some interesting hints concerning hierarchy and chronology of processes occurring in restructuring active regions can be found in histograms of time differences between the xpe start and the extrapolated cme onset for particular subclasses of xpes ( figs .
[ cme - xpe - cl]-[cme - xpe - sr ] ) .
the more energetic subclasses 2 of xpes show close relationship with their associated cmes .
they seem to start almost simultaneously , and small deviations from the zero value are probably caused by unrealistic extrapolation of the cme onsets .
the start of xpes of the less energetic subclasses 1 shows a much looser connection with the cme start . very often the cme seems to occur first . keeping in mind that xpes are usually caused by magnetic reconnection
, we suggest that in the case of the subclasses 2 , the reconnection and loss - of - equilibrium of magnetic structure , thus a cme development , occur almost simultaneously . on the other hand , in the case of the subclasses 1 ,
the reconnection is usually a consequence of destabilization of magnetic structure , which may occur earlier .
the results summarized in tables [ flare ] and [ cmes ] strongly suggest that total amount of energy , converted from the magnetic field in an active region during its magnetic reconfiguration , determines characteristics of events including flares , cmes , and xpes , which are thought to be consequences of this common reconfiguration .
thus , more energetic xpes are associated with more energetic flares and cmes , while less energetic ones seem to occur commonly . this statistically averaged picture does not exclude , for sure , exceptions in partitioning of magnetic energy .
for example , there are x - class confined flares completely devoid of any cme @xcite .
these flares are probably also devoid of xpes but this research is beyond the scope of this work .
our investigation shows that characteristics of flares and cmes associated with particular subclasses of xpes are different .
we found that a scale of differences is higher for flares than for cmes .
we also found that the recurrence criterion proposed in our xpe classification scheme does not separate the associated events as strongly as the morphological and kinematical criteria .
in our catalogue we have collected the most extensive database of xpes so far .
images from the sxt onboard _
yohkoh _ have been organized into movies in the mpeg format .
the events have been classified on the basis of elementary and uniform criteria .
the catalogue also gives a piece of information concerning the associated flares and cmes by using entries to the _ yohkoh _ flare catalogue ( hxt / sxt / sxs / hxs ) and the _ soho _ lasco cme catalog , respectively .
the collected data allow us to study xpes more comprehensively as a separate solar activity phenomena and also as elements of more complex processes occurring in the solar corona .
xpes constitute a strongly inhomogeneous group of events .
their appearances include expanding loop structures , moving blobs , rising columns , and so on .
their strong inhomogeneity is responded by wide range of values of basic parameters : altitude ( @xmath22 cm ) , volume ( @xmath23 @xmath24 ) , duration ( @xmath25 s ) , velocity ( @xmath26 kms@xmath27 ) , acceleration ( @xmath28 ms@xmath29 ) , mass ( @xmath30 g ) , energy ( @xmath31 ergs ) .
it is difficult to point out a universal mechanism responsible for all the events presented in the catalogue .
there is no doubt that the majority of xpes is connected somehow with magnetic reconnection .
however , in many events , the evolution is far from what may be expected from the canonical cshkp model , suggesting the existence of more complex 3d quadrupolar reconnection @xcite .
we often observe an xpe as a result of magnetic reconnection that leads to chromospheric evaporation as a hydrodynamic response of intensified plasma heating or non - thermal electron beams in a flare magnetic structure . on the other hand
, the close morphological and kinematical connection of some xpes with cmes , together with the similar start time , suggests a mechanism of loss - of - equilibrium type common for cmes .
[ indeed , movies illustrating evolution of some xpes resemble cartoons presenting the tether release model or the tether straining model leading to the magnetic breakout model . ] finally , some movies in the catalogue give an impression that sxr plasma leaks out from the magnetic structure probably under low-@xmath32-plasma conditions .
for proper interpretation of the data , we need to identify the mechanism responsible for the observed xpe .
an inappropriate choice of the mechanism can lead to meaningless and erroneous conclusions regarding processes occurring in the solar corona . in the context of strong inhomogeneity of xpes and several possible mechanisms of their origin ,
it is not advised to routinely interpret all xpes in terms of a single and same mechanism .
the similar conclusion were given by @xcite who criticized the tendency to employ the cshkp model for description all `` masuda - type '' flares @xcite . if we consider a one particular xpe , it is basically difficult to decide which mechanism is responsible for its occurrence without the complete quantitative analysis including plasma diagnostics and a modeling of magnetic field structure .
these conditions were unreachable in practice for the majority of events in the catalogue .
therefore , in advent of new observations of xpes derived by modern instruments onboard _ hinode _ , _ solar - terrestrial relations observatory _
( _ stereo _ ) , and the _ solar dynamics observatory _
( _ sdo _ ) , we have been trying to give some solutions that would be correct at least in statistical sense . we have shown that the subclasses of xpes separated on the basis of our simple observational criteria have different levels of correlation with other solar - activity phenomena . the difference is also seen if we consider basic parameters describing these flares and cmes .
however , the association of xpes with different flares or cmes does not mean a specific physical mechanism as far as these flares or cmes represent physically different groups .
in the meantime , discussions concerning a difference in observational characteristics to justify separate physical mechanisms responsible for flares or cmes are still open .
are there two different classes of flares @xcite or all flares can be explained by only one mechanism @xcite ? are there two kinematically different classes of cmes @xcite or the division is artificial @xcite ?
we have found that more energetic xpes are better correlated with flares and cmes and that more energetic xpes correlate with more energetic flares and cmes .
virtually the effect of observational conditions works in the same way and we can not resolve correctly the influence of the effect on our conclusion .
the most promising way in the investigation of xpes is to deal them as an element of a larger ensemble .
the usage of observations made in temporal , spatial , and spectral ranges broader than those needed for direct monitoring of xpes allows the better understanding of processes in which xpes participate .
recently , a similar picture of flares as global events was presented by @xcite .
our experience is that xpes are strongly coupled with flare hxr quasi - periodic oscillations @xcite , probably because the reconnection rate is controlled by plasmoid generation @xcite .
we also found that xpes are somehow associated with progressive spectral hardening in hxrs @xcite , thus with solar energetic particles @xcite . in the future we are going to upgrade the xpes catalogue by adding entries devoted to associated prominences and radio bursts .
moreover , the _ trace _ movies will be added , if available .
we also are going to perform the comprehensive analysis of several , very interesting events from the catalogue that have been omitted by other _
yohkoh _ researchers .
_ yohkoh _ is a project of the institute for space and astronautical sciences , japan , with substantial participation from other institutions within japan and with important contributions from the research groups in the us and the uk under the support of nasa and serc .
the sxt instrument was jointly developed by the lockheed palo alto research laboratory and the national astronomical observatory of japan .
collaborators include the university of tokyo , stanford university , the university of california at berkeley , and the university of hawaii .
the cme catalog used in this work is generated and maintained at the cdaw data center by nasa and the catholic university of america in cooperation with the naval research laboratory .
soho is a project of international cooperation between esa and nasa .
the _ yohkoh _
flare catalogue ( hxt / sxt / sxs / hxs ) used in this work was developed by j. sato , k. yoshimura , t. watanabe , m. sawa , m. yoshimori , y. matsumoto , s. masuda and t. kosugi .
we acknowledge solar data collected and distributed by u. s. national geophysical data center .
we thank dr .
masamitsu ohyama for providing us his list of xpes from years 1991 - 1998 .
we appreciate the valuable remarks of the referee , which helped us to improve this paper .
this work was supported by polish ministry of science and high education grant no .
nn203193733 .
alexander , d. , metcalf , t.r .
, & nitta , n.v .
2002 , , 29 , 1403 aschwanden , m.j . , kosugi , t. , hanaoka , y. , nishio , m. , & melrose , d.b .
1999 , , 526 , 1026 b@xmath33rta , m. , vr@xmath34nak , b. , & karlick@xmath35 , m. 2008 , , 477 , 649 bak - ste@xmath36licka , u. , ko oma@xmath37ski , s. , & mrozek , t. 2011 , cent .
, 35 , 135 benz , a.o .
2008 , living rev .
solar phys . , 5 , 1 burkepile , j.t . , & st .
cyr , o.c .
1993 , a revised and expanded catalogue of mass ejections observed by the _ solar maximum mission _ coronagraph , high altitude observatory , boulder , colorado chen , p.f .
2011 , living rev .
solar phys . , 8 , 1 cheng , x. , zhang , j. , ding , m.d . , guo , y. , & su , j.t .
2011 , , 732 , 87 chertok , i.m .
2000 , j.atm .
solar terr.phys . , 62 , 1545 chmielewska , e. 2010 , master thesis , university of wrocaw ( in polish ) chmielewska , e. , & tomczak , m. 2012 , cent .
astrophys . bull . , submitted dauphin , c. , vilmer , n. , & krucker , s. 2006 , , 455 , 339 falewicz , r. , tomczak , m. , & siarkowski , m. 2002 , esa - sp 506 , proc .
10th european solar physics meeting , a.wilson , noordwijk : esa publications division , 601 forbes , t.g .
2000 , , 105 , 23153 gallagher , p.t . ,
lawrence , g.r . & dennis , b.r .
2003 , , 588 , l53 gopalswamy , n. , yashiro , s. , michaek , g. , stenborg , g. , vourlidas , a. , freeland , s. , & howard , r. 2009 , earth moon planet , 104 , 295 grayson , j.a . , krucker , s. , & lin , r.p .
2009 , , 707 , 1588 hara , h. , tsuneta , s. , lemen , j.r . , acton , l.w .
, & mctiernan , j.m .
1992 , , 44 , l135 harrison , r.a . ,
waggett , p.w . , bentley , r.d . ,
phillips , k.j.h .
, bruner , m. , dryer , m. , & simnett , g.m .
1985 , , 97 , 387 hori , k. 1999 , nro report no .
479 , proc .
nobeyama symposium , t.s .
bastian , n. gopalswamy , & k. shibasaki , 267 howard , r.a . ,
sheeley , n.r .
michels , d.j .
, & koomen , m.j .
1985 , , 90 , 8173 hudson , h.s .
2011 , , 158 , 5 hudson , h.s . ,
kosugi , t. , nitta , n.v . , & shimojo , m. 2001 , , 561 , l211 khan , j.i . ,
vilmer , n. , saint - hilaire , p. , & benz , a.o .
2002 , , 388 , 363 kim , y .- h . , moon , y .-
cho , k .- s . , bonk , s .- ch .
, & park , y .- d .
2004 , j.korean astron.soc .
, 37 , 171 kim , y .- h . , moon , y .- j . ,
cho , k .- s . ,
kim , k .- s . , & park , y.d .
2005a , , 622 , 1240 kim , y .- h . , moon , y .-
cho , k .- s . , bong , s .- ch . , & park , y.d .
2005b , , 635 , 1291 kim , y .- h . , bong , s .- ch . , park , y .- d . ,
cho , k .- s . , & moon , y .- j .
2009 , , 705 , 1721 kiplinger , a.l .
1995 , , 453 , 973 kliem , b. , karlick@xmath35 , m. , & benz , a.o .
2000 , , 360 , 715 klimchuk , j.a . ,
acton , l.w .
, harvey , k.l . ,
hudson , h.s . , kluge , k.l . ,
sime , d.g . , strong , k.t . , & watanabe , t. 1994 , x - ray solar physics from _ yohkoh _ , y. , uchida , t. , watanabe , k. , shibata , & h.s . ,
hudson , universal academy press , 181 ko oma@xmath37ski , s. , tomczak , m. , ronowicz , p. , karlick@xmath35 , m. , & aurass , h. 2007 , cent .
astrophys . bull .
, 31 , 125 kosugi , t. , masuda , s. , makishima , k. , inda , m. , murakami , t. , dotani , t. , ogawara , y. , sakao , t. , kai , k. , & nakajima , h. 1991 , , 136 , 17 kundu , m.r .
, nindos , a. , vilmer , n. , klein , k .-
l . , shibata , k. , & ohyama , m. 2001 , , 559 , 443 masuda , s. , kosugi , t. , hara , h. , tsuneta , s. , & ogawara , y. 1994 , , 371 , 495 michaek , g. , gopalswamy , n. , & yashiro , s. 2003 , , 584 , 472 morrison , m. 1994 , _ yohkoh _ analysis guide , lmsc - p098510 munro , r.h . , & sime , d.g .
1985 , , 97 , 191 nakariakov , v.m . , & melnikov , v.f .
2009 , , 149 , 119 nishida , k. , shimizu , m. , shiota , d. , takasaki , h. , magara , t. , & shibata , k. 2009 , , 690 , 748 nizhizuka , n. , takasaki , h. , asai , a. , & shibata , k. 2010 , , 711 , 1062 nitta , n. , & akiyama , s. 1999 , , 525 , l57 nitta , n.v .
, cliver , e.w . , & tylka , a.j .
2003 , , 586 , l103 nitta , n.v . ,
freeland , s.l . , & liu , w. 2010 , , 725 , l28 ohyama , m. , & shibata , k. 1997 , , 49 , 249 ohyama , m. , & shibata , k. 1998 , , 499 , 934 ohyama , m. , & shibata , k. 2000 , j.atm .
solar terr.phys . , 62 , 1509 ohyama , m. , & shibata , k. 2008 , , 60 , 85 pallavicini , r. , serio , s. , & vaiana , g.s .
1977 , , 216 , 108 ronowicz , p. 2007
, master thesis , university of wrocaw ( in polish ) saint - hilaire , p. , & benz , a.o .
2003 , , 216 , 205 sato , j. , matsumoto , y. , yoshimura , k. , kubo , s. , kotoku , j. , masuda , s. , sawa , m. , suga , k. , yoshimori , m. , kosugi , t. , & watanabe , t. 2006 , , 236 , 351 shanmugaraju , a. , moon , y .-
j . , kim , y .- h . , cho , k .- s . , dryer , m. , & umapathy , s. 2006 , , 458 , 653 sheeley , n.r . ,
jr . , walters , j.h . , wang , y .- m . , & howard , r.a .
1999 , , 104 , 24739 shibata , k. , masuda , s. , shimojo , m. , hara , h. , yokoyama , t. , tsuneta , s. , kosugi , t. , & ogawara , y. 1995 , 451 , l83 shimizu , m. , nishida , k. , takasaki , h. , shiota , d. , magara , t. , & shibata , k. 2008 , , 683 , l203 @xmath38vestka , z. , & cliver , e.w .
1992 , in iau coll .
133 , eruptive solar flares , ed .
z. @xmath38vestka , b.v .
jackson , & m.e .
machado ( new york : springer ) , 1 takeda , a. , acton , l. , mckenzie , d. , yoshimura , k. , & freeland , s. 2009 , data sci .
j. , 8 , 24 september 2009 tandberg - hanssen , e. 1995 , the nature of solar prominences , kluwer , dordrecht tomczak , m. 2003 , esa sp-535 , proc .
iscs 2003 symposium , a. , wilson , noordwijk : european space agency , 465 tomczak , m. 2004 , , 417 , 1133 tomczak , m. 2005 , adv .
space res . , 35 , 1732 tomczak , m. 2008 , cent .
astrophys . bull .
, 32 , 59 tomczak , m. 2009 , , 502 , 665 tomczak , m. , & ronowicz , p. 2007
, 31 , 115 tsuneta , s. 1997 , , 483 , 507 tsuneta , s. , acton , l. , bruner , m. , lemen , j. , brown , w. , caravalho , r. , catura , r. , freeland , s. , jurcevich , b. , morrison , m. , ogawara , y. , hirayama , t. , & owens , j. 1991 , , 136 , 37 vr@xmath34nak , b. , sudar , d. , & ru@xmath39djak , d. 2005 , , 435 , 1149 wang , y. , & zhang , j. 2007 , , 665 , 1428 yashiro , s. , michaek , g. , akiyama , s. , gopalswamy , n. , & howard , r.a .
2008 , , 673 , 1174 rcccclllcl 001 & 91/10/22 & 06:42.2 - 07:28.8 & 2,1,2 & b & 6891 & m1.2 & s11e85 & & 19 + 002 & 91/11/02 & 16:31.2 - 16:57.3 & 2,2,1 & b & 6891 & m4.8 & s10w84 & & 19 + 003 & 91/11/17 & 18:32.9 - 18:41.8 & 1,1,2 & c & 6929 & m1.9 & s12e78 & & 19 + 004 & 91/12/02 & 04:50.6 - 05:21.0 & 2,2,1 & a & 6952 & m3.6 & n18e92 & & 18,19,27,28,54 + 005 & 91/12/03 & 16:35.0 - 17:04.5 & 2,1,1 & b & 6952 & x2.2 & n17e72 & & 19 + 006 & 91/12/09 & 02:02.7 - 02:06.5 & 2,1,1 & c & 6966 & m1.4 & s06e91 & & 19 + 007 & 91/12/10 & 04:03.0 - 04:10.1 & 1,1,1 & b & 6968 & c9.3 & s14e93 & & 19 + 008 & 92/01/13 & 17:27.9 - 17:35.2 & 2,2,2 & c & 6994 & m2.0 & ( s15w89 ) & & 18,19,27,28 + 009 & 92/01/13 & 19:04.1 - 19:13.8 & 1,2,1 & c & 7012 & m1.3 & s10e95 & & 19 + 010 & 92/01/14 & 19:29.0 - 19:32.6 & 2,2,1 & b & 7012 & m1.7 & s11e89 & & 19 + 011 & 92/01/15 & 18:56.1 - 19:04.5 & 2,2,1 & b & 7012 & m2.0 & ( s09e72 ) & & 19 + 012 & 92/01/30 & 17:07.6 - 17:17.6 & 2,1,1 & c & 7042 & m1.6 & s13e84 & & 19 + 013 & 92/02/06 & 03:17.4 - 03:36.6 & 2,2,1 & b & 7030 & m7.6 & n05w82 & & 18,19,27,28 + 014 & 92/02/06 & 20:52.7 - 21:24.8 & 2,2,1 & b & 7030 & m4.1 & n05w94 & & 19 + 015 & 92/02/09 & 03:01.0 - 03:10.7 & 2,1,1 & c & 7035 & m1.2 & s17w74 & & 19 + 016 & 92/02/17 & 15:41.8 - 16:25.0 & 1,2,2 & c & 7050 & m1.9 & n16w81 & & 18,19,27,28 + 017 & 92/02/18 & 18:00.1 - 18:29.9 & 2,1,1 & c & 7067 & + & ( n05e89 ) & & 13 + 018 & 92/02/19 & 14:45.4 - 15:39.1 & 2,2,1 & b & 7067 & m1.2 & n06e94 & & 19 + 019 & 92/02/21 & 03:11.9 - 03:19.2 & 2,1,1 & c & 7070 & m3.2 & ( n09e80 ) & & 19 + 020 & 92/02/21 & 22:04.6 - 22:08.0 & 2,1?,1 & c & 7070 & m2.2 & n05e65 & & 19 + 021 & 92/04/01 & 10:12.8 - 10:22.5 & 2,2,1 & b & 7123 & m2.3 & ( s03e89 ) & & 18,27,28 + 022 & 92/06/05 & 18:08.7 - 19:08.9 & 2,1,1 , & c & 7186 & c2.6 & n07e28 & & 13 + 023 & 92/06/07 & 01:41.0 - 01:50.0 & 2,2,1 & b & 7186 & m2.7 & n09e10 & & 4 + 024 & 92/07/20 & 17:18.1 - 17:49.6 & 1,1,1 & c & 7222 & & ( s06w88 ) & & 13 + 025 & 92/07/29 & 20:19.7 - 21:09.8 & 2,2,2 & b & 7236 & & ( n19w88 ) & & 13 + 026 & 92/08/25 & 19:02.6 - 19:36.4 & 1,1,1 & b & 7260 & c8.7 & n13w98 & & 19 + 027 & 92/09/09 & 02:06.2 - 02:18.6 & 1,1,1 & c & 7270 & m3.1 & s10w72 & & 19 + 028 & 92/09/09 & 17:57.9 - 18:07.2 & 1,2?,1 & c & 7270 & m1.9 & s11w78 & & 19 + 029 & 92/10/04 & 22:14.0 - 22:32.4 & 2,2,2 & b & 7293 & m2.4 & s05w90 & & 18,19,24,27,28,53 + 030 & 92/10/05 & 09:24.3 - 09:52.0 & 1,2,1 & b & 7293 & m2.0 & s08w90 & & 10,12,18,19,21,24,53 + 031 & 92/11/05 & 06:19.0 - 06:40.7 & & d & 7323 & m2.0 & s16w90 & & 18,19,27,28 + 032 & 92/11/05 & 20:30.1 - 21:08.6 & & d & 7323 & c8.7 & s17w92 & & 19 + 033 & 93/02/14 & 12:51.9 - 12:59.4 & 2,2,1 & b & 7427 & m2.0 & s22e78 & & 19 + 034 & 93/02/17 & 10:35.4 - 10:53.5 & 1,2,1 & b & 7420 & m5.8 & s07w87 & & 14,18,19,24,27,28,53 + 035 & 93/02/21 & 00:31.2 - 00:45.2 & 2,2,1 & b & 7433 & m1.4 & n13e75 & & 19 + 036 & 93/03/15 & 20:31.9 - 21:15.2 & 2,2,1 & b & 7440 & m2.9 & s03w93 & & 4 + 037 & 93/03/23 & 01:21.0 - 01:29.5 & 2,2,1 & b & 7448 & m2.3 & n18w78 & & 4 + 038 & 93/05/07 & 20:56.6 - 21:30.6 & 2,2,1 & b & 7500 & m1.6 & n14e41 & & 4 + 039 & 93/05/14 & 22:00.1 - 22:10.3 & 2,2,1 & a & 7500 & m4.4 & n19w48 & & 23 + 040 & 93/06/25 & 03:13.8 - 03:40.3 & 2,2,1 & b & 7530 & m5.1 & s09e88 & & 19 + 041 & 93/06/28 & 01:06.7 - 01:23.0 & 2,2,2 & b & 7535 & c6.5 & n03e69 & & 6 + 042 & 93/09/26 & 17:26.2 - 17:28.3 & 2,1,1 & b & 7590 & c3.4 & n14e94 & & 19 + 043 & 93/09/27 & 12:07.5 - 12:17.8 & 2,2,1 & b & 7590 & m1.8 & n08e90 & & 18,19 + 044 & 93/10/01 & 23:51.3 - 00:01.3 & 2,1,1 & c & 7592 & c8.5 & s14e69 & & 19 + 045 & 93/11/11 & 11:15.4 - 11:31.8 & 1,2,1 & b & 7618 & c9.7 & n10e95 & & 10,14,20 + 046 & 93/11/13 & 06:38.6 - 06:48.9 & 1,2?,1 & b & 7618 & m2.1 & n08e73 & & 19 + 047 & 94/01/05 & 06:49.4 - 06:59.6 & 2,2?,1 & b & 7647 & m1.0 & s13w23 & & 4 + 048 & 94/01/16 & 23:09.6 - 23:22.9 & 2,2,1 & b & 7654 & m6.1 & n05e71 & & 19 + 049 & 94/01/27 & 03:47.8 - 03:59.0 & 1,1,1 & b & 7654 & c4.6 & n08w68 & & 4 + 050 & 94/01/28 & 16:53.2 - 17:26.5 & 2,2,2 & b & 7654 & m1.8 & n08w85 & & 19 + 051 & 94/02/27 & 09:02.7 - 09:18.9 & 2,2,1 & b & 7671 & m2.8 & n08w98 & & 4 + 052 & 94/08/30 & 08:20.5 - 08:41.4 & 1,1,1 & b & 7773 & m1.1 & s06e82 & & 19 + 053 & 96/04/20 & 06:51.7 - 07:04.4 & 1,2,1 & a & 7956 & b2.9 & n04w68 & & 24,36,53 + 054 & 96/08/22 & 07:42.5 - 07:52.1 & 2,2,1 & b & 7986 & c4.5 & s14e107 & + & 3,50,52 + 055 & 97/02/23 & 01:30.2 - 02:15.2 & 2,2?,1 & b & 8019 & b7.2 & ( n31e90 + ) & & 2 + 056 & 97/02/23 & 02:58.2 - 03:31.0 & 2,1,1 & c & 8019 & b7.2 & ( n33e81 ) & + & 4 + 057 & 97/05/16 & 11:44.0 - 12:10.7 & 1,2,1 & c & 8038 & & ( n20w73 ) & + & 4 + 058 & 97/08/09 & 16:32.9 - 16:35.3 & 1,2?,1 & c & 8069 & c8.5 & n19w85 & + & 16,19 + 059 & 97/09/17 & 11:38.4 - 11:49.2 & 1,2,1 & b & 8084 & m1.7 & n21w82 & + & 16,19 + 060 & 97/09/17 & 17:48.4 - 18:24.6 & 2,2,2 & b & 8084 & m1.0 & n21w84 & + & 19 + 061 & 97/11/06 & 11:50.8 - 12:06.3 & 2,2,1 & c & 8100 & x9.4 & s18w63 & + & 17,48 + 062 & 97/11/14 & 09:11.0 - 09:20.8 & 2,1,1 & b & 8108 & c2.5 & n21e70 & + & 24,48,53 + 063 & 97/11/27 & 13:10.4 - 13:25.2 & 2,2,2 & a & 8113 & x2.6 & n17e63 & + & 29,48 + 064 & 98/03/23 & 02:45.9 - 03:12.1 & 2,2,2 & b & 8179 & m2.3 & s22w99 & + & 16,19,50 + 065 & 98/03/25 & 13:04.1 - 13:17.1 & 2,2,1 & c & 8180 & c5.3 & ( s37w90 + ) & + & 2 + 066 & 98/04/20 & 09:43.8 - 10:12.5 & 2,2?,2 & c & 8194 & m1.4 & s30w90 & + & 17 + 067 & 98/04/23 & 05:29.4 - 05:46.8 & 2,2,2 & a & 8210 & x1.2 & s18e104 & + & 1,10,16,18,19,24,26 , + & & & & & & & & & 50,51,52,53 + 068 & 98/04/24 & 08:47.6 - 08:55.2 & 2,1,1 & c & 8210 & c8.9 & s20e91 & + & 19 + 069 & 98/04/25 & 14:21.9 - 14:49.2 & 2,2,2 & b & 8210 & c3.6 & s19e73 & + & 2 + 070 & 98/04/27 & 08:50.0 - 08:55.3 & 2,2?,1 & b & 8210 & x1.0 & s16e50 & + & 34 + 071 & 98/05/03 & 21:17.4 - 21:25.5 & 2,2,2 & a & 8210 & m1.4 & s13w34 & + & 30 + 072 & 98/05/06 & 07:54.3 - 08:11.0 & 2,2,2 & a & 8210 & x2.7 & s11w65 & + & 17,24,35,48,53 + 073 & 98/05/08 & 01:50.4 - 02:20.1 & 2,2,2 & a & 8210 & m3.1 & ( s16w90 + ) & + & 16,18,19 + 074 & 98/05/08 & 14:21.2 - 14:49.4 & 2,2,2 & b & 8210 & m1.8 & s17w95 & + & 19 + 075 & 98/05/09 & 02:04.7 - 02:20.1 & 1,2,1 & c & 8210 & c7.0 & ( s15w90 + ) & & 16 + 076 & 98/05/09 & 03:17.7 - 03:37.3 & 2,2,2 & a & 8210 & m7.7 & s17w102 & + & 16,18,19,50 + 077 & 98/05/10 & 13:25.4 - 13:33.2 & 2,2?,1 & b & 8220 & m3.9 & s27e89 & & 19 + 078 & 98/05/28 & 06:22.5 - 06:28.2 & 1,1,1 & b & 8226 & c1.2 & n20w78 & & 19 + 079 & 98/05/28 & 19:02.4 - 19:17.6 & 2,2,2 & b & 8226 & c8.7 & ( n16w87 ) & + & 16,19 + 080 & 98/05/29 & 01:02.8 - 01:12.0 & 1,2,1 & c & 8226 & m6.7 & ( n15w89 ) & + & 19 + 081 & 98/06/11 & 10:00.8 - 10:18.3 & 2,2,2 & c & 8243 & m1.4 & ( n21e89 ) & + & 16,50 + 082 & 98/06/15 & 07:26.8 - 07:34.3 & 2,2,1 & c & 8232 & c1.4 & ( s16w87 ) & + & 2 + 083 & 98/07/03 & 01:06.5 - 01:12.7 & 2,2?,1 & c & 8256 & m1.2 & s26w14 & & 4 + 084 & 98/08/14 & 08:25.5 - 08:31.7 & 2,2,1 & b & 8293 & m3.1 & s23w74 & & 19 + 085 & 98/08/18 & 08:17.5 - 08:26.5 & 2,2,2 & b & 8307 & x2.8 & n33e68 & & 18,19,48 + 086 & 98/08/18 & 22:15.1 - 22:20.7 & 2,2,1 & b & 8307 & x4.9 & n33e87 & & 18,19,31,48 + 087 & 98/09/06 & 02:12.3 - 02:18.5 & 1,1?,1 & b & 8323 & c2.3 & s22w35 & & 4 + 088 & 98/09/20 & 02:38.5 - 03:19.1 & 2,2,2 & b & 8340 & m1.8 & n22e62 & & 4 + 089 & 98/09/23 & 00:30.2 - 00:47.6 & 2,2,1 & b & 8344 & c9.3 & s20e22 & & 4 + 090 & 98/09/23 & 06:45.3 - 07:05.2 & 2,2,2 & b & 8340 & m7.1 & n18e09 & & 40 + 091 & 98/09/28 & 16:07.9 - 16:11.7 & 2,2,1 & c & 8339 & c6.8 & ( s15w87 ) & & 4 + 092 & 98/09/30 & 13:16.4 - 13:40.5 & 1,2,2 & b & 8340 & m2.8 & n23w81 & & 4 + 093 & 98/10/07 & 17:08.6 - 17:44.8 & 2,2,1 & a & 8355 & m2.3 & s23e68 & & 32 + 094 & 98/10/20 & 20:34.5 - 21:16.1 & 2,2,2 & c & 8360 & c7.4 & ( s21w86 ) & & 4 + 095 & 98/11/03 & 19:14.3 - 19:36.3 & 2,1?,2 & b & 8375 & m1.0 & n21e02 & + & 4 + 096 & 98/11/05 & 01:03.4 - 01:12.2 & 2,1,2 & c & 8375 & c7.1 & ( n20w14 ) & + & 33 + 097 & 98/11/06 & 02:42.1 - 02:51.8 & 1,1,1 & c & 8375 & c4.4 & n19w24 & & 4 + 098 & 98/11/06 & 09:09.9 - 09:16.6 & 1,1?,1 & c & 8375 & c4.2 & n19w27 & + & 4 + 099 & 98/11/07 & 17:52.1 - 18:01.1 & 1,2?,1 & b & 8375 & c5.3 & n19w48 & & 4 + 100 & 98/11/16 & 23:11.4 - 23:18.7 & 2,2,2 & b & 8385 & c6.8 & ( n24w83 ) & & 4 + 101 & 98/11/22 & 06:39.6 - 06:47.2 & 2,2,2 & c & 8384 & x3.7 & s27w82 & & 28,34 + 102 & 98/11/22 & 16:21.5 - 16:31.1 & 1,2,2 & b & 8384 & x2.5 & s30w89 & & 28,34 + 103 & 98/11/24 & 02:19.6 - 02:28.3 & 2,2,1 & c & 8384 & x1.0 & s30w103 & + & 4 + 104 & 98/11/25 & 14:00.8 - 14:07.5 & 2,2,1 & c & 8395 & c6.4 & n18e68 & + & 4 + 105 & 98/11/28 & 05:29.5 - 05:49.6 & 2,2,2 & b & 8395 & x3.3 & n17e32 & + & 34,48 + 106 & 98/12/18 & 17:17.5 - 17:27.9 & 2,2,1 & c & 8414 & m8.0 & ( n34e40 ) & + & 4 + 107 & 98/12/23 & 05:53.5 - 06:02.9 & 2,1,1 & c & 8414 & m2.3 & ( n28e70 ) & & 4 + 108 & 98/12/24 & 01:22.2 - 01:29.1 & 2,1,1 & c & 8421 & c6.2 & n29e85 & & 4 + 109 & 98/12/25 & 06:18.4 - 06:35.0 & 2,1?,2 & b & 8421 & m1.2 & n30e66 & & 4 + 110 & 98/12/28 & 05:16.2 - 05:31.9 & 2,1,2 & b & 8416 & m1.4 & n28e26 & & 4 + 111 & 99/01/03 & 08:15.4 - 09:03.8 & 2,2,2 & a & 8420 & c8.2 & n16w70 & & 4 + 112 & 99/01/06 & 23:59.7 - 00:05.4 & 1,2,1 & b & 8422 & c8.0 & ( s23w90 ) & & 4 + 113 & 99/01/14 & 10:06.9 - 10:17.0 & 2,2,1 & b & 8439 & m3.0 & n18e64 & & 4 + 114 & 99/02/10 & 23:09.1 - 23:25.3 & 2,2?,1 & c & 8457 & c3.7 & ( n18e57 ) & + & 4 + 115 & 99/02/12 & 15:23.7 - 15:32.1 & 2,2?,1 & b & 8456 & c5.7 & n12e27 & & 4 + 116 & 99/02/16 & 21:19.7 - 21:27.8 & 2,2,2 & c & 8462 & c5.3 & n19w12 & & 4 + 117 & 99/02/21 & 13:16.0 - 13:41.5 & 1,1?,2 & b & 8462 & m1.3 & n24w81 & & 4 + 118 & 99/03/07 & 04:05.5 - 04:11.9 & 2,2,1 & c & 8477 & c2.9 & n24w03 & + & 4 + 119 & 99/04/03 & 23:02.0 - 23:10.5 & 1,1,1 & b & 8508 & m4.3 & n29e81 & + & 9 + 120 & 99/05/07 & 04:31.5 - 04:45.0 & 2,2,2 & a & 8535 & m3.2 & n20e87 & + & 9,10 + 121 & 99/05/08 & 14:33.0 - 14:52.1 & 1,1,1 & c & 8526 & m4.6 & n23w75 & + & 9 + 122 & 99/05/09 & 12:34.7 - 13:05.5 & 2,1?,2 & c & 8526 & m1.0 & ( n22w84 ) & + & 9 + 123 & 99/05/09 & 18:04.2 - 18:17.3 & 1,2,1 & b & 8526 & m7.6 & ( n22w86 ) & + & 8,9 + 124 & 99/05/11 & 21:46.9 - 22:02.7 & 2,1,1 & c & 8542 & c4.7 & s19e79 & + & 9 + 125 & 99/05/16 & 17:26.2 - 17:32.9 & 1,1,1 & b & 8534 & m1.1 & s17w76 & & 9 + 126 & 99/05/17 & 03:43.2 - 03:51.6 & 1,1,2 & b & 8534 & c3.9 & ( s16w82 ) & & 8,9,24,37,53 + 127 & 99/05/17 & 04:57.5 - 05:01.0 & 1,1,1 & c & 8534 & m2.3 & ( s15w82 ) & & 9 + 128 & 99/05/29 & 03:08.1 - 03:14.2 & 2,2,1 & b & 8557 & m1.6 & ( s21e65 ) & + & 9 + 129 & 99/06/11 & 11:25.4 - 11:38.4 & 2,2,1 & b & 8585 & c8.8 & ( n46e90 ) & + & 9 + 130 & 99/06/19 & 22:31.8 - 22:42.8 & 2,1,1 & c & 8592 & c4.0 & ( n25e86 ) & + & 9 + 131 & 99/06/23 & 00:41.2 - 00:51.0 & 2,1,1 & b & 8583 & c7.9 & s12w78 & + & 9 + 132 & 99/06/30 & 03:11.0 - 03:18.3 & 1,1,1 & c & 8613 & c4.5 & ( n19e90 ) & + & 9 + 133 & 99/06/30 & 04:36.2 - 05:00.8 & 2,2,2 & b & 8611 & m2.1 & s26e28 & + & 4 + 134 & 99/07/07 & 06:20.7 - 06:30.3 & 1,2?,1 & c & 8611 & c1.2 & s26w76 & + & 9 + 135 & 99/07/07 & 09:19.1 - 09:24.4 & 1,1,1 & c & 8611 & c1.2 & ( s26w76 ) & + & 9 + 136 & 99/07/09 & 22:38.5 - 22:48.2 & 1,1,1 & c & 8629 & c6.9 & n22w69 & & 9 + 137 & 99/07/12 & 23:31.9 - 23:38.0 & 1,1,1 & b & 8626 & c2.1 & ( s19w76 ) & & 9 + 138 & 99/07/16 & 15:50.8 - 15:55.3 & 1,1,1 & c & 8635 & m3.1 & n43w71 & + & 9 + 139 & 99/07/23 & 04:45.9 - 05:03.8 & 2,1,1 & b & 8645 & c9.3 & s23e97 & & 9 + 140 & 99/07/23 & 15:56.1 - 16:26.3 & 1,1,1 & c & 8645 & m1.0 & s26w92 & & 9 + 141 & 99/07/23 & 22:58.1 - 23:10.9 & 2,1,1 & c & 8644 & m1.2 & s26e87 & & 9 + 142 & 99/07/24 & 03:31.7 - 04:06.6 & 2,1,1 & c & 8636 & m1.7 & s29e87 & & 9 + 143 & 99/07/24 & 08:00.4 - 08:16.7 & 2,1?,1 & b & 8636 & m3.3 & s28e78 & + & 9 + 144 & 99/07/25 & 13:08.8 - 13:14.7 & 2,2,1 & b & 8639 & m2.4 & n38w81 & + & 8,9,10,24,38,53 + 145 & 99/08/04 & 05:58.5 - 06:11.5 & 2,1,1 & c & 8647 & m6.0 & s16w64 & + & 9 + 146 & 99/08/04 & 21:54.2 - 22:03.2 & 1,1,1 & c & 8647 & c3.3 & ( s18w80 ) & & 9 + 147 & 99/08/06 & 10:05.3 - 10:18.0 & 2,1,1 & c & 8647 & c7.1 & s28w81 & + & 9 + 148 & 99/08/07 & 20:52.0 - 21:10.9 & 2,2,2 & b & 8645 & m1.7 & s29w104 & + & 9 + 149 & 99/08/10 & 06:02.4 - 06:13.2 & 1,1,1 & c & 8656 & c4.8 & n15w74 & & 9 + 150 & 99/08/13 & 15:07.9 - 15:16.2 & 1,1,1 & b & 8668 & c5.3 & ( n17e90 ) & & 9 + 151 & 99/08/14 & 12:06.0 - 12:16.9 & 1,1,1 & c & 8668 & c7.2 & n23e72 & & 9 + 152 & 99/08/20 & 12:40.4 - 13:36.6 & 2,2?,2 & c & 8674 & m1.8 & s28e76 & + & 9 + 153 & 99/08/28 & 17:54.1 - 18:04.6 & 2,2,2 & b & 8674 & x1.1 & s26w14 & + & 48 + 154 & 99/09/08 & 12:13.6 - 12:14.5 & 2,2?,1 & c & 8690 & m1.4 & n12e53 & + & 25 + 155 & 99/09/21 & 10:16.0 - 10:34.6 & 2,2,1 & b & 8692 & c6.4 & ( s25w84 ) & + & 9 + 156 & 99/10/25 & 19:37.8 - 19:44.1 & 2,2,1 & c & 8737 & c4.8 & s19w69 & + & 9 + 157 & 99/10/26 & 21:12.0 - 21:27.6 & 2,2,2 & b & 8737 & m3.7 & ( s16w86 ) & + & 9,10,11,18,28,39 + 158 & 99/10/27 & 09:11.2 - 09:13.9 & 2,2,1 & b & 8737 & m1.0 & s12w88 & & 9 + 159 & 99/10/27 & 13:35.2 - 14:04.9 & 2,2,1 & b & 8737 & m1.8 & s15w90 & + & 9 + 160 & 99/10/27 & 15:21.1 - 15:43.6 & 2,1,1 & b & 8737 & m1.4 & s14w92 & + & 9 + 161 & 99/11/05 & 18:22.5 - 18:48.6 & 2,2?,1 & b & 8759 & m3.0 & n12e96 & + & 9,10 + 162 & 99/11/06 & 06:32.5 - 06:35.1 & 1,1,1 & c & 8759 & c4.6 & ( n12e86 ) & & 8,9 + 163 & 99/11/06 & 17:04.4 - 17:14.4 & 1,1,1 & b & 8759 & c5.0 & ( n12e86 ) & + & 9 + 164 & 99/11/08 & 06:06.0 - 06:16.0 & & d & 8749 & c5.9 & s18w82 & + & 9 + 165 & 99/11/13 & 02:22.7 - 03:09.4 & 2,2?,1 & c & 8763 & m1.3 & s15e44 & + & 9 + 166 & 99/11/27 & 12:12.1 - 12:23.3 & 1,1,1 & c & 8771 & x1.4 & s15w68 & + & 9,48,49 + 167 & 99/12/18 & 01:26.8 - 01:39.1 & 2,2,1 & c & 8806 & c9.4 & n19e82 & & 9 + 168 & 00/01/12 & 20:50.0 - 20:56.9 & 2,2,2 & b & 8829 & m1.1 & n13e67 & + & 9 + 169 & 00/01/18 & 09:36.0 - 09:58.8 & 1,2,1 & a & 8827 & m1.2 & s15w106 & + & 9,10,24,53 + 170 & 00/01/18 & 17:12.1 - 17:18.1 & 2,2?,1 & c & 8831 & m3.9 & s19e11 & + & 4 + 171 & 00/01/22 & 17:58.9 - 18:08.8 & 2,1,2 & c & 8831 & m1.0 & s23w50 & + & 4 + 172 & 00/02/04 & 09:14.6 - 09:57.4 & 2,1,1 & c & 8858 & m3.0 & n25e71 & & 28 + 173 & 00/02/04 & 19:28.3 - 19:43.5 & 1,2,2 & b & 8858 & c7.0 & n25e71 & + & 9 + 174 & 00/02/05 & 19:33.9 - 19:43.9 & 2,1?,2 & b & 8858 & x1.2 & n26e52 & + & 48 + 175 & 00/02/22 & 20:26.2 - 21:24.0 & 2,1,1 & c & 8882 & c9.2 & ( s18e90 + ) & + & 9 + 176 & 00/02/26 & 23:38.5 - 23:43.8 & 2,2,2 & b & 8889 & m1.0 & n29e50 & & 4 + 177 & 00/03/02 & 08:23.7 - 08:28.7 & 2,2,1 & b & 8882 & x1.1 & ( s18w55 ) & + & 48,49 + 178 & 00/03/02 & 13:12.5 - 13:20.5 & 1,2,1 & c & 8882 & c5.5 & s19w60 & + & 4 + 179 & 00/03/03 & 02:11.6 - 02:14.6 & 2,2,1 & b & 8882 & m3.8 & s15w60 & + & 4 + 180 & 00/03/06 & 10:47.9 - 10:50.0 & 1,1,1 & b & 8889 & c4.5 & ( n20w75 ) & & 9 + 181 & 00/03/06 & 16:20.0 - 16:28.4 & 2,1,1 & b & 8889 & c3.9 & n20w78 & & 9 + 182 & 00/03/18 & 20:50.7 - 20:57.3 & 2,2,1 & b & 8906 & m2.1 & ( s15w68 ) & + & 4 + 183 & 00/03/18 & 21:53.5 - 21:59.0 & 1,1,1 & b & 8906 & c4.2 & s19w67 & + & 4 + 184 & 00/03/18 & 23:16.9 - 23:27.4 & 2,1,1 & b & 8902 & + & ( s18e90 + ) & + & 9 + 185 & 00/03/27 & 13:59.1 - 14:25.5 & & d & 8926 & c8.4 & s09w69 & & 9 + 186 & 00/03/27 & 15:32.8 - 15:41.9 & 2,1,1 & c & 8926 & c8.9 & s10w69 & & 24 + 187 & 00/03/31 & 06:27.8 - 06:50.8 & 2,2?,1 & b & 8936 & m1.2 & s15e55 & + & 9 + 188 & 00/04/06 & 02:22.4 - 02:31.9 & 2,2,1 & b & 8948 & m1.8 & s15e53 & + & 4 + 189 & 00/04/08 & 02:37.5 - 02:47.5 & 2,2,1 & c & 8948 & m2.0 & s15e26 & + & 4 + 190 & 00/05/02 & 14:53.2 - 15:03.0 & 1,1,2 & c & 8971 & m2.8 & n22w68 & + & 9 + 191 & 00/05/05 & 15:18.3 - 15:47.5 & 2,2,2 & b & 8977 & m1.5 & s18w110 & + & 9 + 192 & 00/05/12 & 08:40.9 - 08:50.7 & & d & 8998 & c8.1 & s14e90 & & 9,41 + 193 & 00/05/12 & 21:32.8 - 21:39.8 & & d & 8998 & c9.8 & s15e81 & & 9 + 194 & 00/05/13 & 01:46.6 - 02:10.1 & 2,2,1 & c & 9002 & m1.1 & n22e109 & & 9 + 195 & 00/05/13 & 23:12.4 - 23:21.8 & 2,1,1 & c & 9002 & c7.4 & n22e96 & + & 9 + 196 & 00/05/15 & 08:26.8 - 08:53.1 & 2,2,1 & c & 9002 & m4.4 & ( n23e87 ) & + & 9 + 197 & 00/05/18 & 15:55.6 - 16:05.0 & 2,2?,1 & c & 9002 & m2.7 & n23e30 & + & 4 + 198 & 00/05/23 & 17:50.1 - 18:00.1 & 1,1,1 & c & 8996 & c4.3 & s22w80 & + & 9 + 199 & 00/05/24 & 00:09.2 - 00:19.1 & 1,1,1 & c & 9017 & c6.8 & ( s12e90 + ) & & 9 + 200 & 00/05/24 & 03:14.3 - 04:02.8 & 1,1,1 & c & 9017 & c7.0 & s12e93 & & 9 + 201 & 00/05/24 & 11:43.3 - 12:13.0 & 1,1,1 & c & 9017 & m1.1 & ( s12e90 ) & & 9 + 202 & 00/05/24 & 21:05.6 - 21:51.1 & 1,1,1 & c & 9017 & c9.7 & ( s12e90 ) & & 9 + 203 & 00/05/26 & 11:33.1 - 11.37.6 & 2,2?,1 & b & 8998 & c6.1 & s13w90 & + & 9 + 204 & 00/05/28 & 10:23.6 - 10:34.6 & 1,1,1 & c & 9002 & c8.6 & ( n21w88 ) & + & 9 + 205 & 00/06/02 & 03:44.6 - 03:50.8 & 2,1,1 & b & 9026 & m1.2 & ( n22e77 ) & + & 9 + 206 & 00/06/02 & 20:32.7 - 21:10.3 & 2,1,1 & b & 9026 & m3.1 & ( n21e67 ) & + & 9 + 207 & 00/06/06 & 15:35.6 - 15:45.6 & 1,1,1 & c & 9026 & x2.3 & ( n20e18 ) & + & 4 + 208 & 00/06/07 & 15:42.9 - 15:52.9 & 2,2?,1 & b & 9026 & x1.2 & n23e03 & + & 4 + 209 & 00/06/12 & 12:24.4 - 12:57.5 & 2,1,1 & c & 9042 & c6.4 & n16e87 & & 9 + 210 & 00/06/17 & 02:29.6 - 02:45.3 & 1,1,2 & c & 9033 & m3.5 & n22w72 & + & 9 + 211 & 00/06/18 & 02:04.4 - 02:14.4 & 1,2,1 & c & 9033 & x1.0 & n23w85 & + & 9 + 212 & 00/06/19 & 04:18.1 - 04:23.0 & 1,1,1 & c & 9033 & c1.7 & ( n21w89 ) & + & 9 + 213 & 00/06/21 & 09:24.5 - 09:34.1 & 1,1,1 & c & 9042 & m1.3 & n24w42 & + & 4 + 214 & 00/06/23 & 14:22.7 - 14:37.9 & 2,2,1 & b & 9042 & m3.0 & n26w72 & + & 9 + 215 & 00/06/23 & 22:14.9 - 22:24.8 & 1,1?,2 & c & 9042 & c7.7 & n22w74 & + & 9 + 216 & 00/06/28 & 12:17.4 - 12:19.5 & 2,1?,1 & b & 9064 & c6.1 & ( s22w73 ) & & 9 + 217 & 00/06/29 & 10:28.6 - 10:38.7 & 1,1,1 & b & 9064 & c4.8 & ( s17w82 ) & & 9 + 218 & 00/07/01 & 12:37.6 - 12:43.1 & 1,1,1 & b & 9054 & c6.0 & ( n15w75 ) & & 9 + 219 & 00/07/01 & 23:21.4 - 23:31.3 & 2,1,1 & c & 9054 & m1.5 & n07w88 & & 9 + 220 & 00/07/10 & 21:14.7 - 21:22.7 & 2,2,2 & b & 9077 & m5.7 & n18e49 & + & 4 + 221 & 00/07/12 & 03:32.6 - 03:37.5 & 1,1,1 & c & 9078 & m1.4 & ( s11e87 ) & + & 9 + 222 & 00/07/12 & 10:26.4 - 10:38.7 & 2,2,1 & c & 9077 & x1.9 & n17e27 & + & 48 + 223 & 00/07/12 & 16:24.9 - 16:32.7 & 2,1,1 & c & 9070 & m1.0 & n17w68 & + & 9 + 224 & 00/07/12 & 19:46.1 - 20:18.1 & 2,2,1 & c & 9077 & m1.5 & ( n17w72 ) & + & 9 + 225 & 00/07/13 & 02:04.5 - 02:14.2 & 1,1,1 & b & 9069 & c6.1 & s16w71 & & 9 + 226 & 00/07/13 & 06:59.8 - 07:09.3 & 2,1,1 & c & 9070 & c6.8 & ( n16w80 ) & & 9 + 227 & 00/07/13 & 18:14.1 - 19:04.8 & & d & 9070 & m1.3 & ( n16w83 ) & & 9 + 228 & 00/07/14 & 00:42.0 - 00:44.6 & 1,1,1 & b & 9070 & m1.5 & ( n17w86 ) & & 9 + 229 & 00/07/14 & 10:20.0 - 10:27.2 & 2,2,2 & b & 9077 & x5.7 & n22w07 & + & 48,49 + 230 & 00/07/14 & 13:45.1 - 13:55.0 & 2,2,1 & b & 9077 & m3.7 & n20w08 & & 4 + 231 & 00/07/16 & 01:24.5 - 01:32.5 & 2,1,1 & b & 9087 & c6.3 & s11e53 & & 4 + 232 & 00/07/18 & 05:02.9 - 05:10.5 & 2,2,2 & b & 9077 & m1.8 & n17w58 & & 4 + 233 & 00/07/20 & 09:46.1 - 10:18.7 & 2,2?,2 & c & 9087 & m3.6 & s12w08 & & 24 + 234 & 00/07/21 & 14:33.3 - 14:40.6 & 2,1,2 & c & 9090 & m5.5 & ( n10e12 ) & & 4 + 235 & 00/07/22 & 11:21.1 - 11:27.1 & 2,2,1 & b & 9085 & m3.7 & n14w56 & + & 4 + 236 & 00/07/25 & 02:46.8 - 02:52.4 & 2,2,1 & b & 9097 & m8.0 & n06w08 & + & 42 + 237 & 00/07/26 & 03:54.6 - 04:13.6 & 1,1,1 & c & 9087 & c8.9 & s13w89 & & 9 + 238 & 00/07/27 & 04:08.1 - 04:17.9 & 2,1,1 & b & 9090 & m2.4 & n10w72 & & 9 + 239 & 00/07/27 & 16:46.0 - 16:52.6 & 2,1,1 & b & 9087 & m1.5 & s09w105 & & 9 + 240 & 00/08/02 & 08:17.8 - 08:27.1 & 1,1,1 & b & 9114 & c7.9 & ( n10e85 ) & + & 9 + 241 & 00/08/12 & 09:49.6 - 10:00.5 & 2,2,2 & b & 9119 & m1.1 & ( s15w84 ) & + & 9 + 242 & 00/08/14 & 05:01.1 - 05:10.8 & 1,1,1 & c & 9126 & c8.1 & n06w75 & + & 9 + 243 & 00/08/24 & 09:01.8 - 09:11.6 & 2,1,1 & c & ? & c6.2 & ( n27w88 ) & & 9 + 244 & 00/08/25 & 14:27.7 - 14:32.5 & 2,2,1 & b & 9143 & m1.4 & s15e67 & + & 7,9,25,43 + 245 & 00/09/07 & 20:38.0 - 20:47.6 & 1,1,2 & b & 9151 & c7.2 & n06w47 & + & 44 + 246 & 00/09/15 & 14:41.2 - 14:43.9 & 1,2?,1 & c & 9165 & m2.0 & n12e07 & + & 24 + 247 & 00/09/22 & 23:46.9 - 23:56.5 & 1,1,1 & c & 9165 & c8.5 & n14w94 & & 9 + 248 & 00/09/30 & 17:53.4 - 18:21.1 & 2,1,1 & b & ? & m1.0 & ( s29e85 ) & + & 9 + 249 & 00/09/30 & 23:17.3 - 23:29.0 & 2,2,1 & b & 9169 & x1.2 & n07w90 + & & 9,18,28,45,48 + 250 & 00/10/01 & 07:01.7 - 07:14.2 & 2,1,1 & c & 9169 & m5.0 & n08w97 & & 9 + 251 & 00/10/01 & 13:59.2 - 14:06.5 & 2,1,1 & a & 9169 & m2.2 & n09w101 & & 9 + 252 & 00/10/16 & 05:35.1 - 06:03.9 & 2,2,1 & a & 9182 & c7.0 & n04w107 & + & 9,24,53 + 253 & 00/10/16 & 06:42.5 - 06:46.9 & 2,2?,1 & b & 9182 & m2.5 & n03w108 & + & 9 + 254 & 00/10/26 & 05:10.2 - 05:18.2 & 1,1,1 & b & 9199 & c3.7 & ( n14w79 ) & + & 8,9 + 255 & 00/10/26 & 06:11.2 - 06:20.7 & 2,1,1 & c & 9209 & c4.3 & s25e71 & & 9 + 256 & 00/10/26 & 11:43.9 - 11:46.9 & & d & 9203 & c6.9 & n17w77 & + & 9 + 257 & 00/10/26 & 15:59.1 - 16:35.9 & 2,1?,2 & b & 9209 & c8.5 & s20e64 & + & 9 + 258 & 00/10/29 & 01:32.8 - 01:57.4 & 2,2,2 & b & 9209 & m4.4 & s25e35 & & 4 + 259 & 00/11/01 & 12:04.5 - 12:21.1 & 1,1,1 & c & 9212 & & ( n10e29 ) & & 24 + 260 & 00/11/08 & 23:18.2 - 23:44.0 & 2,1,1 & c & 9213 & m7.4 & ( n10w77 ) & + & 9,17 + 261 & 00/11/09 & 03:03.2 - 03:12.9 & 2,1?,1 & c & 9213 & m1.2 & ( n08w86 ) & & 9 + 262 & 00/11/09 & 06:28.9 - 06:38.6 & 1,1,1 & c & 9213 & m1.2 & ( n12w88 ) & & 9 + 263 & 00/11/14 & 16:26.2 - 16:35.9 & 2,2,2 & b & 9232 & m1.0 & ( n14e90 ) & + & 9 + 264 & 00/11/18 & 16:48.3 - 16:57.9 & 2,1,1 & c & 9227 & c5.9 & ( s16w79 ) & & 9 + 265 & 00/11/24 & 14:55.0 - 15:20.5 & 2,2,2 & a & 9236 & x2.3 & n22w07 & + & 46,48 + 266 & 00/11/24 & 21:50.2 - 22:03.5 & 2,2,2 & b & 9236 & x1.8 & n21w14 & + & 15,48 + 267 & 00/11/25 & 01:00.1 - 01:10.4 & 2,2,2 & a & 9240 & m8.2 & n07e50 & + & 4 + 268 & 00/11/25 & 09:10.0 - 09:21.0 & 2,2,2 & b & 9236 & m3.5 & n18w24 & + & 4 + 269 & 00/11/25 & 18:36.5 - 18:46.4 & 2,1,1 & c & 9236 & x1.9 & n20w23 & + & 48 + 270 & 00/11/26 & 16:38.3 - 16:47.4 & 2,1,2 & b & 9236 & x4.0 & n18w38 & + & 48 + 271 & 00/11/30 & 08:58.0 - 09:23.5 & 2,2,1 & a & 9236 & m1.0 & ( n17w88 ) & & 4 + 272 & 00/12/06 & 22:22.0 - 22:47.5 & 2,2?,1 & c & 9246 & m1.6 & s10w66 & + & 9 + 273 & 00/12/18 & 08:58.6 - 09:06.2 & 1,1,1 & c & 9276 & c5.2 & s14w76 & & 9 + 274 & 00/12/19 & 10:23.5 - 11:05.1 & 1,1,1 & c & 9276 & + & ( s12w85 ) & + & 9 + 275 & 00/12/23 & 04:57.9 - 05:07.5 & 1,1,1 & c & 9283 & c3.4 & s12e76 & & 9 + 276 & 00/12/23 & 08:20.2 - 08:23.1 & 1,1,1 & c & 9283 & c3.7 & s13e75 & & 9 + 277 & 00/12/24 & 01:06.7 - 01:15.0 & 2,1,1 & c & 9283 & c7.0 & s15e66 & + & 9 + 278 & 00/12/26 & 23:47.9 - 23:56.1 & 1,1,1 & b & 9289 & c4.0 & ( s06e87 ) & & 9 + 279 & 00/12/27 & 15:52.0 - 16:05.3 & 2,1,1 & c & 9289 & m4.3 & s07e73 & & 9 + 280 & 01/01/04 & 08:56.9 - 09:06.0 & 2,1,1 & b & 9302 & c4.5 & n25e87 & & 9 + 281 & 01/01/05 & 18:29.5 - 18:39.3 & 2,1,1 & a & 9302 & c5.8 & n20e72 & & 9 + 282 & 01/01/08 & 10:36.5 - 10:53.3 & 2,1?,1 & b & 9302 & c5.1 & ( n21e36 ) & + & 24 + 283 & 01/01/09 & 08:50.0 - 08:59.5 & 2,1,1 & b & 9297 & c5.1 & ( n21w87 ) & & 9 + 284 & 01/01/19 & 17:09.2 - 18:08.4 & 1,1,1 & c & 9313 & m1.0 & s07e61 & + & 9 + 285 & 01/01/24 & 14:42.7 - 15:11.4 & & d & 9311 & m1.0 & n06w77 & & 9 + 286 & 01/01/25 & 07:10.8 - 07:18.5 & 2,2,1 & c & 9325 & c7.4 & ( n10e74 ) & + & 9 + 287 & 01/01/30 & 00:57.1 - 01:03.2 & 2,2,1 & c & 9313 & c3.7 & ( s08w89 ) & + & 9 + 288 & 01/02/19 & 20:55.4 - 21:05.1 & 2,1,1 & b & 9360 & c5.4 & ( s09e89 ) & + & 9 + 289 & 01/03/06 & 10:10.1 - 10:19.7 & 2,2,1 & b & 9364 & c6.7 & ( s11w89 ) & + & 9 + 290 & 01/03/07 & 14:49.2 - 14:59.1 & 2,1,1 & c & 9371 & c5.8 & n23w75 & + & 9 + 291 & 01/03/09 & 19:54.4 - 20:00.5 & 1,1,1 & c & 9371 & c4.2 & ( n22w84 ) & + & 9 + 292 & 01/03/10 & 17:14.3 - 17:23.8 & 1,1,1 & c & 9365 & c5.9 & ( s12w90 ) & + & 9 + 293 & 01/03/11 & 08:47.4 - 08:57.2 & 2,1,1 & b & 9376 & c5.0 & s14e87 & + & 9,24 + 294 & 01/03/21 & 02:36.6 - 02:44.4 & 2,2?,2 & b & 9373 & m1.8 & s05w65 & + & 9 + 295 & 01/03/21 & 11:25.2 - 11:31.5 & 1,1,1 & b & 9373 & c9.8 & s05w70 & + & 9 + 296 & 01/03/24 & 01:36.5 - 01:40.5 & 2,2?,1 & c & 9376 & m1.2 & s14w82 & + & 9 + 297 & 01/03/24 & 23:32.6 - 00:08.8 & 1,1,1 & c & 9393 & m1.1 & n19e60 & & 9 + 298 & 01/03/28 & 10:56.0 - 11:05.7 & 2,1,1 & b & 9397 & m4.3 & s09e29 & + & 24 + 299 & 01/03/29 & 01:37.4 - 01:58.0 & 1,1,1 & c & 9393 & & ( n15w09 ) & & 24 + 300 & 91/03/29 & 12:47.1 - 13:02.3 & 2,1,1 & c & 9393 & c7.6 & ( n16w11 ) & & 24 + 301 & 01/04/01 & 10:59.8 - 11:03.5 & 1,2,1 & b & 9415 & m5.5 & s21e107 & + & 4 + 302 & 01/04/03 & 03:37.6 - 03:49.3 & 1,1?,1 & c & 9415 & x1.2 & s21e83 & + & 4 + 303 & 01/04/05 & 08:30.7 - 08:49.5 & 2,2,1 & b & 9393 & m8.4 & n14w103 & + & 24,53 + 304 & 01/04/06 & 19:12.5 - 19:23.5 & 2,2,2 & b & 9415 & x5.6 & s21e31 & + & 48 + 305 & 01/04/09 & 15:23.6 - 15:24.8 & 2,1,1 & c & 9415 & m7.9 & s21w04 & + & 24 + 306 & 01/04/10 & 05:19.2 - 05:33.9 & 1,1,1 & c & 9415 & x2.3 & s23w09 & + & 4 + 307 & 01/04/12 & 10:12.8 - 10:22.0 & 2,2,1 & b & 9415 & x2.0 & s19w43 & + & 47,48 + 308 & 01/04/15 & 13:29.7 - 13:53.0 & 2,2,2 & b & 9415 & y1.4 & s20w85 & + & 17,18,48 + 309 & 01/04/17 & 12:15.4 - 12:20.7 & 2,2?,1 & c & 9415 & c1.9 & ( s21w90 ) & & 4 + 310 & 01/04/20 & 21:30.7 - 21:35.5 & 2,2,2 & b & 9433 & c8.0 & ( n17e45 ) & & 4 + 311 & 01/04/23 & 10:15.7 - 10:23.8 & 2,1?,1 & b & 9433 & c9.1 & n17e12 & & 4 + 312 & 01/04/24 & 05:36.1 - 05:52.0 & 2,1,2 & b & 9433 & m2.1 & n18e01 & & 4 + 313 & 01/04/25 & 13:43.5 - 13:56.7 & 1,1,2 & b & 9433 & m2.7 & n18w09 & + & 4 + 314 & 01/04/26 & 13:10.7 - 13:18.2 & 1,1,1 & c & 9433 & m7.8 & n17w31 & + & 4 + 315 & 01/05/08 & 00:41.8 - 01:09.9 & 2,1,1 & b & 9445 & c9.9 & n23w43 & & 4 + 316 & 01/05/12 & 23:26.8 - 23:39.7 & 2,2,2 & b & 9455 & m3.0 & s17e00 & & 4 + 317 & 01/05/15 & 02:57.4 - 03:04.4 & 2,2,1 & b & 9455 & m1.0 & s17w29 & + & 4 + 318 & 01/06/19 & 23:19.8 - 23:29.5 & 2,1,1 & b & 9501 & c4.2 & s10w37 & + & 4 + 319 & 01/06/22 & 20:25.5 - 20:35.1 & 2,2,1 & b & 9511 & c5.5 & n09e28 & & 4 + 320 & 01/07/16 & 03:17.0 - 03:23.3 & 2,2?,1 & c & 9539 & m1.2 & s18w20 & + & 4 + 321 & 01/07/23 & 06:22.7 - 06:29.5 & 2,1,1 & b & 9545 & c5.0 & n10w65 & & 4 + 322 & 01/07/30 & 20:40.6 - 20:47.7 & 2,2,1 & b & 9562 & c6.0 & ( n05e79 ) & & 4 + 323 & 01/07/31 & 04:02.5 - 04:09.4 & 1,1,1 & b & 9562 & c6.0 & ( n05e76 ) & & 4 + 324 & 01/08/08 & 07:09.4 - 07:17.7 & 1,1,1 & b & 9557 & c3.9 & ( s18w86 ) & + & 4 + 325 & 01/08/09 & 18:27.7 - 18:37.4 & 2,2?,2 & b & 9570 & c7.8 & s17e19 & & 4 + 326 & 01/08/11 & 01:19.3 - 01:28.0 & 1,1,1 & b & 9563 & c5.2 & ( n20w83 ) & + & 4 + 327 & 01/08/25 & 16:25.8 - 16:34.6 & 1,1,2 & b & 9591 & x5.3 & s17e34 & + & 5 + 328 & 01/08/26 & 13:18.9 - 13:58.9 & 2,1,1 & c & ? & m1.3 & ( n16e89 ) & + & 4 + 329 & 01/08/31 & 10:38.1 - 10:47.8 & 2,2,1 & b & 9601 & m1.6 & n15e37 & + & 4 + 330 & 01/09/02 & 13:44.8 - 14:04.5 & 2,1,2 & b & 9591 & m3.0 & s20w53 & & 24,53 + 331 & 01/09/03 & 18:19.1 - 18:34.0 & 2,2,2 & b & 9608 & m2.5 & s22e96 & + & 4 + 332 & 01/09/07 & 15:28.9 - 15:33.8 & 1,1,1 & c & 9601 & m1.2 & n19w65 & + & 4 + 333 & 01/09/08 & 16:43.6 - 16:49.7 & 2,1,1 & b & 9608 & c5.1 & s23e32 & & 4 + 334 & 01/09/12 & 21:44.0 - 21:48.6 & 1,2,1 & c & 9606 & c9.6 & ( s17w63 ) & + & 4 + 335 & 01/09/13 & 19:50.7 - 19:54.2 & 1,2?,1 & b & 9606 & c5.8 & ( s18w79 ) & & 4 + 336 & 01/09/14 & 21:44.6 - 21:45.9 & 2,2,1 & b & 9616 & m3.7 & ( s14e39 ) & + & 4 + 337 & 01/09/17 & 08:21.0 - 08:29.8 & 2,2,2 & b & 9616 & m1.5 & s14e04 & + & 4 + 338 & 01/09/17 & 21:04.7 - 21:10.2 & 2,2,1 & c & 9616 & m1.0 & s11w06 & & 4 + 339 & 01/09/20 & 18:15.3 - 18:18.7 & 2,2,2 & b & 9631 & m1.5 & n09w11 & + & 4 + 340 & 01/09/21 & 04:52.8 - 05:01.5 & 2,2,1 & b & 9620 & c4.1 & n10e12 & & 4 + 341 & 01/09/22 & 18:09.9 - 18:15.6 & 2,2?,1 & c & 9633 & c5.4 & ( n14e88 ) & & 4 + 342 & 01/09/24 & 09:50.1 - 10:13.7 & 2,1,1 , & c & 9632 & x2.6 & s18e27 & + & 17 + 343 & 01/09/30 & 11:34.1 - 11:41.0 & 1,1,1 & b & 9628 & m1.0 & s20w75 & & 28 + 344 & 01/10/01 & 04:45.8 - 04:50.8 & 1,1,2 & b & & c4.7 & ( n19w89 ) & & 28 + 345 & 01/10/01 & 04:57.9 - 05:17.9 & 2,2,2 & b & 9628 & m9.1 & ( s21w84 ) & + & 28 + 346 & 01/10/02 & 17:11.5 - 17:50.5 & 2,1,2 & a & 9628 & c4.7 & ( s21w85 ) & & 4 + 347 & 01/10/03 & 06:42.5 - 06:47.4 & 1,2,2 & b & 9636 & c6.1 & n19w46 & + & 4 + 348 & 01/10/09 & 07:36.8 - 07:46.7 & 2,2,2 & b & 9645 & c7.0 & ( s19w86 ) & + & 4 + 349 & 01/10/19 & 16:25.2 - 16:35.9 & 2,2,2 & b & 9661 & x1.6 & n15w29 & + & 4 + 350 & 01/10/20 & 21:11.2 - 21:17.6 & 2,2,1 & b & 9674 & c4.6 & s09e24 & & 4 + 351 & 01/10/22 & 00:38.0 - 00:43.4 & 1,1,1 & c & 9658 & m1.0 & s16w86 & + & 4 + 352 & 01/10/22 & 14:40.6 - 14:58.5 & 2,2,2 & b & 9672 & m6.7 & s21e18 & + & 4 + 353 & 01/10/22 & 17:52.1 - 18:01.8 & 2,2,2 & c & 9672 & x1.2 & s18e16 & + & 4 + 354 & 01/10/25 & 22:52.6 - 23:00.9 & 1,2,2 & b & 9678 & c6.1 & ( n08e24 ) & & 4 + 355 & 01/10/29 & 08:14.6 - 08:21.7 & 1,1,1 & c & 9672 & m1.0 & s18w82 & + & 4 + 356 & 01/11/01 & 14:03.6 - 14:29.6 & 2,2?,2 & b & 9687 & m1.7 & s19e77 & + & 2 + 357 & 01/11/04 & 16:05.3 - 16:19.8 & 2,2,2 & b & 9684 & x1.0 & n06w18 & + & 17 + 358 & 01/11/06 & 03:01.3 - 03:03.6 & 2,1,1 & b & 9687 & m2.0 & s19e10 & & 4 + 359 & 01/11/08 & 15:12.9 - 15:45.1 & 2,1?,1 & b & 9690 & m4.2 & s17e36 & & 4 + 360 & 01/11/09 & 18:32.3 - 18:42.1 & 1,1,1 & c & 9687 & m1.9 & s21w42 & & 4 + 361 & 01/11/17 & 05:17.3 - 05:49.3 & 2,2,1 & b & 9704 & m2.8 & s13e42 & + & 4 + 362 & 01/11/28 & 15:41.4 - 15:44.4 & 1,2,1 & c & 9715 & c7.7 & ( n05e17 ) & + & 4 + 363 & 01/11/28 & 15:51.9 - 16:03.1 & 1,1,1 & c & 9715 & c2.1 & ( n05e17 ) & + & 4 + 364 & 01/11/28 & 16:14.9 - 16:22.3 & 1,1,1 & c & 9715 & m6.9 & n04e16 & + & 4 + 365 & 01/11/29 & 01:45.7 - 01:49.0 & 2,2,1 & c & 9715 & m1.1 & n04e12 & & 4 + 366 & 01/11/30 & 01:03.7 - 01:08.8 & 2,1,2 & b & 9718 & m3.5 & s06e57 & & 4 + 367 & 01/12/02 & 21:38.0 - 22:25.5 & 2,2,2 & b & 9714 & m2.0 & ( s09w88 ) & + & 4 + 368 & 01/12/10 & 22:46.1 - 22:55.9 & 2,1,2 & b & 9733 & c7.0 & n10e52 & & 4 + llrr 1,1,1 & collimated , confined , single & 74 & 14 + 1,1,2 & collimated , confined , recurrent & 10 & 5 + 1,2,1 & collimated , eruptive , single & 24 & 6 + 1,2,2 & collimated , eruptive , recurrent & 6 & 5 + 2,1,1 & loop - like , confined , single & 72 & 7 + 2,1,2 & loop - like , confined , recurrent & 15 & 7 + 2,2,1 & loop - like , eruptive , single & 94 & 30 + 2,2,2 & loop - like , eruptive , recurrent & 64 & 52 + & total & 359 & 126 + lcccccc + 1 ( collimated ) & 67/114 & * 24/42 * & c8.7 & * c6.1 * & 70 & * 51 * + 2 ( loop - like ) & 168/245 & * 102/136 * & m1.6 & * m1.8 * & 120 & * 110 * + 2 to 1 ratio & & & 1.8 & * 3.0 * & 1.7 & * 2.2 * + + 1 ( confined ) & 97/171 & * 32/56 * & c9.3 & * c6.1 * & 75 & * 45 * + 2 ( eruptive ) & 138/188 & * 80/98 * & m1.8 & * m2.3 * & 125 & * 120 * + 2 to 1 ratio & & & 1.9 & * 3.8 * & 1.7 & * 2.7 * + + 1 ( single ) & 165/264 & * 43/57 * & m1.1 & * m1.4 * & 90 & * 75 * + 2 ( recurrent ) & 70/95 & * 53/69 * & m1.9 & * m2.1 * & 135 & * 155 * + 2 to 1 ratio & & & 1.7 & * 1.5 * & 1.5 & * 2.1 * + + ( 1,1,1 ) & 42/74 & * 8/14 * & c7.9 & * c5.2 * & 45 & * 42 * + ( 2,2,2 ) & 50/64 & * 38/44 * & m2.1 & * m3.1 * & 160 & * 155 * + ( 2,2,2 ) to & & & 2.7 & * 6.0 * & 3.6 & * 3.7 * + ( 1,1,1 ) ratio & & & & & & + lcccccc + 1 ( collimated ) & 54/90 & * 13/26 * & 89@xmath13 & * 61@xmath13 * & 522 & * 444 * + 2 ( loop - like ) & 128/185 & * 49/67 * & 113@xmath13 & * 125@xmath13 * & 649 & * 547 * + 2 to 1 ratio & & & 1.3 & * 2.0 * & 1.2 & * 1.2 * + + 1 ( confined ) & 75/142 & * 17/39 * & 90@xmath13 & * 83@xmath13 * & 522 & * 450 * + 2 ( eruptive ) & 107/133 & * 38/43 * & 125@xmath13 & * 126@xmath13 * & 629 & * 642 * + 2 to 1 ratio & & & 1.4 & * 1.5 * & 1.2 & * 1.4 * + + 1 ( single ) & 125/200 & * 18/27 * & 93@xmath13 & * 104@xmath13 * & 528 & * 518 * + 2 ( recurrent ) & 57/75 & * 24/30 * & 155@xmath13 & * 126@xmath13 * & 662 & * 602 * + 2 to 1 ratio & & & 1.7 & * 1.2 * & 1.3 & * 1.2 * + + ( 1,1,1 ) & 33/63 & * 7/12 * & 83@xmath13 & * 61@xmath13 * & 500 & * 357 * + ( 2,2,2 ) & 42/52 & * 20/23 * & 168@xmath13 & * 126@xmath13 * & 718 & * 613 * + ( 2,2,2 ) to ( 1,1,1 ) & & & 2.0 & * 2.1 * & 1.4 & * 1.7 * + ratio & & & & & & + | a catalogue of x - ray plasma ejections ( xpes ) observed by the soft x - ray telescope onboard the yohkoh satellite has been recently developed in the astronomical institute of the university of wrocaw . the catalogue contains records of 368 events observed in years 1991 - 2001 including movies and crossreferences to associated events like flares and coronal mass ejections ( cmes ) .
163 xpes from 368 in the catalogue were not reported until now .
a new classification scheme of xpes is proposed in which morphology , kinematics , and recurrence are considered .
the relation between individual subclasses of xpes and the associated events was investigated .
the results confirm that xpes are strongly inhomogeneous , responding to different processes that occur in the solar corona .
a subclass of erupting loop - like xpes is a promising candidate to be a high - temperature precursor of cmes . |
the filamentation instability occurring when two plasmas shells collide has attracted considerable interest in recent years @xcite . in inertial confinement fusion ( icf ) ,
the so - called fast ignition scenario @xcite implies the unstable transport of a relativistic electron beam passing through the fusion plasma . in astrophysics , it is thought that collisionless shocks resulting from the encounter of two collisionless plasma shells could be triggered by the same instability @xcite .
the interest of these collisionless shocks is that they are able to accelerate particles up to high energies @xcite . because they also host magnetic turbulence ,
energetic particles which have not yet escaped the shock environment radiate .
these two features , acceleration and radiation in shock , could explain both high - energy - cosmic - rays ( hecr ) and gamma - ray - bursts ( grb ) @xcite . in collisionless conditions when the mean free path is much larger than the system size , colliding plasma shells should pass through each other without anything happening .
the reason why something happens instead of nothing is that as the shells interpenetrate , the overlapping region forms a counter - streaming system notoriously unstable @xcite .
when analysing the full unstable spectrum in the relativistic regime , one finds the filamentation instability tends to govern the unstable spectrum for shells of similar densities @xcite .
the instability , frequently referred to as weibel in literature , is found with a wave vector normal to the flow .
physically , it simply stems from the fact that whenever small perturbations occur that trigger the apparition of flow - aligned micro - currents in the shells , the resulting opposite currents repel each other , amplifying the perturbation .
while the occurrence of the instability is deleterious in icf @xcite , its presence in astrophysical settings guarantees the formation of collisionless shocks @xcite , hence grb s and hecr s .
great attention should then be drawn to the factors capable of quenching it . in this respect , it has been long for now that a flow - aligned magnetic field can completely suppress the instability @xcite .
but magnetic field could well be oblique with respect to the flow @xcite .
for such configurations , it was proved recently in 2d that in the cold regime , the filamentation instability can be canceled if and only if the field is perfectly aligned with the flow @xcite . in other words ,
the instability can not be suppressed by an oblique field , no matter how high .
the goal of the present paper is to extend this result to the full 3d case .
due to the complexity of the full kinetic theory , we address the cold ( i.e. mono - kinetic ) case . our setup is sketched in fig . [ fig:1 ] .
the two homogenous and infinite shells have the initial velocities @xmath0 for the rightward one , and @xmath1 for the leftward .
their initial densities are @xmath2 respectively .
the magnetic field lies within the plan @xmath3 , and the angle @xmath4 is such that @xmath5 .
we study perturbations with a wave vector within the plan @xmath6 , and the angle @xmath7 is defined by @xmath8 .
when the field is flow - aligned , i.e. @xmath9 , the system is symmetric around the @xmath10 axis and one can freely align the wave vector with the @xmath11 axis .
the growth - rate has no @xmath7 dependance and the system is 2d like @xcite . here
, the oblique magnetic field breaks the symmetry and the 3d problem demands accounting for the two finite perpendicular @xmath12 components .
both plasma shells are composed of electrons and protons , and we neglect the proton motion due to their inertia . at any rate , our goal here , besides evaluating the growth - rate , is to determine whether of not filamentation grows , and the corresponding @xmath7 range .
the filamentation of the electrons will thus be a sign that indeed , the system is unstable and that a shock may form .
since we work in the cold regime , we can write the conservation equation and the relativistic momentum equation for both species as , @xmath13 , \nonumber\end{aligned}\ ] ] with @xmath14 and with @xmath15 , @xmath16 being the electron mass . because both shells are charge and current neutral by their own , the whole system is charge and current neutral in any reference frame .
hence , we choose the reference frame where the initial velocities match the relation , @xmath17 that is , the reference frame where the center of mass is at rest .
this allows to switch continuously from a symmetric collision , where plasmas come from each side at the same velocities ( @xmath18 ) , to the case where a diluted beam enters a plasma nearly at rest ( @xmath19 ) .
in addition , working in the center of mass frame allows direct comparison with computer simulations , usually performed in this very frame .
though lengthy , the derivation of the dispersion equation is quite standard .
equations ( [ eqs ] ) are linearized assuming all quantities vary from their equilibrium value by small perturbations @xmath20 .
when coupled with maxwell s equations , these linearized equations allow to derive the dielectric tensor which determinant is the dispersion equation .
calculations are conducted analytically using a _
notebook described elsewhere @xcite in terms of the variables , @xmath21 where @xmath22 is the electronic cyclotron frequency and @xmath23 the electronic plasma frequency of the leftward shell .
for clarity , we will treat the symmetric and diluted cases separately .
a word of caution is needed regarding the validity of the growth - rate we are about to derive .
because the @xmath24 field is not flow - aligned , it acts on the flow even at zeroth order .
the present calculation are therefore meaningful provided the instability governs the early dynamic of the system , instead of the field .
this implies a growth - rate @xmath25 ( in @xmath26 units ) larger than the cyclotron frequency due to the transverse component of the field , i.e. , @xmath27 where @xmath28 refers to the lorentz factor of the two species . for the symmetric case ,
both lorentz factors are equal . in the diluted case where @xmath19 , eq . ( [ eq : ratio ] ) ensures @xmath29 , yielding a more stringent condition ( [ eq : valid ] ) .
when is the inequality satisfied ? in the ultra - relativistic regime and for the symmetric case , for example , the system is like field - free .
thus , the growth - rate approaches its field - free value @xmath30 @xcite , ensuring the fulfillment of the inequality above . beyond the nearly trivial result , the inequality is not easily solved .
we ll thus indicate on the forthcoming plots the limit it sets .
indeed , our main goal here is to find out whether or not filamentation can be quenched , and which values of @xmath7 are the most unstable , as long as the instability governs the early system dynamic .
units , as a function of @xmath31 for 3 field obliquities and @xmath32 .
the horizontal plane indicates the limit sets by eq .
( [ eq : valid ] ) . ]
labeling @xmath25 the growth - rate in @xmath26 units , the symmetries @xmath33 and @xmath34 allows to restrict the study to the quadrant @xmath35 .
it is known that for the 2d - like case , i.e. @xmath36 , the instability vanishes for @xmath37 @xcite . setting @xmath32 gives a critical field amplitude @xmath38 .
we thus chose @xmath39 and vary the field obliquity @xmath4 to observe ( first , numerically ) how the instability cancelation evolves for @xmath40 . for @xmath9 , one can check on figure [ fig : sym ] how the growth - rate vanishes for @xmath41 . with such field orientation ,
the system is symmetric around the @xmath10 axis , and its 2d nature is recovered .
then , for @xmath4 as small as @xmath42 , the growth - rate still vanishes for small @xmath43 s but is clearly finite for other orientations .
it can be noticed that although the lower part of the plane ( @xmath44 ) yields a growth - rate violating eq .
( [ eq : valid ] ) , the large @xmath45 part of the spectrum fulfills it and remains therefore unstable . for @xmath46 and @xmath32 .
the angle @xmath7 is from bottom to top : 0 , @xmath47 and @xmath48.,scaledwidth=70.0% ] as is the case for the 2d problem , we observe the growth - rate saturates for @xmath49 . this can be checked in 3d on figure [ fig : symz ] where we focus on @xmath46 and plot the growth - rate in terms of @xmath50 for various orientations of the wave vector .
this feature allows to write the dispersion equation in terms of @xmath51 and derive the asymptotic dispersion equation for large @xmath45 from the coefficient of the higher degree in @xmath45 .
the resulting dispersion equation at @xmath52 reads , @xmath53 -\beta ^2 \omega_b^2 \sin\theta_b^2 \cos(2\theta_k ) -
\beta ^2 \omega_b^2 \sin\theta_b^2 = 0.\ ] ] it can be solved exactly with , @xmath54 because this quantity is the high-@xmath45 growth - rate , it is also the maximum growth - rate in any given direction @xmath7 . therefore , canceling @xmath55 implies canceling the instability . setting @xmath56 now yields , @xmath57 and taking the square of both side gives @xmath58
note that although necessary , this condition may not be sufficient .
indeed , inserting ( [ eq : cancelsym ] ) into eq .
( [ eq : delasymzinf ] ) gives , @xmath59 let us now examine each possibilities offered by eq .
( [ eq : cancelsym ] ) : * eq . ( [ eq : cancelsym ] )
can be fulfilled through @xmath60 . with then @xmath61 , eq .
( [ eq : cancelsym1 ] ) gives @xmath56 .
this is the trivial case where there is no drift between the flows . *
( [ eq : cancelsym ] ) can be fulfilled through @xmath62 . with now @xmath63 , eq .
( [ eq : cancelsym1 ] ) gives @xmath64 .
this is the maximum growth - rate for the field - free case @xcite .
( [ eq : cancelsym ] ) can be fulfilled through @xmath65 , that is , @xmath66 $ ] . here
if @xmath61 , that is , @xmath37 .
this is the condition for the cancellation of the instability in the flow - aligned field case @xcite .
such cancellation is valid for all @xmath7 , so that the entire 3d @xmath12 spectrum is stabilized . * finally , eq . ( [ eq : cancelsym ] )
can also be fulfilled through @xmath67 , that is , @xmath68 $ ] .
the growth - rate vanishes now for @xmath61 , namely @xmath69 in this case , we find the instability can be canceled only for discrete orientations of the wave vector , namely , @xmath70 .
the system remains therefore unstable for every other directions .
the proof has thus been made that it is impossible to stabilize the entire @xmath12 spectrum unless the field is along the flow , with @xmath66 $ ] . for an oblique field
, it is quite straightforward that the growth - rate ( [ eq : delasymzinf ] ) is a decreasing function of @xmath7 over @xmath71 $ ] .
therefore , even when @xmath72 , we always have @xmath73 .
this means that filamentation modes with a wave vector pertaining to the plan @xmath74 grow faster than the others .
the orientation of the filament generated is therefore related to the orientation of the field .
also , the 2d regime ( with necessarily @xmath36 ) usually studied in theory or simulations pertains indeed to the fastest growing configuration .
although the large @xmath75 limit necessarily pertains to a regime governed by the field instead of the instability , the growth - rate in this limit is worth investigating in order to bridge with previous results @xcite .
the expansion of eq .
( [ eq : delasymzinf ] ) is straightforward and gives , @xmath76 this result is simply the product of the field - free growth - rate for our present symmetric system , @xmath77 , times a geometric factor .
and the geometric factor is just the 2d result ( @xmath36 ) , times a @xmath78 factor accounting for the orientation of @xmath12 . for @xmath9 ,
we find @xmath79 , which means a flow - aligned field always ends - up canceling the instability . here again , we recover the fact that the growth - rate can be canceled only for @xmath80 $ ] or @xmath81 $ ] . units , as a function of @xmath31 for 3 field obliquities , @xmath82 and @xmath32 .
the horizontal plane indicates the limit sets by eq .
( [ eq : valid ] ) . ]
we now treat the @xmath19 , which amounts to a diluted beam passing to a plasma nearly at rest . in this respect , only the beam is relativistic now .
in the diluted beam regime , the threshold for complete cancelation of the filamentation instability when @xmath9 is @xcite , @xmath83 figure [ fig : dilu ] features the growth - rate in terms of @xmath45 for various field obliquities and @xmath84 close to the threshold above .
the patterns observed in the symmetric case repeat here , with in particular , a saturation at large @xmath45 and an uniform cancelation of the instability only achieved for @xmath9 . still , the dispersion equation for @xmath52 , which gives the maximum growth - rate , in not amenable to a simple 4th order polynomial like eq .
( [ eq : symzinf ] ) . instead , we find a 8th order polynomial without simple solution .
it is nevertheless possible to guess a good approximation of @xmath55 from previous works , and numerically check its relevance . in the 2d case @xcite ,
the large @xmath75 limit of the growth - rate in the diluted beam regime can be deduced from the symmetric case multiplying the result by @xmath85 .
also , the threshold for cancelation goes from @xmath86 in the symmetric case to @xmath87 in the diluted one . here also , multiplying the result by @xmath85 allows to switch between diluted and symmetric cases . denoting @xmath88 the symmetric result of eq .
( [ eq : delasymzinf ] ) , the proposed _ ansatz _ for the diluted counterpart @xmath89 is @xmath90 for the diluted case , exact value vs. _ ansatz _ ( [ eq : dilu_zinf ] ) .
parameters are @xmath91 and @xmath32 .
the last row displays the value of @xmath92 times a large amplification factor , where @xmath89 is the _ ansatz _ ( [ eq : dilu_zinf ] ) , and @xmath93 the dispersion equation . ]
figure [ fig : dilu_compa ] compares for various field obliquities @xmath4 the exact value of @xmath89 with the _
ansatz_. the agreement is very good all over the plane @xmath94 . with the dispersion equation reading @xmath93 , the last row displays the value of @xmath92 , with @xmath89 given by eq .
( [ eq : dilu_zinf ] ) .
we therefore find here a successful test of the accuracy of the _
ansatz_. here again @xmath95 is singled out as it is the only orientation for which the instability can be canceled . at any rate ,
the results obtained for the symmetric case are recovered . for an oblique field
, the instability is canceled only for @xmath95 .
furthermore , the maximum growth - rate is larger at @xmath36 , which again will determine the orientation of the generated filaments .
the 3d filamentation instability of two cold counter - streaming beams in the presence of an oblique magnetic field has been examined for every possible orientations of the wave vector @xmath12 .
our main result is that as is the case for previous 2d treatments @xcite , the instability can never be completely quenched for a non flow - aligned field .
although the formalism is only valid when filamentation does govern the initial beam dynamic , the result remains .
results are rigorous for the symmetric beams case , and rely on an _ ansatz _ for the diluted one .
this 3d study brings an element by definition out of reach of the 2d problem : we find a systematically larger growth - rate for @xmath36 .
this bears consequences on the orientation of the filaments with respect to the field and the flow .
also , 2d simulations of the system with @xmath96 can be performed without missing the most unstable mode .
the regime where the early dynamic imposed by the field can compete with the instability is out of reach of the present study and could definitely be the subject of forthcoming works . also , temperature effects should be investigated .
while the field tends to stabilize first the long wavelengths @xcite , temperature stabilizes the short ones .
it could be that with the help of kinetic effects , and oblique field can still cancel filamentation within some parameters range yet to determine .
the case where one of the two shells is composed of electrons / positrons instead of electrons / protons is worth investigating , in particular in connection with the problem of blazar - induced pair beams propagating in the intergalactic medium @xcite . at present time
, it seems difficult to qualitatively predict the behavior of such a system because electrons and positrons react differently to the magnetic field .
future works will have to clarify this issue . | the filamentation ( weibel ) instability plays a key role in the formation of collisionless shocks which are thought to produce gamma - ray - bursts and high - energy - cosmic - rays in astrophysical environments .
while it has been known for long that a flow - aligned magnetic field can completely quench the instability , it was recently proved in 2d that in the cold regime , such cancelation is possible if and only if the field is perfectly aligned . here , this result is finally extended to a 3d geometry .
calculations are conducted for symmetric and asymmetric counter - streaming relativistic plasma shells .
2d results are retrieved in 3d : the instability can never be completely canceled for an oblique magnetic field .
in addition , the maximum growth - rate is always larger for wave vectors lying in the plan defined by the flow and the oblique field . on the one hand , this bears consequences on the orientation of the generated filaments . on the other hand
, it certifies 2d simulations of the problem can be performed without missing the most unstable filamentation modes . |
the thermal control of a spacecraft ensures that the temperatures of its various parts are kept within their appropriate ranges @xcite .
the simulation and prediction of temperatures in a spacecraft during a mission are usually carried out by commercial software packages .
these software packages employ `` lumped parameter '' models that describe the spacecraft as a discrete network of nodes , with one energy - balance equation per node .
the equations for the thermal state evolution are coupled nonlinear first - order differential equations , which can be integrated numerically .
given the thermal parameters of the model and its initial thermal state , the numerical integration of the differential equations yields the solution of the problem , namely , the evolution of the node temperatures .
however , a detailed model with many nodes is difficult to handle , and its integration for a sufficiently long time of evolution can take considerable computer time and resources .
therefore , it is very useful to study simplified models and approximate methods of integrating the differential equations . many spacecraft missions , in particular , satellite missions , consist of an initial transient part and then a stationary part , in which the spacecraft just goes around a closed orbit , in which the heat inputs are periodic .
these periodic heat inputs are expected to induce periodic temperature variations , with a maximum and a minimum temperature in each orbit .
this suggests a conservative approach that consists in computing only the temperatures for the _ hot _ and _ cold cases _ of the given orbit , defining them as the two steady cases with the maximum and minimum heat loads , respectively . naturally , the real temperature variations in the orbit are smaller , because there is not enough time for the hot and cold cases to establish themselves .
in fact , the temperature variations can be considerably smaller , to such a degree that it is necessary to integrate the differential equations , at least approximately .
the differential equations for energy balance are nonlinear due to the presence of radiation couplings , which follow the stefan - boltzmann quartic law .
a common approach to these equations involves a linearization of the radiation terms that approximate them by heat conduction terms @xcite .
this approach transforms the nonlinear equations into standard linear heat conduction equations . but
this approach has not been sufficiently justified , is of a heuristic nature and does not constitute a systematic approximation .
in fact , nonlinear equations are very different from linear equations and , in particular , a periodic driving may not induce periodic solutions but much more complex solutions , namely , _ chaotic _ solutions .
therefore , we have carried out in preceding papers a full nonlinear analysis of spacecraft thermal models @xcite .
the conclusion of the analysis is that the complexities of nonlinear dynamics , such as multiple equilibria and chaos , do not appear in these models .
while the existence of only one equilibrium state can be proved in general , the absence of chaos under driving by variable external heat loads can only be proved for a limited range of magnitudes of the driving loads .
this range presumably includes the magnitudes involved in typical spacecraft orbits .
the proofs in refs . and are constructive and are based on a perturbation method that is expected to be sound when the _ linear _ equations corresponding to the first perturbative order constitute a good approximation of the nonlinear equations .
this implies that the fully nonlinear solution describes a weakly nonlinear oscillator .
since the perturbative approximation is mathematically rigorous and systematic , it is worthwhile to study in detail the scope of the perturbative linear equations and , furthermore , to compare them with previous linear approaches of a heuristic nature .
the main purpose of this paper is to study the linear method of predicting the thermal behavior of spacecraft in stationary orbits ( sect .
[ linear ] and [ fourier - anal ] ) and to test it on a minimally realistic thermal model of a satellite in a circular orbit . since the general one and two - node models analyzed in refs . and
, respectively , are too simple , we define in this paper a ten - node thermal model of a small moon - orbiting satellite ( sect .
[ ten - node ] ) .
this model is simple enough to allow us to explicitly show all the quantities involved ( thermal couplings and capacities , heat inputs , etc . ) and it is sufficient for illustrating the main features of the linear approach . as realistic thermal models have many more nodes , we consider in sect .
[ scale ] the important issue of scalability of the method and , hence , its practical applications .
computational aspects of the steady - state problem have been studied by krishnaprakas @xcite and by milman and petrick @xcite , while computational aspects of the direct integration of the nonlinear equations for the unsteady problem have been studied by krishnaprakas @xcite . here
we focus on the linear equations for the stationary but unsteady case and survey its computational aspects . a note on notation : in the equations that contain matrix or vector quantities , sometimes we use component notation ( with indices ) while other times we use compact matrix notation ( without indices ) , according to the nature of the equations .
a lumped - parameter thermal model of a continuous system consists of a discrete network of isothermal regions ( _ nodes _ ) that represent a partition of the total thermal capacitance and that are linked by thermal conduction and radiation couplings @xcite .
this discretization reduces the integro - differential heat - transfer equations to a set of energy - balance odes , one per node , which control the evolution of the nodes temperatures @xcite : @xmath66 - r_i\,({t_i}^4- t_{0}^4 ) , \quad i= 1,\ldots , n,\end{aligned}\ ] ] where @xmath30 is the number of nodes and @xmath34 contains the total heat input to the @xmath17th - node from external radiation and from internal energy dissipation ( if there is any ) .
the conduction and radiation coupling matrices are denoted by @xmath24 and @xmath67 , respectively ; they are symmetric ( @xmath68 and @xmath69 ) and @xmath70 ; so there are @xmath71 independent coupling coefficients altogether , but many vanish , usually .
the temperature @xmath72 k is the temperature of the environment , namely , the cosmic microwave background radiation . the @xmath17th - node coefficient of radiation to the environment is given by @xmath73 , where @xmath0 denotes the outward facing area , @xmath74 its ( infrared ) emissivity , and @xmath75 is the stefan - boltzmann constant . the constant term @xmath76 can be included in @xmath34 or ignored altogether , if each @xmath77 .
equations ( [ ode ] ) coincide with the ones implemented in commercial software packages , for example , esatan@xmath44 @xcite .
there is no systematic procedure for finding the analytical solution of a system of nonlinear differential equations , except in some particularly simple cases .
of course , nonlinear systems can always be integrated numerically with finite difference schemes .
methods of this kind are employed in commercial software packages .
when a nonlinear system can be approximated by a linear system and , hence , an approximate analytic solution can be found , this solution constitutes a valuable tool .
actually , one can always resort to some kind of perturbation method to linearize a nonlinear system .
therefore , we now study the rigorous linearization of eqs .
( [ ode ] ) based on a suitable perturbation method , and we also describe , for the sake of a comparison , a heuristic linearization , which actually is best understood in light of the results of the perturbation method .
if we assume that the heat inputs @xmath78 in the energy - balance eqs .
( [ ode ] ) are periodic , namely , that there is a time interval @xmath79 such that @xmath80 , then it seems sensible to study first the effect of the mean heat inputs in a period .
this _ averaging _ method , introduced in refs . and
, relies on the fact that the _ autonomous _ nonlinear system of odes for constant @xmath81 can be thoroughly analyzed with analytical and numerical methods .
for example , it is possible to determine that there is a unique steady thermal state and that it is ( locally ) stable @xcite .
the actual values of the steady temperatures can be found efficiently with various numerical methods @xcite .
furthermore , the eigenvalues and eigenvectors of the jacobian matrix of the nonlinear system of odes provides us with useful information about the dynamics , in particular , about the approach to steady - state : the eigenvectors represent independent thermal modes and the eigenvalues represent their relaxation times @xcite .
once the averaged equations are solved , the variation of the heat inputs can be considered as a driving of the averaged solutions .
thus , we can define the driving function @xmath82 where @xmath83 denotes the mean value of @xmath78 over the period of oscillation .
a weak driving function must not produce a notable deviation from the averaged dynamics .
in particular , the long - term thermal state of an orbiting spacecraft must oscillate about the corresponding steady - state .
to embody this idea , we introduce a formal perturbation parameter @xmath84 , to be set to the value of unity at the end , and write eqs .
( [ ode ] ) as @xmath85 - \frac{r_{i}}{c_i } \,{t_i}^4 , \quad i= 1,\ldots , n,\end{aligned}\ ] ] then , we assume an expansion of the form @xmath86 when we substitute this expansion into eqs . ( [ eode ] ) , we obtain for the zeroth order of @xmath84 @xmath87 - \frac{r_{i}}{c_i } \,t_{(0)i}^4\ , , \quad i= 1,\ldots , n,\ ] ] that is to say , the averaged equations .
the initial conditions for these equations are the same as for the unaveraged equations .
for the first order in @xmath84 , we obtain the following system of linear equations : @xmath88 here , @xmath89 is the jacobian matrix @xmath90 where @xmath91 is the solution of the zeroth order equation
. equations ( [ linode - pert ] ) are to be solved with the initial condition @xmath92 .
the elements of the jacobian matrix at a generic point in the temperature space are calculated to be : @xmath93.\end{aligned}\ ] ] this matrix has interesting properties .
first of all , it has negative diagonal and nonnegative off - diagonal elements . in other words
, @xmath94 is a @xmath95-matrix @xcite .
furthermore , it fulfills a semipositivity condition that qualifies it as a nonsingular @xmath96-matrix @xcite .
since the eigenvalues of an @xmath96-matrix have positive real parts , the opposite holds for @xmath20 , namely , its eigenvalues have negative real parts .
one more interesting property of @xmath94 , related to semipositivity , is that it possesses a form of _ diagonal dominance _
: it is similar to a diagonally dominant matrix and the similarity is given by a positive diagonal matrix .
naturally , this property is shared by @xmath20 .
these properties are useful to prove some desirable properties of the solutions of eqs .
( [ linode - pert ] ) .
the chief property of @xmath20 is that @xmath94 is a nonsingular @xmath96-matrix . in particular
, it implies that @xmath97 is non - negative and , therefore , that the perron - frobenius theory is applicable to it @xcite .
the relevant results to be applied are : ( i ) perron s theorem , which states that a _
strictly _ positive matrix has a unique real and positive eigenvalue with a positive eigenvector and that this eigenvalue has maximal modulus among all the eigenvalues ; ( ii ) a second theorem , stating that if a @xmath95-matrix that is a nonsingular @xmath96-matrix is also `` irreducible '' , then its inverse is strictly positive .
the irreducibility of @xmath20 follows from the symmetry of the matrices @xmath27 and @xmath39 @xcite .
as the positive ( perron ) eigenvector of @xmath97 is the eigenvector of @xmath20 that corresponds to its smallest magnitude eigenvalue , it defines the slowest relaxation mode ( for a given set of temperatures ) .
therefore , in the evolution of temperatures given by eqs .
( [ bode - pert ] ) , steady - state is eventually approached from the zone corresponding to simultaneous temperature increments ( or decrements ) .
the matrix @xmath98 in eqs .
( [ linode - pert ] ) is obtained by substituting @xmath99 for @xmath100 in eqs .
( [ jij ] ) and ( [ jii ] ) .
then , the nonhomogeneous linear system with variable coefficients , eqs .
( [ linode - pert ] ) , can be solved by variation of parameters @xcite , yielding the expression : @xmath101 where @xmath102 is a matrix formed by columns that are linearly independent solutions of the corresponding homogeneous equation , with the condition that @xmath103 ( the identity matrix ) .
the difficulty in applying this formula lies in computing @xmath102 , that is , in computing the solutions of the homogeneous equation .
moreover , this computation demands the previous computation of the solution for @xmath91 . since we are only interested in the stationary solutions of the heat - balance equations rather than in transient thermal states , it is possible to find an expression of these solutions that is more manageable than eq .
( [ t_1 ] ) .
the transient thermal state relaxes exponentially to the stationary solution , which is a _
limit cycle _ of the nonlinear equations , technically speaking @xcite .
therefore , the stationary solution is given by the solution of eqs .
( [ linode - pert ] ) with the _ constant _ jacobian matrix calculated at the steady - state temperatures , which we name @xmath104 . ) in which @xmath105 . ]
this solution is simply @xcite : @xmath106 \cdot f(t-\t ) \,d\t , \label{t_1_jcst}\ ] ] with @xmath20 calculated at the point @xmath45 .
furthermore , the periodic stationary solution is obtained by extending the upper integration limit from @xmath41 to infinity : @xmath107 \cdot f(t-\t ) \,d\t . \label{t_1_lim}\ ] ] this function is indeed periodic , unlike the one defined by eq .
( [ t_1_jcst ] ) , so it is determined by its values for @xmath108 $ ] .
note that @xmath109 . for numerical computations
, it can be convenient to express the integral from 0 to @xmath110 as an integral from 0 to @xmath79 , taking advantage of the periodicity as follows : @xmath111 \cdot f(t-\t)\,d\t = \sum_{n=0}^\infty \int_{n\t}^{(n+1)\t } \exp\left[\t j \right ] \cdot f(t-\t)\,d\t = \\ & & \sum_{n=0}^\infty \exp(n\t j ) \int_0^{\t } \exp\left[\t j \right ] \cdot
f(t-\t)\,d\t = [ i-\exp(\t j)]^{-1 } \int_0^{\t } \exp\left[\t j \right ] \cdot f(t-\t)\,d\t\end{aligned}\ ] ] ( the series converges because the eigenvalues of @xmath20 have negative real parts ) . in the last integral ,
the argument of @xmath112 can be transferred to the interval @xmath113 $ ] : @xmath114 \cdot f(t-\t)\,d\t = \int_0^{t } \exp\left[\t j \right ] \cdot f(t-\t)\,d\t + \int_t^{\t } \exp\left[\t j \right ] \cdot f(t-\t+\t)\,d\t,\ ] ] where @xmath108 $ ] .
note that the one - period shift in the argument of the last @xmath112 is necessary for the argument to be in @xmath113 $ ] .
some remarks are in order .
first of all , we have assumed that there is one asymptotic periodic solution of the nonlinear eqs .
( [ eode ] ) and only one ( a unique limit cycle ) .
equivalently , we have assumed that the perturbation series converges .
this assumption holds in an interval of the amplitude of heat input - variations @xmath112 @xcite .
besides , for the integrals in eq .
( [ t_1_lim ] ) and the following equations to make sense , it is required that @xmath115 \ra 0 $ ] as @xmath116 .
this is guaranteed , because the eigenvalues of @xmath20 have negative real parts , as is necessary for the steady - state to be stable .
in fact , the eigenvalues are expected to be negative real numbers and @xmath20 is expected to be diagonalizable but both properties are not rigorously proven @xcite ( however , see sect . [ heur ] ) .
if @xmath20 is diagonalizable , that is to say , there is a real matrix @xmath31 such that @xmath117 is diagonal , then the calculation of the integrals is best carried out on the eigenvector basis , given by the matrix @xmath31 . using this basis , eq . ( [ t_1_lim ] )
is expressed as @xmath118_i(t ) = \sum_{a=1}^n p_{ia } \int_0^\infty \exp\left[\t \l_a \right ] \sum_{j=1}^n p^{-1}_{aj } f_j(t-\t)\,d\t , \quad i= 1,\ldots , n , \label{t_1_eigenvec}\ ] ] where the first sum runs over the eigenvectors and their corresponding eigenvalues @xmath119 .
expression ( [ t_1_eigenvec ] ) allows us to compare the contribution of the different thermal modes . in particular , for the fast modes , such that @xmath120 is large
, we can use watson s lemma @xcite to derive the asymptotic expansion : @xmath121 f_a(t-\t)\,d\t = \frac{f_a(t)}{-\l_a } - \frac{\dot{f}_a(t)}{\l_a^2 } + \mfunction{o}\left(\frac{1}{\l_a^3 } \right),\ ] ] where @xmath122 .
when @xmath120 is large , the first term suffices ( unless @xmath14 is also large , for some reason ) ; and the first term is small , unless @xmath12 is large .
in essence , if the fast modes are not driven strongly , they can be neglected in the sum over @xmath13 in eq .
( [ t_1_eigenvec ] ) . for second order in @xmath84 , a straightforward calculation @xcite yields
the following linear equation : @xmath123 where @xmath98 is the same jacobian matrix that appears in the first - order eq .
( [ linode - pert ] ) and @xmath124 the initial condition for eq .
( [ linode - pert2 ] ) is @xmath125 , as for eqs .
( [ linode - pert ] ) .
therefore , the first - order and second - order equations have identical solutions in terms of their respective driving terms , although @xmath126 , eqs .
( [ tf ] ) , is a known function of @xmath41 only when the lower order equations have been solved . the integral expression , eqs . ( [ t_1_lim ] ) , of the stationary solution
@xmath127 is also valid for @xmath128 , after replacing @xmath112 with @xmath126 and using in eq .
( [ tf ] ) the stationary values @xmath129 and @xmath130 ( which make @xmath126 periodic ) .
it is possible to carry on the perturbation method to higher orders , and it always amounts to solving the same linear equation with increasingly complicated driving terms that involve the solutions of the lower order equations .
the example of sect .
[ ten - node ] shows that , in a typical case , @xmath128 is a small correction to @xmath127 , and further corrections are not necessary .
this confirms that the perturbation method is reliable for a realistic case .
a linearization procedure frequently used in problems of radiation heat transfer @xcite consists of using the algebraic identity @xmath131 to define an effective conductance for the radiation coupling between nodes @xmath17 and @xmath132 .
the equation @xmath133 defines the effective conductance @xmath134 for specified values of the node temperatures @xmath46 and @xmath100 . for an orbiting spacecraft ,
the natural base values of the node temperatures are the ones that correspond to the steady - state solution of the averaged equations , namely , @xmath135 in the special case of radiation to the environment , @xmath136 can be replaced with linear terms @xmath137 such that @xmath138 for @xmath139 the resulting linear equations are : @xmath140 these equations have only conduction couplings , so they are a discretization of the partial differential equations of heat conduction . as a linear system of odes ,
the standard form is @xmath141 where @xmath20 ( the jacobian matrix ) is now given by : @xmath142.\end{aligned}\ ] ] the linear system of eqs .
( [ stdlinode ] ) can be solved in the standard way , yielding : @xmath143 \left({t}(0 ) + \int_0^t \exp\left[-\t j \right ] \cdot q(\t)\,d\t \right),\ ] ] where we have introduced the vector @xmath144 , with components @xmath145 .
we can also express the solution in terms of the driving function @xmath146 : @xmath147 { t}(0 ) + \int_0^t \exp\left[(t-\t ) j \right ] \cdot \left(f(\t ) + \langle q \rangle \right ) d\t = \\
\exp\left[t j \right ] { t}(0 ) + \int_0^t \exp\left[(t-\t ) j \right ] \cdot f(\t ) d\t + j^{-1}(\exp\left[t j \right ] -i ) \langle q \rangle\end{aligned}\ ] ] for large @xmath41 , this solution tends to the periodic stationary solution @xmath148
\cdot f(t-\t ) d\t - j^{-1 } \langle q \rangle , \label{t_lim}\ ] ] assuming that @xmath149 \ra 0 $ ] as @xmath150 .
this is a consequence of the structure of @xmath20 , as in the preceding section . in the present case ,
the eigenvalues of @xmath20 , beyond having negative real parts , are actually negative real numbers , as we show below .
the total conductance matrix @xmath151 is symmetric but this does not imply that @xmath20 is symmetric .
nevertheless , if we define @xmath152 , the matrix @xmath153 is symmetric , because its off - diagonal matrix elements are : @xmath154 hence , the matrix @xmath155 , similar to @xmath20 , has real eigenvalues .
furthermore , @xmath155 is diagonalized by an orthogonal transformation ; that is to say , there is an orthogonal matrix @xmath156 such that @xmath157 is diagonal .
therefore , the thermal modes are actually _ normal _ ; that is to say , the modes , which are the eigenvectors of @xmath20 and hence the columns of the matrix @xmath158 , are related to the eigenvectors of @xmath159 , which are normal and are given by the columns of @xmath156 .
alternatively , one can say that the eigenvectors of @xmath20 are normal in the `` metric '' defined by @xmath2 ; namely , @xmath160 which can be written in matrix form as @xmath161 .
naturally , the orthogonality of modes greatly simplifies some computations .
furthermore , the symmetry of the conductance matrix implies that the sum in eq .
( [ linode ] ) can be written as the action of a _
graph laplacian _
@xcite on the temperature vector . naturally , the graph is formed by the nodes and the linking conductances .
a graph laplacian is a discretization of the ordinary laplacian and is conventionally defined with the sign that makes it positive semidefinite .
the zero eigenvalue corresponds to a constant function , that is , a constant temperature , in the present case .
a vector with equal components , say , equal to @xmath162 , is the positive ( perron ) eigenvector of the matrix . with more generality , the laplacian of a graph can be defined as a symmetric matrix with off - diagonal entries that are negative if the nodes are connected and null if they are not @xcite .
this definition does not constrain the diagonal entries and , therefore , does not imply that a graph laplacian is positive semidefinite .
it can be made positive definite ( or just semidefinite ) by adding to it a multiple of the identity matrix , which does not alter the eigenvectors .
of course , the eigenvector corresponding to the smallest eigenvalue does not have to be constant , but the perron - frobenius theorem @xcite tells us that it is positive . by this general definition of a graph laplacian ,
the matrix @xmath163 is a different laplacian for the same graph , and eqs .
( [ stdlinode ] ) contain the action of this laplacian on the vector @xmath164 .
notice that this general definition of a graph laplacian is connected with the definition of a @xmath95-matrix @xcite and , actually , a symmetric @xmath95-matrix is a graph laplacian . if such a matrix is positive definite , then it is equivalent to a _ stieltjes matrix _ , namely , a symmetric nonsingular @xmath96-matrix @xcite .
the general jacobian obtained in sect .
[ pert - eq ] is also such that @xmath94 and also @xmath165 are both nonsingular @xmath96-matrices , but they need not be symmetric . to investigate the accuracy of the approximation of the radiation terms by conduction terms , let us compare the periodic solution given by eq .
( [ t_lim ] ) with the first - order perturbative solution found in sect .
[ pert - eq ] , namely , @xmath166 .
of course , the jacobian matrices in the respective integrals differ , as do the temperature vectors added to the integrals , namely , @xmath45 or @xmath167 . while @xmath45 corresponds to the authentic steady - state of the nonlinear averaged equations , @xmath168 corresponds to the steady - state of eqs .
( [ linode ] ) after averaging , which is a state without significance , since we have already used the set of temperatures @xmath45 of the authentic steady - state to define the radiation conductances @xmath169 in eq .
( [ linode ] ) . therefore , the only sensible linear solution is the perturbative solution @xmath166 , even if we replace the jacobian matrix given by eq .
( [ jij ] ) and ( [ jii ] ) with the one given by eq .
( [ njij ] ) and ( [ njii ] ) . in our context
, the notion of radiation conductance actually follows from the symmetry of the matrices @xmath170 or @xmath155 .
therefore , the most natural definition of radiation conductance probably is @xmath171 , that is , the symmetrization of the term @xmath172 in eq .
( [ jij ] ) .
this symmetrization has been tested by krishnaprakas @xcite , considering the steady - state problem for models with up to @xmath173 nodes and working with various resolution algorithms .
he found that the effect of symmetrization is not appreciable . to estimate the effect of the antisymmetric part of the matrix @xmath174 , namely , @xmath175 , on the eigenvalue problem for the jacobian
, we proceed as follows .
we formulate this eigenvalue problem in terms of the matrix @xmath155 , so that it is an eigenvalue problem for a symmetric matrix perturbed by a small antisymmetric part .
this problem is well conditioned , because the eigenvectors of the symmetric matrix ( the columns of the matrix @xmath156 ) are orthogonal . in particular , the perturbed eigenvalues are still real .
furthermore , the first - order perturbation formula for the eigenvalue @xmath119 associated with an eigenvector @xmath4 @xcite yields : @xmath176 vanishing because the perturbation matrix @xmath177 is antisymmetric .
so the nonvanishing perturbative corrections begin at the second order in the perturbation matrix , and , in this sense , they are especially small .
given that @xmath178 is a periodic function , it can be expanded in a fourier series . to derive this series ,
let us first introduce the fourier series of @xmath7 , @xmath179 inserting this series in the integral of eq .
( [ t_1_lim ] ) and integrating term by term , we obtain the fourier series for @xmath178 .
alternatively , we can substitute the fourier series for both @xmath178 and @xmath7 into eqs .
( [ linode - pert ] ) , where @xmath20 is taken to be constant ; then we can solve for the fourier coefficients of @xmath178 .
the result is @xmath180 the fourier coefficients @xmath8 are obtained by integration : @xmath181 given that @xmath7 is a real function , @xmath182 furthermore , @xmath183 implies @xmath184 so @xmath7 is defined by the sequence of fourier coefficients for positive @xmath9 .
this sequence must fulfill the requirement that @xmath185 so a limited number of the initial coefficients may suffice . actually , for numerical work
, eq . ( [ fourier - coef ] ) can be conveniently replaced by the discrete fourier transform @xmath186 which only requires sampling of the values for @xmath7 , but also only defines a finite number of independent fourier coefficients , because @xmath187 .
notice that we usually have available just a sampling of the heat inputs at regular time intervals , rather than the analytical form of @xmath34 . to calculate the exact number of independent fourier coefficients provided by eq .
( [ d_fourier - coef ] ) , we must take into account eqs .
( [ f * ] ) and ( [ f0 ] ) .
if @xmath29 is an odd number , the independent fourier coefficients @xmath8 are the ones with @xmath188 ; that is to say , there are @xmath189 independent real numbers .
if @xmath29 is even , the independent fourier coefficients are the ones with @xmath190 , and @xmath191 is real , so there are @xmath189 independent real numbers as well . for definiteness ,
let @xmath29 be odd .
then , we can express @xmath7 as @xmath192.\ ] ] of course , the values of @xmath7 at @xmath193 are the sampled values employed in eq .
( [ d_fourier - coef ] ) , but the expression is valid for any @xmath108 $ ] and constitutes an interpolation of the sampled values .
naturally , the higher the sampling frequency @xmath29 , the more independent fourier coefficients we have and the more accurate the representation of @xmath7 is . as is well known , the fourier series of a function @xmath7 that is piecewise smooth converges to the function , except at its points of discontinuity , where it converges to the arithmetic mean of the two one - sided limits @xcite .
however , the convergence is not _ uniform _ , so that partial sums oscillate about the true value of the function near each point of discontinuity and `` overshoot '' the two one - sided limits in opposite directions .
this overshooting is known as the gibbs phenomenon , and , in our case , produces typical errors near the discontinuities of the driving function @xmath112 .
these discontinuities are due to the sudden obstructions of the radiation on parts of the aircraft that occur at certain orbital positions , for example , when the sun is eclipsed .
is always continuous but it undergoes sharp variations at some times .
these sharp variations can be considered as discontinuities , especially , if the function is sampled . ]
section [ dns ] shows that the gibbs phenomenon at eclipse points can be responsible for the largest part of the error of the linear method when the discrete fourier transform is used .
the approximation of @xmath178 provided by the @xmath29 samples of @xmath7 is , of course , @xmath194 , \label{d_fourier - sol}\ ] ] and is valid for any @xmath108 $ ] .
however , if we are only interested in @xmath195 at @xmath196 , @xmath197 we can compute these values with the inverse discrete fourier transform @xmath198 i/\t - j\right)^{-1 } \cdot \hat{f}(m)\ , , \label{i_d_fourier - tr}\ ] ] where , for @xmath199 @xmath200 , and where @xmath201 gives the remainder of the integer division by @xmath29 .
this inverse discrete fourier transform can be more convenient for a fast numerical computation . regarding computational convenience , the discrete fourier transform ,
be it direct or inverse , is best performed with a fast fourier transform ( fft ) algorithm .
the classic fft algorithm requires @xmath29 to be a power of two @xcite ; in particular , it has to be even .
the function @xmath178 , computed by fourier analysis from @xmath29 samples of @xmath34 , is to be compared with the one computed by a numerical approximation of the integral formula , eq .
( [ t_1_lim ] ) , in terms of the same samples .
naturally , we can use instead of the integral over @xmath202 $ ] the integral over @xmath203 $ ] below eq .
( [ t_1_lim ] ) . this integral can be computed from the @xmath29 samples of @xmath7 by an interpolation formula , say the trapezoidal rule .
it is not easy to decide whether this procedure is more efficient than fourier transforms .
considering that the substitution of the continuous fourier transform , eq .
( [ fourier - coef ] ) , by the discrete transform , eq .
( [ d_fourier - coef ] ) , is equivalent to computing the former with the trapezoidal rule , the integral formula may seem more direct .
in particular , this formula allows us to select the values of @xmath41 for which we compute @xmath178 independent of the sampling frequency , so we can choose just a few distinguished orbital positions and avoid the computation of all the @xmath189 integrals ( one is removed by the condition @xmath204 ) .
note that the computation of all of the independent @xmath8 with eq .
( [ d_fourier - coef ] ) is equivalent to the computation of precisely @xmath189 integrals .
however , the efficiency of the fft reduces the natural operation count of this computation , of order @xmath205 , to order @xmath206 ; so its use can be advantageous , nevertheless .
it goes without saying that the second - order perturbative contribution @xmath207 to the stationary solution is given by the right - hand side of eq .
( [ d_fourier - sol ] ) with the fourier coefficients @xmath8 replaced by the fourier coefficients of the function @xmath19 defined in sect .
[ second - order ] .
to test the previously explained methods , we construct a small thermal model of a simple spacecraft , namely , a ten - node model of a moon - orbiting satellite .
our satellite ten - node model supports a basic thermal structure and is simple enough for allowing one to explicitly display the main mathematical entities , e.g. , the matrices @xmath24 , @xmath67 and @xmath20 .
the satellite consists of a rectangular parallelepiped ( a cuboid ) of square base plus a small cylinder on one of its sides that simulates an observation instrument , as represented in fig .
in addition , at a height of two thirds of the total height , there is an inner tray with the electronic equipment .
the dimensions of the cuboid are @xmath208 , and the cylinder has a length of 0.1 m and a radius of 0.04 m. the satellite s frame is made of aluminum alloy , using plates 1 mm thick , except the bottom plate , which is 2 mm thick .
this plate plays the role of a radiator and its outer surface is painted white to have high solar reflectance .
the cylinder is made of the same aluminum alloy , as well as the tray ; they are 0.5 mm and 2 mm thick , respectively .
the sides of the satellite , except the one with the instrument , are covered with solar cells , which increase the sides thickness to 2.25 mm .
the thermal model of the satellite assigns one node to each face of the cuboid , one more to the cylinder and another to the tray , that is , eight nodes altogether .
furthermore , to conveniently split the total heat capacitance of the electronic equipment , it is convenient to add two extra nodes with ( large ) heat capacitance but with no surface that could exchange heat by radiation . nodes of this type are called `` non - geometrical nodes '' . in the present case , they represent two boxes with equipment placed above and below the tray , respectively .
we order the ten nodes as shown in fig .
the lower box ( node 10 ) is connected to the radiator by a thermal strap .
given the satellite s structure and assuming appropriate values of the specific heat capacities , it is possible to compute the capacitances @xmath209 with the result given in table [ tab1 ] .
using the value of the aluminum - alloy heat conductivity and assuming perfect contact between plates , we compute the conduction coupling constants @xmath27 between nodes @xmath210 .
the remaining conduction coupling constants are given reasonable values , shown in eq .
( [ kij ] ) .
the computation of the radiation coupling constants @xmath211 and @xmath212 and indeed the computation of the external radiation heat inputs requires a detailed _ radiative model _ of the satellite , consisting of the geometrical _ view factors _ and the detailed thermo - optical properties of all surfaces .
this radiative model allows us to compute the respective _ absorption factors _ @xcite . .node capacities and mean heat inputs with their associated steady - state temperatures .
[ cols=">,>,>,>",options="header " , ] the thermo - optical properties of the surfaces are assumed to be as realistic as possible , given the simplicity of the thermal model .
all radiation reflection is assumed to be diffuse , as is common for many types of surfaces .
the inner surfaces are painted black and have high emissivity , @xmath213 , to favor the uniformization of the interior temperature .
the outer surfaces are of three types .
the three sides covered with solar cells also have high emissivity , @xmath214 , to favor the cooling of the solar cells . on the other hand
, they have high solar absorptivity , @xmath215 . of this 0.75
, 0.18 is processed into electricity and the remaining 0.57 dissipates as heat in the solar cells .
the top surface , the surface with the cylinder , and the cylinder itself ( its two sides ) have low emissivity , @xmath216 , and low solar absorptivity , @xmath217 , which are chosen to simulate the effect of a _ multilayer insulator_. in contrast , the bottom surface simulates a radiator , with @xmath218 and @xmath217 ( like an _ optical solar reflector _ ) .
all of these thermo - optical properties are summarized in fig .
[ opt ] . for the computation of the corresponding absorption factors , we employ the ray - tracing monte - carlo simulation method provided by esarad@xmath44 ( esatan@xmath44 s radiation module ) @xcite .
taking into account the above information , one obtains the following conduction ( in w / k ) and radiation ( in w / k@xmath38 ) matrices : @xmath219 the satellite s thermal characteristics are defined by the data set @xmath220 but the radiation heat exchange depends on the nodal temperatures , which in turn depend on the heat input . as explained in sect .
[ pert - eq ] , the appropriate set of nodal temperatures corresponds to the steady - state for averaged heat inputs , given by the algebraic equation that results from making @xmath221 in eq .
( [ bode - pert ] ) .
since we need the external heat inputs and , therefore , the orbit , we proceed to define the orbit characteristics .
we choose a circular equatorial orbit @xmath222 m above the moon s surface , such that @xmath223 s. the radiation heat input to the satellite consists , on the one hand , of the solar irradiation and the moon s albedo , and , on the other hand , of the moon s constant ir radiation .
we take 0.12 for the mean moon s albedo and 270 k for the black - body equivalent temperature of the moon .
there is also heat produced by the dissipation of electrical power in the equipment ( nodes 9 and 10 ) . for the sake of simplicity
, the dissipation rate is assumed to be constant , equal to the mean electrical power generated in an orbit . in a part of the orbit ,
the moon eclipses the sun , so the satellite receives no direct sunlight or albedo , although there is always ir radiation from the moon .
the satellite is stabilized such that the cylinder ( the `` observation instrument '' ) always points to the moon and the longer edges are perpendicular to the orbit .
the radiation heat input can be computed by taking into account the given orbital characteristics and the satellite s thermo - optical characteristics , in particular , the absorption factors .
it has been computed with esarad@xmath44 , taking 111 positions on the orbit , that is , at intervals of one minute . in fig .
[ orbit ] , all 111 positions are plotted .
the initial position of the satellite is at the subsolar point and it moves towards the east .
the total external radiation heat input to the first eight nodes ( the ones that receive radiation ) is plotted in fig .
[ heat - in ] ( only at every other position , for clarity ) .
note the symmetry between nodes 1 and 3 , which denote the lateral faces , covered with solar cells .
node 4 corresponds to the back side , also covered with solar cells .
so the external radiation load on it has a similar time variation , but it is displaced . the solar radiation absorbed by all solar cells results in an orbital mean power rate of 10.5 w , dissipated in the equipment and split between nodes 9 and 10 , which receive 4.35 and 6.15 w , respectively . the external radiation absorbed by the side with the cylinder ( node 2 ) is considerably smaller than the radiation absorbed by the sides with solar cells , due to the low value of @xmath224 ( and of @xmath225 , as well ) for the corresponding surface .
the bottom and top outer surfaces , which belong to nodes 5 and 6 , respectively , have view factors for the external radiation that are much less favorable than those of the side surfaces . nevertheless , the amount of lunar ir radiation absorbed by the bottom surface , due to its high @xmath225 , is such that the orbital mean of the external heat input to node 5 is , in fact , larger than the one for node 2 ( see table [ tab1 ] ) .
naturally , node 7 , with no outer surfaces , does not absorb any external radiation .
+ to determine the hot and cold cases of the orbit , we compute the total heat load on the satellite for each position in the orbit , finding a maximum of 90.59 w at position 14 and a minimum of 18.65 w at any position in the eclipse , during which all the heat loads stay constant .
the solution of the corresponding steady - state problems at position 14 and at a position in the eclipse yields the two sets of nodal temperatures ( for the given node order ) : hot : : @xmath226 .
cold : : @xmath227 . the results in sect .
[ j&p - sol ] show that the periodic thermal state does not reach these extreme temperatures , which could endanger the performance of the satellite .
we compute the averages @xmath83 and substitute for them and the data set @xmath228 in eq .
( [ bode - pert ] ) to find the steady state temperatures .
the results are given in table [ tab1 ] .
then , according to eqs .
( [ jij ] ) and ( [ jii ] ) , the jacobian matrix is ( in s@xmath21 ) @xmath229 by inspection , one can check that it has nonnegative off - diagonal and negative diagonal elements , that is to say , @xmath94 is a @xmath95-matrix .
it is also _ diagonally dominant _ , namely , @xmath230 their inverses ( in absolute value ) give us the typical relaxation times of the corresponding thermal modes .
thus , we deduce that relaxation time of the fastest mode is about 55 s , whereas the relaxation time of the slowest one is @xmath231 s. the latter time is similar to @xmath232 s. the eigenvalues are real numbers and , furthermore , @xmath20 is diagonalizable , because the eigenvalues are different .
both properties also follow from @xmath155 being almost symmetric : its antisymmetric part , @xmath233 , is relatively small , namely , @xmath234 , where the matrix norm is the frobenius norm ( other standard matrix norms yield similar values ) .
therefore , the notion of `` radiation conductance '' ( sect . [ heur ] ) is appropriate in this case , as concerns its use in the linear equations . the thermal modes are almost normal , namely , the eigenvector matrix @xmath31 is such that @xmath161 with an error @xmath235 the most interesting eigenvector of @xmath20 is , of course , the positive ( perron ) eigenvector , which corresponds to the slowest mode .
the normalized positive eigenvector is @xmath236 note that the temperature increments are of a similar magnitude , except the ones of node 7 and , especially , nodes 9 and 10 , which are associated , respectively , for the tray and the boxes of electronic equipment .
the next mode , corresponding to the eigenvalue @xmath237 , has one negative component ( the ninth ) , and the remaining modes have more than one . to calculate @xmath178 , we choose the fourier series of eq .
( [ d_fourier - sol ] ) or , rather , the inverse discrete fourier transform of eq .
( [ i_d_fourier - tr ] ) , which can be computed with a fft algorithm .
the fourier coefficients @xmath8 can also be computed with the fft , according to eq .
( [ d_fourier - coef ] ) .
once the vector @xmath238 at the 111 positions is available , the set of nodal temperatures corresponding to the first - order perturbative solution is @xmath239 plotted in fig .
[ tevol ] .
a measure of the accuracy of this perturbative calculation is given by the second - order calculation in the next section .
the truncation of the fourier series imposed by the sampling of @xmath112 also is a source of error , unrelated to perturbation theory .
the piecewise smoothness of the function @xmath178 suggests that the error is small ( but see sect .
[ dns ] ) .
+ it is also interesting to see if the first - order perturbative calculation is affected by neglecting the fastest modes : according to the analysis at the end of sect .
[ pert - eq ] , these modes are expected to contribute in proportion to their relaxation times .
the fastest mode relaxes in about 55 s , a short but non - negligible time . as a consequence ,
its contribution to @xmath238 , which we find to have a maximum magnitude of 0.8 k , is small but non - negligible .
but we can deduce that still faster modes , which would appear in a thermal model of the satellite with more nodes , are hardly necessary . from the engineering standpoint , note that this satellite thermal model is successful , insofar as it predicts that all nodal temperatures stay within adequate ranges .
in particular , nodes 9 and 10 , corresponding to the boxes with electronic equipment , stay within the range from @xmath240 to 23 @xmath241 .
these nodes are inner nodes with large thermal capacity and , hence , are protected against the larger changes in the external heat inputs .
in contrast , the outer nodes are very exposed and undergo considerable variation in temperature , with especially sharp changes at the beginning and end of the eclipse . + according to sect .
[ second - order ] , the second - order perturbative correction @xmath242 to the periodic stationary solution is obtained by the same procedure as that for @xmath243 but using a different driving function @xmath126 that is computed from @xmath45 and from @xmath244 itself .
the computations are straightforward and they yield the correction plotted in fig .
[ t2evol ] .
this correction is always negative , because the negative term in the expression for @xmath126 , eq .
( [ tf ] ) , dominates over the positive term . the equation ( [ t_1_lim ] ) for @xmath238 and the corresponding equation for @xmath242 are both linear , so @xmath238 and @xmath242 are proportional to the respective driving functions ; and we can compare their magnitudes by comparing those driving functions , say , comparing typical values of @xmath245 and @xmath246 .
this latter quantity can be roughly estimated as @xmath247 w @xmath248 w , whereas @xmath249 w ( table [ tab1 ] ) .
their ratio is about 25 , which roughly agrees with the ratio of @xmath238 to @xmath242 , as can be seen by comparing fig .
[ tevol ] to fig .
[ t2evol ] .
the order of magnitude of the second - order correction suggests that higher orders are not necessary , as we show next .
of course , the periodic solution @xmath250 can also be obtained by direct integration of the nonlinear eqs .
( [ ode ] ) with an adequate solver , based on runge - kutta or other methods @xcite .
since the nonlinear analysis proves that the solution of eqs .
( [ ode ] ) will converge to the stationary periodic solution @xcite , the numerical solver must be run until this periodicity is established .
periodicity can be enforced by comparing the nodal temperatures at the beginning and the end of every period and demanding that they be equal , within some tolerance . to do this
, esatan@xmath44 provides the routine solcyc @xcite .
we have employed this routine to obtain the nonlinear equations cyclic solution ( which is established in 10 periods if the tolerance is set to 0.001 ) . to compare the cyclic solution obtained by the linear method with the `` exact '' solution @xmath250 obtained by solcyc
, we quantify the deviation by the vector of `` errors '' @xmath251.\ ] ] the largest components of @xmath252 are plotted in fig .
[ terror ] ( the remaining components stay in the range @xmath253\,^{\circ}\mathrm{c}$ ] ) .
there seem to be two branches for each node , but , in fact , it is an effect produced by the high - frequency oscillations of each @xmath254 .
notice that the error is generally small when compared with @xmath207 ; namely , @xmath255 for each @xmath256 , except near the two positions corresponding to the beginning and end of the eclipse , where the heat inputs have discontinuities ( fig .
[ heat - in ] ) .
these discontinuities induce oscillations of certain amplitude about the true values of @xmath250 , due to the gibbs phenomenon in the discrete fourier transform ( sect .
[ fourier - anal ] ) . to suppress the oscillations , we would need a specific method for treating the gibbs phenomenon ; for example , we could use a smoother method for fourier series summation , such as cesro summation @xcite
the ten - node model studied in sect .
[ ten - node ] is too small to pose a computational problem , whether we employ the linear method or directly integrate nonlinear eqs .
( [ ode ] ) . to assess the practical applicability of the linear method
, we must study how it scales to realistic sizes and then compare it with the direct integration of eqs .
( [ ode ] ) .
naturally , the first step in the method is to compute the steady - state temperature @xmath257 but this computation , arguably , is not a substantial part of the whole process and we do not consider it ( it may take , say , between a few percent and one fith of the whole process , depending on the circumstances ) .
the size of an orbiting spacecraft thermal model can be scaled with respect to the number of nodes , @xmath30 , or the number of different positions in the orbit , @xmath29 . however , while realistic models must have many nodes , an @xmath29 of about one hundred can generally be suitable .
note , in particular , that the condition that @xmath258 in the ten - node model is likely to be reversed in realistic models .
for these reasons , the scaling with respect to @xmath30 is more relevant . for the moment ,
let us neglect any special feature of eqs .
( [ ode ] ) , such as possible coefficient sparsity .
then , the computational complexity of numerically integrating those equations is of order @xmath259 , where @xmath23 is the number of time steps taken . on the other hand , the complexity of the numerical integration of linear eqs .
( [ linode - pert ] ) is also of order @xmath260 [ excluding the computation of @xmath98 ] .
we can employ the explicit integral , eq . ( [ t_1 ] ) , which can be calculated with the trapezoidal rule , for example , but this does not reduce the complexity of the computation . however , for the stationary solution , eq . ( [ t_1_lim ] ) , its expression as an integral over a period @xmath79 sets the number of time steps to @xmath29 , which is advantageous if @xmath262 , where @xmath23 now is the number of steps necessary for some initial temperatures to relax to the stationary solution . since we have to evaluate the integral in eq .
( [ t_1_lim ] ) for several times @xmath41 , it is preferable to use the fft , as discussed in sect .
[ fourier - anal ] , so that the total operation count is of order @xmath263 .
however , we have not taken into account matrix operations of considerable complexity .
for example , the discrete fourier transform , eq .
( [ d_fourier - sol ] ) , involves the _ inverse _ of an @xmath264 matrix ( @xmath94 plus a multiple of the identity matrix ) , and the inversion of a matrix is generally a process of order @xmath265 @xcite . if we have to employ a process of order @xmath265 , we may as well diagonalize @xmath20 , because the diagonalization of a matrix also is generally a process of order @xmath265 @xcite and the diagonalization of @xmath20 has several uses .
let us assume that we carry out this diagonalization and determine the independent thermal modes , which can then be employed to express eq .
( [ t_1_lim ] ) as eq .
( [ t_1_eigenvec ] ) or to simplify the matrix operations in eq .
( [ d_fourier - sol ] ) .
if we roughly compare the computational complexity of order @xmath265 with the numerical integration complexity of order @xmath259 , we deduce that the diagonalization is worthwhile when @xmath266 . using the above estimate @xmath267 and taking as relaxation time @xmath268 ( the rough value for the ten - node model )
, we deduce that the determination of thermal modes can be useful just as a computational procedure for models with a few hundred nodes .
let us now consider that @xmath20 is surely a sparse matrix , so iterative matrix methods can take advantage of this characteristic .
in fact , there are two degrees of sparsity in @xmath20 , associated with conduction or radiation coupling terms .
the conductance matrix @xmath24 has to be very sparse , because conduction is a local process , so each node can only be coupled to a few nodes .
in contrast , radiation is a nonlocal process and couples any pair of nodes that has a nonvanishing view factor .
in addition to the different sparsity of conduction and radiation coupling matrices , there are two other circumstances that make them different : ( i ) we are assuming that the conductive coupling matrix just depends on material properties , so it is independent of the reference temperatures @xmath52 ; ( ii ) the conduction coupling terms are significantly larger than the radiation coupling terms for the natural values of those temperatures .
all of this suggests separating the conduction and radiation parts of @xmath20 , and then diagonalizing the conduction part .
based on the study in sect .
[ heur ] , the diagonalization of the conduction part of @xmath20 boils down to the diagonalization of the ( generalized ) laplacian matrix @xmath269 and this matrix is sparse .
suitable iterative algorithms to perform this diagonalization are , for example , the lanczos @xcite or the davidson @xcite algorithms . these algorithms are particularly useful when only a few of the largest or smallest eigenvalues are needed .
this is indeed our case , as only the slower modes are expected to contribute to @xmath238 and @xmath270 once the conduction part of @xmath20 has been diagonalized , in the sense that the lowest eigenvalues and eigenvectors of the corresponding laplacian matrix are known , the radiation part of @xmath20 can be treated as a perturbation , using matrix perturbation methods @xcite .
iterative matrix methods , combined with matrix perturbation methods or other methods , if necessary , can reduce the order @xmath265 to @xmath271 or even to almost linear and so allow us to diagonalize the jacobians for the largest values of @xmath30 that appear in current thermal spacecraft models .
of course , the sparsity of thermal coupling matrices also facilitates the direct integration of the nonlinear eqs .
( [ ode ] ) .
however , their nonlinearity prevents one from taking advantage of the above mentioned approximations methods , for example , the reduction to the small subspace of slow modes , or the splitting into conduction and radiation in which the latter is treated as a matrix perturbation .
moreover , the linearization is useful in various respects .
for example , the overall relaxation time , given by the eigenvalue of smallest magnitude , can be effectively bounded by inequalities @xcite and some of these bounds can be found with little computational effort .
we have studied the evolution of the thermal state of an orbiting spacecraft and developed a linear approach to this problem that is based on a rigorous perturbative treatment of the exact nonlinear equations .
the first - order perturbation equations , eqs .
( [ linode - pert ] ) , constitute the basic linear system , which can be applied to higher orders after calculating the corresponding driving terms .
as the jacobian matrix of the nonlinear equations has negative eigenvalues , the linear equations describe the relaxation to a stationary thermal state , namely , a periodic solution that is independent of the initial conditions and only depends on the external heat input .
this relaxation is similar to the relaxation to steady - state under constant external heat load .
we have shown that the perturbative treatment reveals the scope of a common linearization procedure of a heuristic nature , in which the nonlinear equations are rendered linear by the definition of radiation conductances ( sect .
[ heur ] ) . if one previously calculates , with the correct nonlinear eqs .
( [ bode - pert ] ) , the reference steady - state condition that corresponds to the average external heat input , the deviation from that steady - state is well approximated by the linear equations with radiation conductances . the jacobian matrix corresponding to radiation conductances , obtained in eqs .
( [ njij ] ) and ( [ njii ] ) , is related to a symmetric matrix and , therefore , is easier to diagonalize .
furthermore , this relation implies that the thermal modes are normal , like the vibrational modes of a mechanical system .
although the notion of radiation conductance is just an approximation , it serves nonetheless to show that the jacobian matrix is diagonalizable and has real eigenvalues .
the diagonalization of the jacobian matrix is useful for the computation of the stationary thermal state and also provides information on the relaxation to that state , because the relaxation times of the thermal modes are the inverses of the eigenvalues .
these times span a considerable range , but the longest times are much more significant than the shortest times , because the latter depend on the details of the lumped - parameter thermal model employed whereas the former are essentially independent of it .
in fact , a thermal model that has more nodes and therefore more details also has more thermal modes ; but the slowest modes , which correspond to temperature changes in large parts of the spacecraft , are hardly affected by the details , whereas the fast modes can be significantly altered .
the slowest mode , in particular , corresponds to a simultaneous but non - uniform increase ( or decrease ) of the temperature throughout the spacecraft and is hardly altered by small - scale changes .
the computation of the stationary thermal state with the linear method relies on an explicit integral , eq . ( [ t_1_lim ] ) , or a fourier expansion , eq .
( [ d_fourier - sol ] ) . given a sampling of the thermal driving function at equal time intervals
, the periodic solution can be obtained through two discrete fourier transforms : a direct transform to get the fourier coefficients of the driving function and an inverse transform of the coefficient vector multiplied by a suitable matrix ( sect .
[ fourier - anal ] ) .
of course , the discrete fourier transforms are best performed with a fft algorithm .
this computation is more efficient than the numerical computation of the integral , eq . ( [ t_1_lim ] ) , if we need the values of the temperatures at all the given sampling times .
however , the fourier transform presents the gibbs phenomenon , associated with sudden variations of the heat loads , as occur at eclipse times , for example .
the gibbs phenomenon introduces errors , but these errors could be suppressed with special methods .
the computation of the thermal modes and the stationary thermal state for a satellite ten - node thermal model confirms the validity of the linear method for a minimal but realistic model .
the relaxation times span a considerable range , between 55 seconds and nearly one hundred minutes . of course
, the latter time must be almost independent of the particular thermal model used , whereas the former has no intrinsic significance , and , if the number of nodes grew , that time would shrink ( thus further expanding the range of relaxation times ) .
the slowest mode corresponds to node temperature increments with the same sign ( positive by convention ) , whereas the increments corresponding to other modes have both signs .
the periodic variation in the external heat input ( fig .
[ heat - in ] ) excites the thermal modes and produces a definite pattern of stationary temperature oscillations , well approximated by the first - order solution ( fig .
[ tevol ] ) .
the second - order correction is small compared to the first - order solution , but it is worth computing , as it reaches 1.7 k. higher order corrections are essentially negligible , but the error due to the gibbs phenomenon at the eclipse positions reaches 0.6 k ( at the most ) . focusing on the computational aspects of the linear approach , we have studied how it scales with the number @xmath30 of nodes and the number @xmath29 of sampling positions on the orbit .
if the jacobian matrix is dense , the complexity of the corresponding matrix operations is of order @xmath265 .
it is convenient to employ just one matrix operation , namely , the diagonalization of the jacobian matrix , because then only a few of the slowest modes are needed for the remaining operations , so these have negligible complexity .
the complexity of a direct numerical integration of the nonlinear equations is of order @xmath259 , @xmath23 being the number of time steps necessary for relaxation . for a low - altitude orbit
, @xmath23 is expected to be on the order of one thousand , as for our moon - orbiting satellite .
therefore , the linear method would be computationally effective as just an integration method only for models with a few hundred nodes . at any rate
, the jacobian matrix can be assumed to be sparse , and its conduction part can be assumed to be especially sparse , in addition to being the larger part of the jacobian matrix and also being independent of the orbit . as the orbit may be subjected to changes in the planning of a mission ,
a convenient strategy probably is to diagonalize the conduction part at the outset and , when needed , add the radiation part within some approximation scheme
. this strategy can be far more efficient than integrating the nonlinear equations each time .
moreover , the strength of the linear approach lies with the insight that it provides about the thermal behavior of the spacecraft , as embodied by the decomposition of its thermal modes , of which only the slowest ones are significant .
these significant modes can actually be obtained with a reduced thermal model using few nodes .
therefore , the linear approach is especially useful in the context of reduced models .
furthermore , it provides a method for model reduction based on the mode decomposition : this decomposition can be used to group nodes .
indeed , there is a technique for graph partitioning based on the eigenvalues and eigenvectors of the laplacian matrix of the graph @xcite . according to sect .
[ heur ] , this technique is applicable to the jacobian matrix , but the details of this application are beyond the scope of the present paper and are left for future work . finally , our linear approach can surely be applied to other cyclic heating processes that involve radiation heat transfer .
we thank isabel prez - grande for bringing ref . to our attention .
f. kreith , _ radiation heat transfer for spacecraft and solar power plant design_. intnal .
textbook co. , scranton , penn . | we develop a linear method for solving the nonlinear differential equations of a lumped - parameter thermal model of a spacecraft moving in a closed orbit .
our method , based on perturbation theory , is compared with heuristic linearizations of the same equations .
the essential feature of the linear approach is that it provides a decomposition in thermal modes , like the decomposition of mechanical vibrations in normal modes .
the stationary periodic solution of the linear equations can be alternately expressed as an explicit integral or as a fourier series .
we apply our method to a minimal thermal model of a satellite with ten isothermal parts ( nodes ) and we compare the method with direct numerical integration of the nonlinear equations .
we briefly study the computational complexity of our method for general thermal models of orbiting spacecraft and conclude that it is certainly useful for reduced models and conceptual design but it can also be more efficient than the direct integration of the equations for large models .
the results of the fourier series computations for the ten - node satellite model show that the periodic solution at the second perturbative order is sufficiently accurate .
lcl @xmath0 & = & outward - facing area of _ _ i__th - node , m@xmath1 + @xmath2 & = & thermal capacitance diagonal matrix , j / k + @xmath3 & = & thermal capacitance of _ _ i__th - node , j / k + @xmath4 & = & eigenvector of @xmath5 , k + @xmath6 & = & _ _ j__th - component of eigenvector @xmath4 , k + @xmath7 & = & driving function vector , k / s + @xmath8 & = & @xmath9th fourier coefficient of @xmath7 , k / s + @xmath10 & = & complex conjugate of @xmath11 , k / s + @xmath12 & = & driving function for @xmath13th - mode , k / s + @xmath14 & = & time derivative of @xmath12 , k / s@xmath15 + @xmath16 & = & driving function for @xmath17th - node , k / s + @xmath18 & = & @xmath9th fourier coefficient of @xmath16 , k / s + @xmath19 & = & driving function for second order temperature vector , k / s + @xmath20 & = & jacobian matrix , s@xmath21 + @xmath22 & = & jacobian matrix _ ij_-element , s@xmath21 + @xmath23 & = & number of steps in numerical integration of odes + @xmath24 & = & conduction coupling matrix , w / k + @xmath25 & = & matrix of conductances from linearized radiation coupling terms , w / k + @xmath26 & = & @xmath17th - node conductance from linearized environment - radiation terms , w / k + @xmath27 & = & conduction coupling matrix _
ij_-element , w / k + @xmath28 & = & _ ij_-element of conductance matrix from linearized radiation terms , w / k + @xmath29 & = & number of samples for discrete fourier transform + @xmath30 & = & number of nodes + @xmath31 & = & eigenvector matrix for _ j _ , k + @xmath32 & = & _ _ i__th - component of @xmath20-eigenvector for @xmath13th - mode , k + @xmath33 & = & auxiliary heat input vector , k / s + @xmath34 & = & heat input to _ _ i__th - node , w + @xmath35 & = & mean value of @xmath34 over period @xmath36 , w + @xmath37 & = & _ _ i__th - node coefficient of radiation to the environment , w / k@xmath38 + @xmath39 & = & radiation coupling matrix , w / k@xmath40 + @xmath41 & = & time , s ( min in figures ) + @xmath42 & = & temperature vector , k + @xmath43 & = & vector of `` errors '' w.r.t .
esatan@xmath44 solution , k + @xmath45 & = & steady - state temperature vector , k + @xmath46 & = & temperature of _ _ i-__node , k + @xmath47 & = & cosmic microwave background radiation temperature , k + @xmath48 & = & _ _ n__th order term in _ _ j__th - node temperature expansion , k + @xmath49 & = & stationary solution temperature vector , k + @xmath50 & = & _ _ n__th order of stationary solution temperature vector , k + @xmath51 & = & time derivative of _ _ i__th - node temperature , k / s + @xmath52 & = & steady - state temperature of _ _ i__th - node , k + @xmath36 & = & orbital period , s + @xmath53 & = & matrix of independent solutions to homogeneous equations , k + @xmath54 & = & solar absorptivity + @xmath55 & = & kronecker s delta + @xmath56 & = & antisymmetric part of @xmath5 , s@xmath21 + @xmath57 & = & _ ij_-element of @xmath56 , s@xmath21 + @xmath58 & = & infrared emissivity of _ _ i__th - node outward - facing surface + @xmath59 & = & perturbation parameter + @xmath60 & = & eigenvalue of _ j _ for @xmath13th mode , s@xmath21 + @xmath61 & = & perturbation of _
j_-eigenvalue for @xmath13th mode ,
s@xmath21 + @xmath62 & = & stefan - boltzmann constant , @xmath63 wm@xmath64k@xmath65 + |
in recent years hundreds of exoplanets have been discovered , many of which ( @xmath021% ) have mass greater than or equal to that of jupiter and orbital semi - major axes of @xmath1 au ( where @xmath2 ) , although a significant fraction ( @xmath061% ) of planets with jupiter s mass or greater have been observed with semi - major axes @xmath3 au ( see , e.g. , the catalogue at exoplanet.eu ) .
the possibility of detection of the auroral radio emissions of ` hot jupiter'-like exoplanets close to their parent star has been considered by a number of authors ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this interest has been sparked in part by the imminent commencement of high sensitivity radio observations by next generation radio telescopes such as the low frequency array ( lofar ) , which will have a detection threshold of 1 mjy ( where @xmath4 ) @xcite .
such previous consideration of hot jupiter - like exoplanets has assumed that the auroral radio emission would be caused by a star - planet interaction reminiscent of either the solar wind - earth interaction or the io - jupiter interaction .
the former is mediated primarily via reconnection of the planetary and interplanetary magnetic fields at the dayside magnetopause @xcite .
this drives plasma flows within the magnetosphere that generate electric currents flowing between the magnetosphere and the resistive ionosphere , the upward magnetic field - aligned component of which , associated with downward - precipitating electrons , produces auroral and radio emissions .
the latter interaction is thought to be mainly associated with the generation of alfvn waves in the vicinity of the moon io , caused by its motion through the rapidly - rotating planetary magnetic field and plasma @xcite .
consideration of such processes has led to the extrapolation of a ` radiometric bode s law ' relating incident magnetic power to output radio power for the case of hot jupiter - like exoplanets orbiting extremely close ( at typically @xmath010 stellar radii ) of their parent stars .
it has been concluded that such interaction may generate emissions at or above the lofar detection threshold @xcite .
+ however , despite the importance placed by previous authors on stellar wind - planet and io - jupiter type interactions , significant components of jupiter s radio emissions , i.e. the b - kom , hom and non - io - dam emissions @xcite , are thought to be generated by the large - scale magnetosphere - ionosphere ( m - i ) coupling current system associated with the breakdown of corotation of iogenic plasma in jupiter s middle magnetosphere , illustrated by fig .
[ fig : miccs ] @xcite .
this process generates intense field - aligned electron beams which drive the brightest and most significant of jupiter s uv auroral emission , i.e. the main auroral oval @xcite , and , coupled with particle mirroring and the absorption of particles in the loss cone , excite the cyclotron maser instability in the high - latitude low-@xmath5 plasma , which gives rise to the above radio emissions .
observationally , the uv aurora and radio emissions of jupiter and saturn have been shown by a number of studies to be closely associated with one another ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
io orbits deep within jupiter s magnetosphere at @xmath05.9 @xmath6 ( where @xmath6 represents the equatorial radius of jupiter equal to 71,373 km ) , and its volcanoes liberate sulphur and oxygen atoms into a torus surrounding the moon s orbit at the rate of @xmath01000@xmath7 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
these atoms are ionised by electron impact ionisation and thus become sensitive to the rotating planetary magnetic field , such that the newly - created plasma is picked up to corotation velocity .
the picked - up plasma is centrifugally unstable and diffuses radially away from the planet , probably via flux - tube interchange motions @xcite , before being lost down the dusk flank of the magnetotail via the pinching off of plasmoids ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . + as the plasma diffuses radially outward , its angular velocity drops ( inversely with the square of the distance if no torques act ) due to conservation of angular momentum , such that a radial gradient of angular velocity is set up in the equatorial plasma .
this angular velocity gradient , when mapped along the magnetic field to the ionosphere , causes an equatorward - directed ( for jupiter s magnetic field polarity ) ionospheric pedersen current to flow , the @xmath8 force of which opposes the drag of the neutral atmosphere on the sub - rotating plasma .
angular momentum is transferred between the ionosphere and the equatorial plasma by the sweep - back of magnetic field lines into a lagging configuration , such that the ionospheric pedersen current is balanced in the equatorial plane by an outward - directed ( again , for jupiter s magnetic field polarity ) radial current , the @xmath8 force of which tends to return the equatorial plasma back to corotation with the planet .
current continuity between these two field - perpendicular currents is maintained via field - aligned ( birkeland ) currents , the inner upward component of which is thought to generate jupiter s main auroral oval emission @xcite .
the current system was studied in detail theoretically by @xcite , who considered the effect of two poorly - constrained but important system parameters , the effective ionospheric pedersen conductance @xmath9 , and the plasma mass outflow rate @xmath10 , and they derived analytic approximations appropriate for small and large radial distances , the former of which will be instrumental in the present work . @xcite
went on to examine the effect on the current system of modulation of the ionospheric pedersen conductance due to auroral electron precipitation , and @xcite and @xcite have studied the effect of field - aligned voltages . @xcite considered the effect on the current system of solar wind - induced expansions and contractions of jupiter s magnetosphere , and recently the modulation of the current system by diurnal variation of the ionospheric pedersen conductance caused by solar illumination has been considered by @xcite .
+ in this paper we consider the application of the model describing jupiter s magnetosphere - ionosphere coupling current system to jupiter - like exoplanets with internal plasma sources such as active moons .
the stability of satellites about exoplanets has been considered ( see e.g. @xcite and references therein ) , since this is an important issue for the habitability of moons orbiting close - in ` hot jupiters ' . however , this is much less of a problem for satellites about more distant - orbiting exoplanets , which will be shown to be important for the present study .
for example , @xcite could not place any mass limits on satellites orbiting planets beyond @xmath00.6 au .
the vulcanism of io is driven by tidal dissipation , which also acts to dampen the eccentricity of the moon s orbit , such that the small eccentricities of the orbits of the galilean satellites of @xmath00.001 - 0.01 are only raised due to laplace resonance @xcite .
the outer limit for stable orbits of a prograde moon was given by @xcite to be @xmath11 where @xmath12 is the semi - major axis of the planet s orbit , @xmath13 is the mass of the planet , @xmath14 is the mass of the parent star , @xmath15 is the eccentricity of the planet s orbit and @xmath16 is the eccentricity of the moon s orbit . for values appropriate to io ( i.e. @xmath17 au , @xmath18 kg , @xmath19 kg , @xmath20 , and @xmath21 ) , @xmath22 , such that volcanic satellites orbiting jovian - like planets are expected to be stable ( as is of course observed in the solar system , e.g. io orbits at 5.9 @xmath6 , well inside @xmath23 ) . indeed , for jovian mass planets orbiting beyond 1 au , satellites with roughly io s orbital distance will be typically stable , and only the planets with very eccentric orbits and high mass parent stars ( e.g. @xmath24 au , @xmath25 , @xmath26 , such that @xmath27 ) would be expected to be devoid of stable satellites at io s orbital distance . + both giant planets in our solar system which have been extensively studied by in - situ spacecraft , i.e. jupiter and saturn , possess moons which actively outgas into the near - planetary space . at jupiter
this is io , which emits material at the rate of @xmath01000 @xmath28 as discussed above , whilst at saturn this is enceladus , whose cryo - volcanoes emit water group ions at estimated rates ranging from a few to a few hundred @xmath28 ( e.g. * ? ? ?
it is thus reasonable to assume that such active moons may be relatively prevalent amongst jupiter - like exoplanets , and the resulting implications for detectability of auroral radio emissions is therefore considered here .
we begin from the assumption that jupiter - like exoplanets are strongly illuminated by their parent star , such that the ionospheric pedersen conductance is high enough to maintain near - rigid corotation throughout the magnetosphere , a condition which , as reviewed below , maximises the field - aligned current density for a given magnetosphere .
we then compute the currents , and thus the resulting radio power output , for varying configurations of jupiter - like exoplanets .
note that such strong illumination could be associated with close ` hot jupiters ' , but equally applies to more distant planets orbiting active stars .
the parameters examined are the plasma mass outflow rate , planetary orbital distance , and rotation rate .
we show that , for planets with host stars more active than the sun , only relatively modest modifications from the jovian system parameters are required to produce potentially - detectable configurations , and by doing so we open up the catalogue of potential candidates for detection by radio telescopes such as lofar to a class of planet previously overlooked .
in nt , ( b ) the associated flux function @xmath29 in @xmath30 , and ( c ) the ionospheric co - latitude @xmath31 to which the field lines threading the equatorial plane at radial distance @xmath32 map .
the red solid lines show the ` baseline ' model values given by eqs .
[ eq : bze ] and [ eq : fe ] , while the black long - dashed lines show the planetary dipole values for comparison .
the other solid lines show the values modified by using eqs .
[ eq : deltab ] and [ eq : deltaf ] for assumed magnetopause standoff distances @xmath33 , and by simply extending eqs .
[ eq : bze ] and [ eq : fe ] for @xmath34 .
the colours blue , green , red , and black correspond to @xmath35 values of 20 , 50 , 85 , and @xmath36 , respectively .
the horizontal dotted lines in panels ( b ) and ( c ) show the value of @xmath37 nt @xmath38 , and the corresponding limiting ionospheric co - latitude of the outer boundary of the model at @xmath39 .
, width=317 ] in this section we discuss the equations which describe the system , along with the approximations appropriate to the case of jupiter - like exoplanets .
the analysis follows that which has been described previously by @xcite , @xcite , @xcite , and @xcite , such that here we provide only a brief outline .
we start by outlining the magnetic field model , which acts as the basis for the current system .
it will become necessary to consider the variation of the current system with sub - solar magnetopause standoff distance @xmath40 , and here we thus describe the ` baseline ' model which is then modified by changing @xmath40 .
the magnetic field model is that which has been used previously for jupiter by @xcite .
it is an axisymmetric poloidal field with the magnetic axis co - aligned with the spin axis , such that it can be described by the flux function @xmath41 , related to the magnetic field @xmath42 by @xmath43 where @xmath44 is the perpendicular distance from the magnetic axis , @xmath45 is the distance along the axis from the magnetic equator , and @xmath46 is the azimuthal angle .
magnetic field lines are thus given by contours of @xmath47 , such that mapping between the ionosphere ( subscript i ) and equator ( subscript e ) is simply achieved by writing @xmath48 . for a near - dipolar ionospheric field
we have @xmath49 where @xmath50 is the perpendicular distance from the magnetic axis in the ionosphere , @xmath51 is the planet s equatorial dipole magnetic field strength ( taken initially to be equal to jupiter s value @xmath52 = 426,400 nt in conformity with the vip4 planetary field model of @xcite ) , @xmath53 is the planetary radius ( taken to be equal to jupiter s radius @xmath54 ) , and @xmath31 is the magnetic co - latitude . in this model ,
the magnitude of the north - south magnetic field threading the equatorial plane is given by @xmath55+b_\circ\left(\frac{r_p}{\rho_e}\right)^m\right\ } , \label{eq : bze}\end{aligned}\ ] ] where @xmath56 nt , @xmath57 @xmath58 , @xmath59 nt , and @xmath60 .
this form closely approximates the field model used by @xcite and @xcite , who employed the ` voyager-1/pioneer-10 ' ( ` can ' ) field model of @xcite in the inner region and the voyager-1 ( ` kk ' ) model of @xcite in the outer region .
this ` baseline ' field model is shown by the solid red line in fig .
[ fig : bf]a .
the values of @xmath61 are always less than those of the pure planetary dipole , shown by the dashed line in fig .
[ fig : bf]b , due to the radial stretching of the field lines associated with the current sheet .
the equatorial flux function is found by integration of eq .
[ eq : bf ] with the use of eq .
[ eq : bze ] , such that @xmath62+\frac{b_\circ}{(m-2)}\left(\frac{r_p}{\rho_e}\right)^{m-2 } , \label{eq : fe}\end{aligned}\ ] ] where @xmath37 nt @xmath38 is the value at infinity , and @xmath63 is the incomplete gamma function .
this flux function is shown by the solid red line in fig .
[ fig : bf]b , and is typically factors of @xmath05 greater than the corresponding dipole flux function shown by the long - dashed line .
we now consider how the above field model is modified by taking different values for the internal field strength and sub - solar magnetopause distance .
as summarised by @xcite , there are a number of models which estimate the planetary field strength given the planetary rotation rate , the radius of the dynamo region , and the mass density and conductivity within the dynamo region , although recently @xcite used an approach which was independent of the planetary rotation rate . here
we will assume for simplicity jovian values for the planetary radius , core conductivity and mass density , such that the planetary field strength is solely dependent on the planetary rotation rate @xmath64 .
the planetary magnetic moment thus scales as @xmath65 , where @xmath66 @xcite , @xmath67 @xcite , or @xmath68 @xcite .
we thus take @xmath69 as representative of these choices , such that @xmath70 and @xmath71 where @xmath72 is jupiter s planetary rotation period given by @xmath73 , and @xmath74 and @xmath75 are given by eqs .
[ eq : bze ] and [ eq : fe ] , respectively .
+ turning to the sub - solar magnetopause distance , this is expected to be dependent on the planet s orbital distance from the star via the stellar wind dynamic pressure @xmath76 , as discussed below , and on the internal field strength . for the former ,
we employ the method of @xcite , who studied the effect on the current system of changes to the magnetospheric size due to solar wind - induced compressions and expansions of jupiter s magnetosphere .
we note , however , that as well as considering the middle magnetosphere current sheet region described by the above field model , those authors also incorporated ` outer magnetosphere ' and open field line regions which are not included in this study since a significant flow shear at the open - closed field line boundary is not expected at planets with high ionospheric conductance , as considered here @xcite .
following @xcite , we set the reference boundary of the ` baseline ' field model to be at 85 @xmath58 , as shown by the red lines in fig . [
fig : bf ] , and modify the magnetic field model for smaller values of @xmath40 , by balancing the magnetic flux lost due to the closer outer boundary with the addition of the equivalent flux via a uniform southward field @xmath77 throughout the system , such that @xmath78 where @xmath77 thus represents the effect of the magnetopause currents as seen inside the magnetosphere .
the flux function @xmath47 is then equivalently modified by the addition of a term @xmath79 given by @xmath80 such that an identical amount of flux ( @xmath81 nt @xmath38 ) is contained within the boundary of the model for any value of @xmath40 smaller than 85 @xmath6 , and thus the ionospheric location of the outer boundary is kept fixed at @xmath82 . for values of @xmath40
larger than 85 @xmath58 , we simply use @xmath83 and @xmath84 , since beyond @xmath0130 @xmath58 the magnitude of @xmath77 becomes larger than the more distant values of @xmath61 . simply using @xmath83 and @xmath84 to larger radial distances
does result in a modest increase in the amount of flux contained within the model , thus modifying the ionospheric location of the outer boundary , but it is obvious from the difference between the lowest values of the red lines and the horizontal dotted lines in figs .
[ fig : bf]b and c that this increase is very small , and the change in location of the open - closed field line boundary is less than @xmath85 , i.e. less than observed changes in the location of jupiter s main auroral oval @xcite . the magnetic field and flux values for these modified field structures are shown for magnetopause standoff distances of 20 , 50 , 85 , and @xmath36 @xmath58 by the blue , green , red , and black lines in fig .
[ fig : bf ] , respectively .
+ we now discuss how the magnetopause standoff distance @xmath40 is expected to depend on planetary orbital distance @xmath86 .
the position of the magnetopause is dependent on the stellar wind dynamic pressure @xmath87 , where @xmath88 and @xmath89 are the stellar wind velocity and density , respectively .
these parameters are related to the mass loss rate of the star , which is measurable by detecting ly-@xmath90 absorption from the collision of the stellar wind with the surrounding interstellar medium @xcite .
these authors showed that stars more active than the sun ( which may be most important for internally - generated exoplanetary radio emission , as discussed below ) exhibit mass loss rates ranging between @xmath00.01 - 100 times the solar value , and conjectured that although younger , more active stars are expected to exhibit greater mass loss rates , the high magnetic field strengths of the most active stars may in fact act to inhibit stellar wind outflow . in view of these results , we employ solar wind velocity and density values as reasonably representative of the large observed range of stellar mass loss rates for active stars .
we thus assume that the stellar wind velocity @xmath88 is constant beyond 10 stellar radii at @xmath91 , a value typical of those observed at both @xmath01 au @xcite and @xmath05 au @xcite .
we then assume that the stellar wind density @xmath89 falls inversely with the square of the distance from the star ( thus neglecting plasma sources in the interplanetary medium such as interstellar pick - up ions which become increasingly significant toward the outer regions of the heliosphere ( e.g. * ? ? ?
* ; * ? ? ?
the observed values of the solar wind density actually vary significantly over a solar rotation period due to the formation of corotating interaction regions in the interplanetary medium , and at jupiter s orbit takes values between @xmath92 @xcite , with typical values being of order @xmath93 .
we thus employ the representative value of @xmath93 at 5.2 au , such that the stellar wind dynamic pressure varies with orbital distance as @xmath94 in the regime where the interplanetary magnetic field ( imf ) strength is negligible with respect to the planetary field , stress balance between the dynamic pressure of the shocked stellar wind plasma in the magnetosheath adjacent to the magnetopause and the magnetic pressure of the compressed planetary vacuum dipole yields @xmath95 where @xmath96 for a high - mach stellar wind @xcite and @xmath97 represents the factor by which the magnetopause field is enhanced by magnetopause currents , given by @xmath98 for a sub - solar boundary of realistic shape , i.e. lying between the values of 2 and 3 appropriate to planar and spherical boundaries , respectively ( e.g. * ? ? ?
noting , however , that the precise values of @xmath99 and @xmath97 are in practise unimportant due to the @xmath100 exponent in eq .
[ eq : rmp ] , this relation is approximated by @xmath101 similar relations to which have been used previously by various authors discussing exoplanetary magnetospheres ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , it is well known that this formula underestimates the size of jupiter s magnetosphere by a factor of @xmath02 , since jupiter s magnetopause is inflated to @xmath045 - 100 @xmath6 , with a mean of @xmath075 @xmath6 , by the presence of internal plasma through the combined actions of thermal pressure and centrifugal force @xcite .
therefore , @xcite presented an empirical relation for the variation of jupiter s sub - solar magnetopause standoff distance @xmath40 with @xmath76 , given by @xmath102 and ( right axis ) stellar wind dynamic pressure versus orbital radius @xmath86 computed using eq .
[ eq : psw ] and the huddleston et al .
( 1998 ) empirical relation for jupiter given by eq .
[ eq : hud ] ( solid line ) and the vacuum dipole relation given by eq .
[ eq : rmpdip ] ( dashed line ) , both for jovian values of the planetary magnetic moment .
also shown by the vertical dotted line at 5.2 au is the range of magnetopause distances observed for jupiter , thus the range over which the huddleston et al .
( 1998 ) relation was defined , i.e. 45 - 100 @xmath6 , and the horizontal dotted lines show the orbital radii to which this range corresponds in the present study , i.e. @xmath01.7 - 10 au .
, width=317 ] exoplanets with active moons would similarly be expected to contain significant internal plasma , and here we thus employ a modification of eq . [
eq : hud ] , taking into account the dependence on the internal field strength in eq .
[ eq : rmpdip ] , to estimate the values of @xmath40 , i.e. @xmath103 we show in fig .
[ fig : rmp ] the variation of sub - solar magnetopause distance @xmath40 and stellar wind dynamic pressure @xmath76 with orbital radius @xmath86 computed using eq .
[ eq : psw ] and the @xcite empirical relation for jupiter given by eq .
[ eq : hud ] ( solid line ) , along with the vacuum dipole relation given by eq .
[ eq : rmpdip ] ( dashed line ) .
the vertical dotted line at 5.2 au shows the range of sub - solar magnetopause distances observed for jupiter , i.e. @xmath045 - 100 @xmath6 corresponding to a solar wind dynamic pressure range of @xmath00.01 - 0.32 npa , hence representing the range over which the empirical @xcite relation is known to be valid .
the horizontal dotted lines then show the orbital radii to which this dynamic pressure range corresponds in the present study through eq .
[ eq : psw ] , i.e. @xmath01.7 - 10 au , although we acknowledge that intrinsic variations in the stellar wind dynamic pressure will actually induce a range in the sub - solar magnetopause standoff distances of an exoplanet at a given orbital distance , as observed at all solar system magnetospheres , an effect which we do not consider here . the validity of an empirical relation such as eq .
[ eq : hud ] beyond the observed values is , of course , not known
. however , the @xcite relation encapsulates in a relatively simple manner the response of a rotationally - driven , plasma - filled magnetosphere to changes in solar wind dynamic pressure , and produces results consistent with the 3d mhd simulations of jupiter s magnetosphere by @xcite and , more crucially , with in situ observations of the boundary location . on the other hand , the vacuum dipole relation given by eq .
[ eq : rmpdip ] represents an approximation at odds with a key property of the systems studied here , i.e. the radial distension of the field through the actions of the centrifugal force and internal plasma pressure , and does not provide results consistent with in situ observations of the jovian magnetosphere for any solar wind dynamic pressure value .
thus , in the absence of any information about the size of extra - solar jovian - type magnetospheres , we use the expression validated by observations of jupiter as a reasonable estimation , even though this sometimes involves extrapolation beyond the originally - observed range .
we next consider the angular velocity profiles arising from the application of newton s second law to a radially - outward steady flow of plasma from the torus source . as derived previously by @xcite and @xcite , and discussed further by @xcite and @xcite , the equatorial plasma angular velocity @xmath104 obeys the ` hill - pontius ' differential equation @xmath105 the left - hand side ( lhs ) of eq .
[ eq : hp ] represents the radial gradient of the plasma angular momentum per unit mass , whilst the right - hand side ( rhs ) is the ionospheric torque on the equatorial plasma per unit mass .
we also note that the effective pedersen conductance @xmath9 is reduced from the true value @xmath106 by @xmath107 , where the parameter @xmath108 is related to the reduction of the angular velocity of the neutral atmosphere ( @xmath109 ) from rigid corotation ( @xmath64 ) via ` slippage ' @xcite , such that @xmath110 .
the value of @xmath108 is somewhat uncertain , so in common with previous works we take @xmath111 .
+ we now discuss the equations which describe the resulting magnetosphere - ionosphere coupling currents .
first , taking the ionospheric field to be vertical and equal to @xmath112 in strength ( an approximation valid to within @xmath05% in our region of interest for planets with near - dipolar ionospheric fields @xcite ) , the equatorward - directed height - integrated pedersen current @xmath113 is given by @xmath114 current continuity and the assumption of north - south symmetry then yields for the equatorial radial current integrated across the width of the current sheet @xmath115 @xmath116 from eqs .
[ eq : ip ] , [ eq : irhoip ] and [ eq : fi ] , and recalling that @xmath48 we have @xmath117 such that the total radial current integrated in azimuth @xmath118 is @xmath119 equal to twice the azimuth - integrated pedersen current @xmath120 flowing in each hemisphere . the field - aligned current density , e.g. , northward out of the northern surface of the equatorial current sheet is then given , through current continuity , by the divergence of the radial current , such that @xmath121 where the factor of a half arises from the assumption of hemispheric symmetry , i.e. an equal and opposite current flows out of the southern face of the current sheet .
noting that the quantity @xmath122 is constant along a field line , such that @xmath123 , we find the following expression for the field - aligned current density at the top of the ionosphere @xmath124 with these currents in mind , we now consider the approximations appropriate to strongly - irradiated jupiter - like exoplanets .
equation [ eq : hp ] is a first order , linear , inhomogeneous ordinary differential equation , which can be solved analytically for power law magnetic field models , such as a dipole field @xcite or the ` kk ' field model valid throughout the majority of jupiter s middle magnetosphere @xcite , and numerically for other field models .
the solutions have the property that they are dependent on the quotient @xmath125 , as shown in fig .
[ fig : omega]a for roughly jovian values of the ionospheric pedersen conductance @xmath9 and equatorial plasma mass outflow rate @xmath10 .
specifically , the typical scale over which the plasma falls from rigid corotation is called the ` hill distance ' @xmath126 after t. w. hill , who first derived this scale length for a dipole magnetic field ( not to be confused with the radius of the hill sphere within which a body s gravitational field is dominant ) .
the scale length was modified by @xcite for the case of a magnetic field that varies as a power law with arbitrary exponent @xmath127 , and for which the field lines map from the equatorial plane to a narrow band in the ionosphere .
this is appropriate for the stretched current sheet magnetic field of jupiter s middle magnetosphere , as can be appreciated from the ionospheric mapping shown in fig .
[ fig : bf]c , in which between 20 - 150 @xmath6 the current sheet field models indicated by the solid lines map between @xmath015 - 17 , whereas the dipole field indicated by the dashed line maps to a much broader region between @xmath05 - 13 .
the current sheet scale length is given by ( solid lines ) and the high - conductance approximation given by eq .
[ eq : omegas ] ( dashed lines ) .
parameters shown are ( a ) the equatorial plasma angular velocity , ( b ) the azimuthally - integrated equatorial radial current , and ( c ) the field - aligned current density .
three results are shown for @xmath128 , with @xmath10 set to 1000 @xmath28 in panels b and c. note only one dashed line is shown in panels b and c since the high - conductance approximations ( given by eqs .
[ eq : totirhos ] and [ eq : jparis ] for the radial and field - aligned currents , respectively ) are independent of @xmath9 .
the tick marks show the limit of validity of the approximation as suggested by nichols % cowley ( 2003 ) .
, width=317 ] @xmath129 where @xmath130 is the value of the flux function at the narrow ionospheric band , taken here to be equal to @xmath131 , shown by the dotted horizontal line in fig .
[ fig : bf ] .
thus , for a given equatorial distance , the plasma angular velocity increases for increasing values of @xmath125 .
@xcite therefore derived analytic approximations to the solution appropriate to both the inner region , where the plasma near - rigidly corotates , and the outer region , where the plasma can be considered to be either free of ionospheric torque or stagnant .
the inner region approximation for the equatorial plasma angular velocity , which @xcite suggested is valid out to a distance @xmath132 of @xmath133 is the leading term in the series solution of eq .
[ eq : hp ] , taking @xmath134 as the formal expansion parameter and @xmath135 at @xmath136 ( see @xcite for further details ) , such that @xmath137 as shown by the dashed lines in fig .
[ fig : omega]a . substitution of eq .
[ eq : omegas ] into eqs .
[ eq : totip ] and [ eq : jpari ] thus yields @xmath138 and @xmath139 for the total radial current and ionospheric field - aligned current density , respectively .
these limiting currents , which are dependent only on @xmath10 and not @xmath9 , are those required to maintain near - rigid corotation throughout the magnetosphere , and can thus be viewed alternatively as a high-@xmath9 approximation for a given value of @xmath10 .
+ the currents @xmath140 and @xmath141 derived from the angular velocity profiles shown in fig . [
fig : omega]a are shown by the solid lines in figs . [ fig : omega]b and c , respectively , for roughly jovian field and current system parameters , i.e. @xmath10=1000 @xmath28 and @xmath9=0.2 , 1 , and 5 mho .
it is apparent that the current magnitudes increase with increasing @xmath9 , asymptoting toward the limiting currents shown by the dashed lines .
thus , the high - conductance approximation represents the condition under which maximum currents , and therefore maximum output radio power , are generated .
the minimum conductance required such that the high - conductance approximation is valid throughout the magnetosphere is found by setting @xmath142 from eq .
[ eq : slim ] , such that @xmath143 which , upon substitution of eq . [ eq : rhoh ] and rearranging for @xmath9 , yields @xmath144 since we are interested in current systems powerful enough to be detected across interstellar distances , as possibly possessed by jupiter - like exoplanets strongly irradiated by their parent stars , such a high - conductance is clearly appropriate and will be used as the basis for the results which follow . + before considering the precipitating electron energy flux and associated radio power output , it is appropriate here to discuss the region of validity of the high - conductance approximation by estimating how the ionospheric pedersen conductance varies with planetary orbital distance @xmath86 , and the type of star .
we use a model for the pedersen conductance which comprises two components , i.e. conductance produced by solar illumination , and auroral particle precipitation . for the former , in the absence of more complex ionospheric models for jupiter - like exoplanets we assume that the pedersen conductance is , like the earth s , linearly proportional to the ionospheric electron number density , a quantity proportional to the square root of the electron production rate , which is in turn linearly proportional to the solar extreme ultraviolet flux @xcite , although we note that actually the earth s upper atmosphere is ionised by solar radiation from x - ray to fuv wavelengths ( a range denoted by the term ` xuv ' ) .
thus , we have for the pedersen conductance generated by stellar illumination @xmath145 where @xmath146 and @xmath147 are the xuv luminosities of the exoplanet s parent star and the sun , respectively , and where the constant of proportionality is set such that @xmath148 mho at 5.2 au for solar xuv luminosity , consistent with estimates for jupiter ( e.g. * ? ? ?
the pedersen conductance is therefore expected to be higher for stars with higher xuv luminosities , such as more active stars .
for example , the @xmath0170 nm flux from f2v - type stars is two orders of magnitude larger than solar - type g2v stars @xcite , since g2v type stars are sub - luminous in the fuv by a factor of ten or more due to absorption in this waveband @xcite .
luminosity at x - ray - euv wavelengths is mainly non - thermal coronal emission , and surveys indicate that most coronal sources in the euv are active stars with coronal x - ray and euv luminosities several orders of magnitude greater than that of the sun @xcite .
stellar age is an important factor which determines stellar x - ray - euv luminosity @xcite , with younger main sequence stars typically exhibiting higher x - ray - euv emissions . in the results which follow ,
we will thus consider stellar xuv luminosities from 1 to @xmath149 , a range consistent with values observed by @xcite for solar - type stars , and thus could be conservative with respect to other spectral types .
+ as well as stellar illumination , pedersen conductance is expected to be generated by auroral electron precipitation .
computations by @xcite indicated that up to @xmath08 mho is generated by auroral precipitation at jupiter , and @xcite used these results to compute the expected variation of conductance with field - aligned current density assuming a representative primary and secondary precipitating electron spectrum . in their model the conductance plateaus at @xmath03 mho at @xmath00.5 @xmath150 , since at higher currents the primary electrons precipitate lower than the pedersen conducting layer and the conductance is generated by the lower energy secondaries . a full analysis of the effect of precipitation induced conductance requires that eq .
[ eq : hp ] and eq .
[ eq : jpari ] are solved simultaneously , as in the computations of @xcite , but here we simply assert that @xmath141 is everywhere greater than 0.5 @xmath150 , as is generally the case for the powerful exoplanetary systems we consider below , such that we assume a constant auroral contribution to the conductance of 3 mho , such that @xmath151 the maximum orbital distance @xmath152 for which the high - conductance approximation is valid is thus obtained simply by setting @xmath153 given by eq .
[ eq : spslim ] equal to @xmath9 given by eq .
[ eq : sps ] and solving for @xmath86 , numerically in practice .
this then limits the maximum power output for a given system , since at larger distances the high - conductance approximation is not valid and the magnetosphere will not rigidly corotate , and , as will be seen below , at smaller distances the power required to maintain the more compressed magnetosphere rigidly rotating is reduced .
normalised to the limiting current @xmath154 produced by field - aligned voltage @xmath155 normalised to the unaccelerated population thermal energy @xmath156 , for the three current - voltage relations given by eqs .
[ eq : jfk ] for @xmath157 , appropriate for an accelerator height of @xmath02.5 @xmath6 ( dashed line ) , [ eq : jkt ] ( solid line ) , and [ eq : jrk ] ( dotted line ) .
`` rel . '' means `` relativistic '' , `` non - rel . ''
means `` non - relativistic '' , and @xmath158 is the ratio of the ionospheric magnetic field strength to the field strength at the top of the voltage drop .
, width=317 ] we now briefly discuss the precipitating electron energy flux and associated radio power output that results from the field - aligned current given by eq .
[ eq : jparis ] .
the maximum field - aligned current that can be carried by unaccelerated precipitating magnetospheric electrons is @xmath159 with a corresponding precipitating energy flux of @xmath160 where @xmath161 is the elementary charge , @xmath162 is the magnetospheric electron number density , @xmath156 is the electron thermal energy , and @xmath163 is the electron mass . in the absence of detailed models or empirical values for exoplanetary magnetospheric plasmas
we take , in common with previous works , jovian values for the plasma parameters outside the current sheet , i.e. @xmath164 and @xmath165 @xcite , such that @xmath166 .
a significant field - aligned voltage is thus required to drive field - aligned currents greater than this value .
the current - voltage relation was first derived by @xcite and is given , for an arbitrary value of the ratio @xmath158 of the ionospheric magnetic field strength to the field strength at the top of the voltage drop , by @xmath167 where @xmath155 is the field - aligned voltage . in the limit that the field - aligned voltage region is located at high altitude ( typically @xmath168 up the field lines at jupiter ) , the ratio @xmath158 becomes large ( e.g. @xmath169 at @xmath170 ) such that eq .
[ eq : jfk ] reduces to @xmath171 the relation used by @xcite . on the other hand
, @xcite suggested that the accelerator would be fixed at low altitude ( 2 - 3 @xmath6 ) , such that they employed eq .
[ eq : jfk ] with @xmath157 . however , the @xcite relations given by eqs .
[ eq : jfk ] and [ eq : jkt ] assume the accelerating voltage is small such that the precipitating particles are accelerated to non - relativistic speeds , an assumption which is not valid when the field - aligned voltage approaches or exceeds the electron rest energy ( @xmath0511 kev ) , as would be the case for powerful exoplanetary systems , as shown below .
this restriction was removed by @xcite , who showed that for a relativistic accelerated population the current - voltage relation becomes , in the high @xmath158 limit , @xmath172}\;\ ; , \label{eq : jrk}\ ] ] where @xmath173 is the speed of light .
the current - voltage relations given by eqs .
[ eq : jfk ] , [ eq : jkt ] , and [ eq : jrk ] are shown in fig .
[ fig : jp ] , in which it is apparent that the relativistic and fixed - height accelerator cases produce competing deviations from the simple relation given by eq .
[ eq : jkt ] . in light of the uncertainty surrounding the exact plasma parameters at jupiter - like exoplanets , we thus employ the limiting @xcite relation given by eq .
[ eq : jkt ] and shown by the solid line in fig .
[ fig : jp ] , as providing a reasonable representation of the available choices for the current - voltage relation , and we simply note that this relation has been successfully used previously to compute auroral brightness and precipitating electron energy values for jupiter that are consistent with observations ( e.g. * ? ? ?
* ) . + the precipitating energy flux associated with the electrons accelerated through voltage @xmath155 as given by eq .
[ eq : jkt ] was shown by @xcite to be @xmath174\simeq \frac{\ensuremath{e_{f\circ}}}{2}\left(\frac{\ensuremath{j_{\|i}}}{\ensuremath{j_{\|i\circ}}}\right)^2\;\ ; , \label{eq : ef}\ ] ] where the approximation on the rhs is valid where @xmath175 , a condition that is valid essentially everywhere in the models we investigate .
the total precipitating power for each hemisphere @xmath176 is thus obtained by integration of @xmath177 over the hemisphere , such that @xmath178 although the exact details of the mechanism by which radio frequency waves are generated from unstable auroral electron distributions remain to be determined , the most likely candidate is the electron cyclotron maser instability ( e.g. * ? ? ?
* ) , which has a generation efficiency of @xmath01% @xcite , such that we take the total emitted radio power to be given by @xmath179 normalised to jupiter s value @xmath52 , and ( on the right axis ) the estimated radio emission bandwidth versus planetary angular velocity over the range @xmath180 , computed from eqs .
[ eq : bzeomega ] and [ eq : delnu ] , respectively .
, width=317 ] we finally assume that the radio emission is beamed into solid angles of 1.6 sr , in conformity with observations of the jovian hom and dam emissions @xcite .
the effect of beaming is to increase the flux at a given distance but decrease the chance of detection , and here we assume that the beam from only one hemisphere is observable at any one time .
assuming such beamed emission over the bandwidth @xmath181 , the spectral flux density at a given distance @xmath182 from the source is @xmath183 where the bandwidth @xmath181 is taken to be equal to the electron cyclotron frequency in the polar ionosphere , i.e. @xmath184 again assuming a uniform polar ionospheric field strength equal to @xmath112 .
we show in fig .
[ fig : obf ] the field strengths and resulting estimated radio bandwidths for the range of planetary angular velocities considered in this paper ( see below ) .
( blue , green , red , and black lines respectively ) , and with @xmath185 and @xmath186 .
the labels in panel ( a ) correspond to the values of @xmath86 which result in magnetospheres of these sizes , according to eqs . [ eq : psw ] and [ eq : hud ] .
parameters shown are ( a ) the azimuthally - integrated ionospheric pedersen current @xmath187 in a as given by eq .
[ eq : totirhos ] , ( b ) the field - aligned current density at the top of the ionosphere @xmath141 in @xmath188 given by eq .
[ eq : jparis ] , ( c ) the minimum field - aligned voltage @xmath189 in v required to drive the current in panel ( b ) as given by eq .
[ eq : jkt ] , and ( d ) the precipitating electron energy flux @xmath177 in @xmath190 as given by eq .
[ eq : ef ] .
the labels in panel ( d ) show the radio power obtained by integrating the energy fluxes using eqs .
[ eq : pe ] and [ eq : pr ] .
note there are no blue lines in panels ( c ) and ( d ) since for this case the current density in panel ( b ) is everywhere below the minimum current @xmath154 for which field - aligned voltages are required ( shown by the horizontal dotted line in panel ( b ) ) .
, width=317 ] we now consider the application of the analysis presented in sec .
[ sec : theory ] to various configurations of exoplanets .
we first show in fig .
[ fig : is ] the effect on the current system of taking different values of @xmath40 , whilst maintaining jovian values for the planetary angular velocity @xmath185 ( and thus @xmath191 ) and mass plasma outflow rate @xmath192 , thus simulating transplanting the jovian system to different orbital distances , assuming that the xuv flux is at all distances sufficiently high such that high - conductance approximation holds everywhere .
it is apparent from fig .
[ fig : is ] that for values of @xmath40 less than the reference value of 85 @xmath6 the current amplitudes are , as expected from e.g. eq .
[ eq : totirhos ] , reduced owing to the elevated magnetic field strength @xmath61 , which we recall is increased via eq .
[ eq : deltab ] for a smaller magnetosphere due to flux conservation . for these model parameters , very compressed magnetospheres with @xmath40 smaller than @xmath020 @xmath6 ,
are such that the field - aligned current density is everywhere less than @xmath154 , hence no field - aligned voltages are required and there is no bright discrete auroral emission . hence , even if we neglect tidal - locking which would in reality act to prevent the fast rotation of such planets ( discussed further below ) , very close - orbiting ( @xmath193 au ) ` hot jupiter ' systems jovian - like in all respects apart from the orbital distance would exhibit no main auroral oval auroras . on the other hand , similar magnetospheres larger than 85 @xmath6 exhibit common current profiles , extending to further distances and thus modestly further toward the pole , depending on the size of the magnetosphere . the monotonically - decreasing field strength with increasing equatorial radius results in an associated increasing of the current magnitudes toward the pole , and thus precipitating electron energy flux , toward the outer boundary of the model .
increasing the size of the magnetosphere thus increases the emitted radio power , as shown in fig .
[ fig : is]d . + and @xmath10,i.e . ( a ) ( @xmath194 , @xmath195 ) , ( b ) ( @xmath196 , @xmath195 ) , ( c ) ( @xmath194 , @xmath197 ) , and ( d ) ( @xmath196 , @xmath197 ) .
each section consists of two panels .
the top shows the limiting conductance given by eq .
[ eq : spslim ] ( solid line ) and the estimated pedersen conductance in mho given by eq .
[ eq : sps ] for four values of the stellar xuv luminosity , i.e. @xmath198 = 1,10,100 , and 1000 ( dashed lines from bottom to top , respectively ) .
the bottom shows with the left axis the radio power in w emitted , obtained by integrating the precipitating electron energy fluxes using eqs .
[ eq : pe ] and [ eq : pr ] , and with the right axis the maximum distance in parsecs at which sources emitting such powers would be detectable , assuming a detection threshold of 1 mjy .
all parameters are plotted versus orbital distance @xmath86 .
the vertical dotted lines in each section indicate where the estimated pedersen conductances equal the limiting conductance , i.e. show the maximum distances @xmath152 to which the high - conductance approximation is valid , and thus the maximum radio powers for these parameters .
the lines in the bottom of each panel are shown dotted beyond the maximum values of @xmath152 for @xmath199 , indicating that this region corresponds to xuv luminosities above the upper limit for a young solar - type star as indicated by ribas et al .
, width=317 ] having briefly examined the effect of magnetospheric size , we now extend the the analysis by considering the parameters @xmath10 and @xmath64 .
there are presently no published theoretical limitations on the maximum plasma mass outflow rate from exoplanetary moons , which presumably would vary with the number of volcanic moons in a given system as well as the orbital and geological parameters of each satellite and planet .
the theory considered in this paper is valid for all non - zero values of the plasma mass outflow rate , and it is worth noting that at jupiter the ionosphere and solar wind also act as plasma sources with rates of @xmath200 and @xmath201 @xmath28 , respectively @xcite , such that in practice the mass outflow rate will probably be non - zero in most systems , even those devoid of active moons .
it therefore seems reasonable to compute the radio power emitted for a range of @xmath10 values within an order of magnitude of the canonical jovian figure of 1000 @xmath28 .
the rotation rate of a planet is governed by its initial angular velocity and subsequent tidal dissipation .
the origin of the former is poorly understood , but candidates include the relative motion of planetesimals during protoplanetary accretion and , for the gas giants specifically , hydrodynamic flows during accretion of hydrogen and helium @xcite .
a crude estimation of the maximum angular velocity allowed by centrifugal stability of a planet is @xmath202 where @xmath203 is the gravitational constant equal to @xmath204 , and which for a planet of jupiter s mass and radius yields @xmath205 .
however , jupiter lies roughly on the boundary between the regime of bodies supported by coulomb pressure , for which @xmath206 , and degeneracy pressure , for which @xmath207 , such that jupiter s radius is near maximal for bodies of solar composition .
the net effect is that all bodies from solar composition giant planets through brown dwarfs to the very lowest mass stars are expected to have radii similar to jupiter , and it is thus possible for planets with mass of , e.g. @xmath208 to have angular velocities up to @xmath011 @xmath72 .
the time scale @xmath209 required to de - spin a planet by tidal dissipation is given by @xmath210 where @xmath211 is the planet s tidal dissipation factor typically given by @xmath212 for jovian planets , @xmath213 is the keplerian angular velocity of the planet , and where we note that factors of order unity have been omitted @xcite . for planets of jovian parameters orbiting solar - type stars at 0.1 , 1 , and
100 au , this expression yields tidal synchronisation time scales of @xmath214 , @xmath215 , and @xmath216 y , respectively , while for planets more susceptible to tidal locking e.g. those with @xmath217 and @xmath218 , the synchronisation time scales are modified to @xmath219 , @xmath220 , and @xmath221 y. thus , close - orbiting ` hot jupiters ' are expected to rapidly become tidally locked , but for planets orbiting beyond 1 au , tidal effects will be negligible over time scales of e.g. the present age of the solar system of @xmath222 y , such that the planetary angular velocity may not be much reduced from its initial value at formation .
+ with the above considerations in mind , we now examine the system using four spot combinations of the parameters @xmath10 and @xmath72 , i.e. ( @xmath194 , @xmath195 ) , ( @xmath196 , @xmath195 ) , ( @xmath194 , @xmath197 ) , and ( @xmath196 , @xmath197 ) .
results are shown in fig .
[ fig : sigpow ] , which is divided into four sections of two panels corresponding to the above pairs of parameters .
the solid line in the top panel of each section shows the limiting conductance given by eq .
[ eq : spslim ] .
the dependence of the limiting conductance on the parameters @xmath86 , @xmath10 , and @xmath64 , obtained from eqs .
[ eq : bze ] , [ eq : fe ] , [ eq : psw ] , [ eq : hudmod ] , and [ eq : spslim ] , is approximately @xmath223 , such that it increases slightly faster than linearly with distance in each panel , and compared to the values shown fig .
[ fig : sigpow]a , those in figs .
[ fig : sigpow]b , c , and d are thus multiplied by factors of 10 , @xmath00.41 , and @xmath04.1 , respectively .
the estimated pedersen conductance is also shown by the dashed lines in each panel for stellar xuv luminosities of @xmath198 = 1,10,100 , and 1000 ( from bottom to top , respectively ) . despite these xuv luminosities
being orders of magnitude larger than the solar value , since @xmath224 the conductance values are multiplied from the solar values by factors of only 1 , 3.2 , 10 , and 32 , respectively .
however , it is obvious that the increased conductance generated by the higher xuv luminosities greatly increases the maximum orbital distance @xmath152 at which the high - conductance approximation holds , as shown in each section by the four vertical dotted lines , for example in section ( a ) at @xmath04.4 , 5.3 , 7.3 , and 10.9 au corresponding to conductances of @xmath03.6 , 4.6 , 6.6 , and 10.6 mho , respectively . as shown in figs .
[ fig : sigpow]b , c , and d , these values of @xmath152 are then modified from , e.g. , the high xuv luminosity case of @xmath010.9 au to @xmath03.5 , 17.7 , and 5.4 au , due to increased @xmath10 , @xmath64 , and both @xmath10 and @xmath64 , respectively .
+ the effects of these changes on the maximum radio power are shown in the bottom panels of each section of fig .
[ fig : sigpow ] .
the left hand axes indicate the emitted powers , obtained by integrating the precipitating electron energy fluxes using eqs .
[ eq : pe ] and [ eq : pr ] , whilst the right hand axes show the distance @xmath182 in parsecs obtained from eqs .
[ eq : fr ] and [ eq : delnu ] with these powers and an assumed spectral flux density of 1 mjy .
note that the right hand axes in figs .
[ fig : sigpow]c and d are different to those in figs .
[ fig : sigpow]a and b , since the spectral flux density is dependent on @xmath225 , which is proportional to @xmath226 through @xmath51 . as also shown in fig .
[ fig : is ] , the radio power increases with orbital distance , with a slight discontinuity in the gradient where @xmath227 .
this discontinuity is simply an artefact introduced by the change in the behaviour of the equatorial magnetic field model at the reference boundary of the ` baseline ' magnetic field model . for @xmath228 , as the radius of the magnetosphere increases with increasing orbital distance @xmath86 , the equatorial magnetic field strength decreases due to flux conservation , such that the current intensities and power output increase as discussed above ( c.f .
the blue and green lines in fig .
[ fig : is ] ) . on the other hand , as the radius of the magnetosphere further increases such
that @xmath229 the current intensity at a given co - latitude remains the same , but extends increasingly toward the pole ( c.f .
the red and black lines in fig .
[ fig : is ] ) .
the latter has less of an effect that the former , such that the power output increases less quickly with @xmath86 after @xmath40 passes 85 @xmath58 . thus , for the values of @xmath10 and @xmath64 in fig .
[ fig : sigpow]a , the maximum radio powers @xmath230 for the four xuv luminosity cases are @xmath231 , @xmath232 , @xmath233 , and @xmath234 w , corresponding to maximum observable distances @xmath182 of @xmath00.4 , 0.7 , 1.5 , and 2.6 pc , respectively . however , again taking the high xuv luminosity case as an example , this power is modified as shown in figs .
[ fig : sigpow]b , c , and d to @xmath235 , @xmath236 , and @xmath237 w , corresponding to maximum observable distances of @xmath02.7 , 10 , and 20 pc , respectively .
thus , from these four spot combinations of @xmath10 and @xmath64 it is apparent that the maximum radio power @xmath230 is relatively insensitive to changes in @xmath10 , since the higher powers available at a given orbital distance @xmath86 for increased @xmath10 are compensated for by a decrease in the maximum distance to which the high - conductance approximation is valid @xmath152 . on the other hand , whilst changing @xmath64 does have competing effects on different components of the system , the overall behaviour is one in which increased @xmath64 results in increased power .
we further note that , although , as discussed above , the compressed vacuum dipole relation for the sub - solar magnetopause standoff distance is not strictly appropriate for the rotationally - driven , plasma - filled magnetospheres we consider here , the @xcite relation was derived using observations of jovian values over the range @xmath045 - 100 @xmath6 , and extrapolation beyond this range may lead to uncertainties .
we have thus also computed the above values using eq .
[ eq : rmpdip ] in place of eq .
[ eq : hudmod ] , and find that the results are qualitatively similar , except that the values of @xmath152 , @xmath230 , and @xmath182 are modified by factors of @xmath02 - 4 , @xmath00.1 - 0.4 , and @xmath00.3 - 0.6 , respectively . as noted by @xcite and @xcite , the effect of the stretching of the planetary field from a dipolar configuration into a magnetodisc structure is to amplify the upward field - aligned current density associated with the aurora and radio emissions by 1 - 2 orders of magnitude , such that the precipitating electron energy flux , a quantity proportional to the square of the field - aligned current density through eq .
[ eq : ef ] , is modified by 2 - 4 orders of magnitude .
using eq .
[ eq : rmpdip ] in place of eq .
[ eq : hudmod ] compresses the equatorial magnetic field to a more dipole - like form , such that the radio power available at a given orbital radius is significantly reduced .
however , the effect of a closer boundary is also to reduce the conductance required to maintain near - rigid corotation throughout the magnetosphere , such that the maximum power reduction is partly mitigated by increased orbital radius at which the maximum powers are available .
however , we again note that eq .
[ eq : rmpdip ] does not reproduce results consistent with observations of jupiter s magnetosphere , such that in the discussion that follows we employ results obtained using the @xcite relation . + available using different pairs of system parameters @xmath10 and @xmath64 for , in panels ( a)-(d ) respectively , the xuv luminosity cases @xmath198 = 1,10,100 , and 1000 . also shown by the black contours are the maximum distances @xmath182 in pc at which sources of these powers are observable , assuming a detection threshold of 1 mjy , and the white contours show the orbital distances @xmath152 in au at which these maximum powers are available .
labelled crosses refer to the spot combinations shown in fig .
[ fig : sigpow ] .
, width=317 ] we now extend the analysis to cover the full parameter space within the considerations discussed above .
we thus examine the behaviour of the system over @xmath238 and @xmath239 .
we should note that angular velocities above @xmath03 @xmath72 would require planets more massive than jupiter and may thus have larger dipole moments , although as discussed above for simplicity we do not take this into account here and will be the subject of future study .
results are shown in fig .
[ fig : grids ] for the four different xuv luminosity cases .
the colour indicates the maximum radio power @xmath230 available for each pair of parameter values , as shown in fig .
[ fig : sigpow ] by the powers at the locations of the vertical dotted lines , and the black contours show the maximum observable distances of sources emitting such radio powers , again assuming a spectral flux density threshold of 1 mjy .
note the kinks are an artefact of the model indicating the change of behaviour where @xmath240 as discussed above .
the white contours indicate the orbital distances @xmath152 at which these powers are available , which we note increase with decreasing @xmath10 and , to a lesser extent with increasing @xmath64 , as previously noted .
this plot confirms the overall behaviour apparent in fig .
[ fig : sigpow ] , i.e. that the radio power is essentially independent of @xmath10 , increases with @xmath72 , and increases with stellar xuv luminosity .
thus , for stars with solar xuv luminosity , planets with @xmath241 are required to produce radio emissions detectable from beyond @xmath010 pc , but this is reduced to e.g. @xmath242 for stars with @xmath198 = 1000 and @xmath243 . in all xuv luminosity cases , a significant number of parameter combinations within an order of magnitude of the jovian values are capable of producing emissions observable beyond 1 pc , in most cases requiring exoplanets orbiting at distances between @xmath01 and 50 au .
for the two higher xuv luminosity cases , parameter combinations within an order of magnitude of jupiter s could generate emissions detectable beyond @xmath050 pc .
in this paper we have provided the first consideration of magnetosphere - ionosphere coupling at jupiter - like exoplanets .
we have estimated the radio power emitted by such systems under the condition of near - rigid corotation throughout ( a condition which maximises the field - aligned currents and thus the radio power for a given magnetosphere ) , in order to examine the behaviour of the best candidates for detecting internally - generated radio emission with next generation radio telescopes such as lofar .
we have thus estimated for different stellar xuv luminosity cases the orbital distances within which the ionospheric pedersen conductance would be high enough to maintain near - rigid corotation , and we have then considered the magnitudes of the large - scale magnetosphere - ionosphere currents flowing within the systems , and the resulting radio powers , at such distances .
we have also examined the effects of two key system parameters , i.e. the planetary angular velocity and the plasma mass outflow rate .
+ the key results of the study can be summarised as follows : 1 .
the radio power emitted increases with increasing system size , and thus increases with orbital distance within the limit of validity of the high - conductance approximation .
2 . the limiting orbital distance , which defines the maximum radio power available for a given set of system parameters , increases with stellar xuv luminosity and planetary rotation rate , and decreases with magnetospheric plasma mass outflow rate .
the overall effect is that the radio power emitted increases with planetary rotation rate , but is essentially independent of plasma mass outflow rate since the higher powers available at a given orbital distance for increased plasma mass outflow rate are compensated for by a decrease in the maximum orbital distance to which the high - conductance approximation is valid .
4 . in all xuv luminosity cases studied , a significant number of parameter combinations within an order of magnitude of the jovian values are capable of producing emissions observable beyond 1 pc , in most cases requiring exoplanets orbiting at distances between @xmath01 and 50 au . for the higher xuv luminosity cases the observable distances for jovian mass planets can reach @xmath020 pc , and massive , rapidly rotating planets could be detectable beyond @xmath050 pc .
however , we should note here the limitations of the simple model used in this study .
first , we have not considered how the structure of the magnetic field changes with planetary angular velocity and plasma mass outflow rate .
as noted above , jupiter s magnetosphere is partly inflated by the centrifugal force of iogenic plasma @xcite , such that systems with higher angular velocity or plasma loading would be expected to be further inflated , although this would be somewhat mitigated by the associated increase of the planetary field strength in the case of the angular velocity . however , as discussed in section [ sec : results ] , the effect of the stretching of the planetary field from a dipolar configuration into a magnetodisc structure is to amplify the upward field - aligned current density associated with the aurora and radio emissions by 1 - 2 orders of magnitude , and therefore the power values derived here may be viewed as lower limits in this regard .
second , we note that the field - aligned voltages obtained by the current - voltage relation used in this study may underestimate the true voltages if the location of the accelerator is fixed at a few planetary radii up the the field lines , such that these power values might again be underestimates . in a related point
, we note that simply defining the plasma angular velocity as we have done ignores the effects of the significant field - aligned voltages , which act to de - couple the equatorial and ionospheric plasma angular velocities @xcite , an issue which should be studied in further works .
further , we note that we have not considered at all the stellar wind interaction mediated by magnetic reconnection at the dayside magnetopause , which at jupiter could be associated with the many variable and sometimes extremely bright polar auroras observed @xcite , and would thus sporadically increase the power output . nor
have we considered the effect of changing stellar wind dynamic pressure , which at jupiter is known to modulate the intensity of the uv and radio emissions by factors of @xmath03 @xcite , and again , although the details of how this affects jupiter s auroral and radio emissions remain to be fully determined , this effect may also act to increase the output powers from those computed here .
however , we also reiterate that our assumption of solar values for the stellar wind velocity and density may represent underestimates for stars more active than the sun , such as those much younger .
higher stellar wind dynamic pressure values would decrease the size of the magnetosphere for a given orbital distance , thus reducing the emitted radio power .
it is clear that the evolution of internally - generated radio emissions over the lifetime of a star should therefore be considered in future studies .
we should also recall that we have only considered here planetary magnetic fields with the same polarity as those of jupiter and saturn , i.e. with magnetic and spin axes co - aligned to first order , since this is the configuration that has been most studied for bodies in our solar system .
however , it is probable that only 50% of jupiter - like exoplanets exhibit this polarity .
the effect of reversed polarity is to reverse the direction of the current system , such that downward currents are replaced with upward currents and vice versa , such that the source population for the field - aligned currents in this case may be very different to that considered here ( see e.g. @xcite for a discussion of field - aligned currents induced at jupiter s dayside magnetopause ) .
the emissions from planets with the opposite polarity should be studied in future works .
+ finally , we note that the implication of these results is that the best candidates for detection of such internally - generated radio emissions are rapidly rotating jupiter - like exoplanets orbiting stars with high xuv luminosity at orbital distances beyond @xmath01 au .
this type of exoplanet has not previously been considered as potential detection candidates for next generation radio telescopes , but searching for such emissions such may offer a new method of detection of more distant - orbiting exoplanets less likely to be detected by those techniques which are biased toward close - orbiting ` hot jupiters ' . however
, dual detections with radio telescopes and conventional methods would best constrain the planetary parameters .
jdn was supported by stfc grant st / h002480/1 , and wishes to thank s. w. h. cowley , m. a. barstow , m. r. burleigh and g. a. wynn for constructive discussions during this study , and also thanks the referee for providing helpful comments on the manuscript .
clarke j. t. , grodent d. , cowley s. w. h. , bunce e. j. , zarka p. , connerney j. e. p. , satoh t. , 2004 , in jupiter .
the planet , satellites and magnetosphere , f. bagenal , t. e. dowling and w. b. mckinnon , ed .
, cambridge .
univ . press , cambridge , uk , pp .
639670 k. k. , kivelson m. g. , vasylinas v. m. , krupp n. , woch j. , lagg a. , mauk b. h. , kurth w. s. , 2004 , in jupiter . the planet , satellites and magnetosphere , f. bagenal , t. e. dowling and w. b. mckinnon , ed .
, cambridge .
press , pp .
593616 r. , sandahl i. , 1978 , in esa special publication , vol .
135 , european sounding rocket , balloon and related research , with emphasis on experiments at high latitudes , halvorsen t. , battrick b. , eds . , | in this paper we provide the first consideration of magnetosphere - ionosphere coupling at jupiter - like exoplanets with internal plasma sources such as volcanic moons .
we estimate the radio power emitted by such systems under the condition of near - rigid corotation throughout the closed magnetosphere , in order to examine the behaviour of the best candidates for detection with next generation radio telescopes .
we thus estimate for different stellar x - ray - uv ( xuv ) luminosity cases the orbital distances within which the ionospheric pedersen conductance would be high enough to maintain near - rigid corotation , and we then consider the magnitudes of the large - scale magnetosphere - ionosphere currents flowing within the systems , and the resulting radio powers , at such distances .
we also examine the effects of two key system parameters , i.e. the planetary angular velocity and the plasma mass outflow rate from sources internal to the magnetosphere . in all xuv luminosity cases studied , a significant number of parameter combinations within an order of magnitude of the jovian values are capable of producing emissions observable beyond 1 pc , in most cases requiring exoplanets orbiting at distances between @xmath01 and 50 au , and for the higher xuv luminosity cases these observable distances can reach beyond @xmath050 pc for massive , rapidly rotating planets .
the implication of these results is that the best candidates for detection of such internally - generated radio emissions are rapidly rotating jupiter - like exoplanets orbiting stars with high xuv luminosity at orbital distances beyond @xmath01 au , and searching for such emissions may offer a new method of detection of more distant - orbiting exoplanets .
[ firstpage ] planetary systems planets and satellites : aurorae , magnetic fields , detection . |
the virial theorem ( vt ) ( chandrasekhar & fermi 1953 ) provides a direct way of analyzing the energy balance of a bounded region in a flow , describing the effect of various forces either in driving changes in the structure of a dynamical system or in determining the character of its equilibrium .
the virial theorem can be cast in either eulerian or lagrangian form .
the latter applies to a fluid parcel following the flow , i.e. , the volume @xmath0 and its bounding surface @xmath1 will generally be time - dependent .
the eulerian version of the virial theorem ( evt ) applies to a fixed volume @xmath0 rather than a fixed mass ( e.g. , mckee & zweibel 1992 ) .
this is best suited for application to fixed - grid , eulerian numerical simulations .
however , when considering a large region of the ism , clouds constitute an ensemble in which each one is _ moving _ and _ morphing _ ( see vzquez - semadeni , this volume ) .
thus , for studying the evt we have two options : \(a ) we can consider all clouds in the simulation , or `` lab '' frame ( ballesteros - paredes & vazquez - semadeni 1997 ) .
however , in this case , the clouds bulk motions will appear as internal kinetic energy , and the energy budget of the volume defining the cloud is not exactly the energy budget of the cloud ; \(b ) we can consider a different frame for each cloud , so that the frame is moving with the cloud s center of mass velocity with respect to the lab frame , and has its origin at the cloud s center of mass . in the present work ,
we briefly discuss the choice of reference frame , and then present preliminary statistical results of virial balance in two three - dimensional simulations of mhd isothermal turbulence with self - gravity at a resolution of @xmath2 grid points , and with rms mach number 2.2 .
one simulation has a box size @xmath3 equal to the jeans length @xmath4 , and the other has @xmath5 .
the transition from the lab frame is not completely straightforward .
we first considered the possibility of staying in the lab frame , but taking a volume that moves with the velocity of the cloud s center of mass , although maintaining a fixed shape in time .
however , in this case , the form of the evt is altered , and several extra terms appear that have no simple interpretation . due to these difficulties ,
we finally have chosen to completely move to each cloud s frame , as described above . in this case
, the evt remains in its usual form , at the expense that plots for the cloud ensemble comparing the various terms in the evt contain data from many different frames
one for each cloud .
the evt is then @xmath6 where @xmath7 is the moment of the inertia of the cloud , @xmath8 is the thermal energy ; @xmath9 is the kinetic energy , @xmath10 is the surface pressure term , @xmath11 is the surface kinetic term ; @xmath12 is the magnetic energy and @xmath13 is the surface magnetic term , with @xmath14 being the maxwell stress tensor ; @xmath15 is the gravitational term ( _ not _ equal to the gravitational energy ) , and @xmath16 is the flux of moment of inertia through the surface of the cloud ( sums over repeated indices are assumed ) .
we define the clouds in the numerical simulations as connected sets of pixels with densities larger than an arbitrary threshold value . in order to measure only the contribution associated to the fluctuations , we substract from the velocity the bulk , mass - averaged velocity of the cloud , defined as @xmath17 , where @xmath18 is mass of the cloud , and substract from the position of each pixel the position of the cloud s center of mass .
we consider several density thresholds to enlarge the cloud ensemble , and consider that a cloud maintains its identity as long as it does not split into several components upon increasing the threshold .
we find the following preliminary results , consistent with those of ballesteros - paredes & vzquez - semadeni ( 1997 ) : the time derivative terms ( i.e. , the lhs and the last term in the rhs of equation [ 1 ] ) are dominant in the overall virial balance , being much larger than the remaining volume and surface terms ( fig.1 ) .
we refer to these terms as `` geometrical '' , since they correspond to the ( time derivatives of the ) mass distribution in the cloud and its flux through the cloud s boundary .
this implies that , far from being in quasi - hydrostatic equilibrium , the clouds are continually changing shape and `` morphing '' .
the surface terms , which are often neglected in virial - balance studies , have magnitudes comparable to those of the volumetric ones ( fig.2 ) . * 3 . *
however , the surface and the volume terms do not cancel out exactly , leaving a net contribution for shaping the clouds and balancing gravity .
in fact , we propose that the correct diagnostic for whether a cloud will undergo gravitational collapse is @xmath19 ; otherwise , the cloud is transient . indeed , in the simulation with @xmath20 , two clouds are observed to be collapsing by the final time .
these two clouds satisfy the above criterion ( fig.3 , left ) . note that comparing the absolute value of gravitational term vs. only the volume term would suggest that none of the clouds be collapsing ( fig.3 , right ) , contrary to what occurs in the simulation .
thus , we conclude that _ the correct diagnostic for determining gravitational binding must include the contribution of the surface terms , or else the gravitational term is underestimated . _ | we discuss the virial balance of all members of a cloud ensemble in numerical simulations of self - gravitating mhd turbulence .
we first discuss the choice of reference frame for evaluating the terms entering the virial theorem ( vt ) , concluding that the balance of each cloud should be measured in its own reference frame .
we then report preliminary results suggesting that a ) the clouds are far from virial _ equilibrium _
, with the `` geometric '' ( time derivative ) terms dominating the vt .
b ) the surface terms in the vt are as important as the volume ones , and tend to decrease the action of the latter .
c ) this implies that gravitational binding should be considered including the surface terms in the overall balance .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
at present , it is not straightforward to describe the transverse momentum ( @xmath1 ) distributions of electroweak bosons produced in hadron - hadron collisions . in the usual collinear approximation ,
the transverse momentum of the incoming partons is neglected and so , for the born level subprocesses @xmath8 ( where @xmath9 ) or @xmath10 , the transverse momentum of the final electroweak boson is zero .
therefore , initial - state qcd radiation is necessary to generate the @xmath1 distributions . both the leading order ( lo ) and next - to - leading order ( nlo ) differential cross sections diverge for @xmath11 , with terms proportional to @xmath12 appearing due to soft and collinear gluon emission , requiring resummation to achieve a finite @xmath1 distribution .
traditional calculations combine fixed - order perturbation theory at high @xmath1 with either analytic resummation or numerical parton shower formalisms at low @xmath1 , with some matching criterion to decide when to switch between the two . in addition
, a parameterisation is needed to account for non - perturbative effects at the lowest @xmath1 values .
analytic resummation can be performed either in the transverse momentum space ( see , for example , @xcite ) or in the fourier conjugate impact parameter space ( see , for example , @xcite ) .
an alternate description is provided in terms of _ unintegrated _ parton distribution functions ( updfs ) , where each incoming parton carries its own transverse momentum @xmath13 , so that the subprocesses @xmath8 and @xmath10 already generate the lo @xmath1 distributions in the so - called @xmath13-factorisation approach .- factorisation , see @xcite . ]
it has been shown in @xcite that updfs obtained from an approximate solution of the ccfm evolution equation @xcite embody the conventional soft gluon resummation formulae .
the updfs that we use are obtained from the familiar dglap - evolved pdfs determined from a global parton analysis of deep - inelastic and related hard - scattering data .
the transverse momentum of the parton is generated entirely in the _ last _ evolution step @xcite .
angular - ordering constraints are imposed which regulate the singularities arising from soft gluon emission , while the virtual terms in the dglap equation are resummed into sudakov form factors .
effects . ] in @xcite it was shown that it is necessary to extend the ` last - step ' formalism of @xcite to consider _
doubly_-unintegrated parton distribution functions ( dupdfs ) in order to preserve the exact kinematics .
it was demonstrated that the main features of conventional higher order calculations can be accounted for within a much simpler theoretical framework , named @xmath0-factorisation .
is the splitting fraction associated with the last evolution step , where a parton with ( light - cone ) momentum fraction @xmath14 splits to a final parton with ( light - cone ) momentum fraction @xmath15 .
the idea is an extension of the original ddt formula @xcite ; however , in comparison with @xcite we go beyond the double leading logarithmic approximation ( dlla ) and account for the precise kinematics of the two incoming partons , as well as the angular ordering of emitted gluons . ] strictly speaking , the integrated pdfs used as input to the last evolution step should themselves be determined from a new global fit to data using the @xmath0-factorisation approach . for the present paper ,
we take the input pdfs from a global fit to data using the conventional collinear approximation @xcite .
this treatment is adequate for these initial investigations .
however , we expect it to lower our predictions for quark - initiated processes by @xmath1610% compared to the case where the input pdfs are determined from a global fit using the @xmath0-factorisation approach .
we will illustrate this point in section 5 by comparing predictions for the proton structure function @xmath17 in the collinear approximation and in the @xmath0-factorisation formalism .
the ` last - step ' prescription has some features in common with the initial - state parton shower algorithms implemented in monte carlo event generators ( for a recent review , see @xcite ) such as the dglap - based herwig @xcite and pythia @xcite programs and the ccfm - based cascade @xcite program .
the main advantage of our approach is that we use simple analytic formulae which implement the crucial physics in a transparent way , without the additional details or tuning which are frequently introduced in monte carlo programs .
the @xmath1 distributions are generated entirely from known and universal dupdfs .
for example , fits to @xmath3 production data at the tevatron using parton showers favour a large intrinsic partonic transverse momentum @xmath18 gev , while confinement of partons inside the proton would imply a @xmath19 gev @xcite .
the dglap - based parton showers used in @xcite are theoretically justified only in the limit of strongly - ordered transverse momentum , since only the collinear divergent part of the squared matrix element is kept in each parton branching .
similarly , the ccfm - based parton shower used in @xcite is strictly justified only in the limit of strongly - ordered angles , which reduces to the limit of strongly - ordered transverse momentum as long as @xmath15 is not too small . in this limit ,
the transverse momentum generated in all evolution steps prior to the last is negligible .
therefore , neglecting transverse momentum in every evolution step prior to the last should be a good approximation to the parton shower algorithms in which finite transverse momentum is generated at _ every _ evolution step .
the goal of the present paper is to demonstrate that the @xmath1 distributions of electroweak bosons can be successfully generated by dupdfs .
we do not aim to produce a _
better _ description of the data than existing calculations , but rather a _
simpler _ analytic description which reproduces the main features . with this approach ,
it is easy to see the physical origin of the @xmath1 distributions and to identify the most important feynman diagrams .
since the dupdfs are universal that is , they apply equally well to all hard hadronic processes it is important to check them in a new kinematic domain .
one topical application is the prediction of the cross section for diffractive higgs boson production at the lhc @xcite , which is driven by the unintegrated domain the skewed effect can be included by the shuvaev prescription @xcite . ]
gluon distribution @xmath20 , where @xmath21 is the hard scale of the subprocess .
at the moment , the only possibility to check the behaviour of updfs in the domain @xmath22 is to compare predictions with the observed @xmath1 distributions of @xmath2 and @xmath3 bosons produced at the tevatron . we will show that the doubly - unintegrated quark distributions , generated directly from the known integrated pdfs under the ` last - step ' prescription , satisfactorily describe these data , including the region of interest , @xmath23 . in section [ sec : zkthad ]
we describe the formalism for @xmath0-factorisation at hadron - hadron colliders , and in section [ sec : apply ] we apply it to calculate the @xmath1 distributions of electroweak bosons .
numerical results are given in section [ sec : numerical ] .
we now extend the formalism of @xcite , which concerned deep - inelastic scattering ( dis ) , to hadron - hadron collisions . the basic idea is illustrated in fig .
[ fig : zkt](a ) , which shows only one of the possible configurations .
all permutations of quarks and gluons must be included .
the arrows show the direction of the labelled momenta .
the blobs represent the familiar integrated pdfs .
the transverse momenta of the two incoming partons to the subprocess , represented by the rectangles labelled @xmath24 in fig .
[ fig : zkt ] , are generated by a single parton emission in the last evolution step .
( a)(b ) + -factorisation : the last evolution step is factorised into @xmath25 , where @xmath26.,scaledwidth=80.0% ] -factorisation : the last evolution step is factorised into @xmath25 , where @xmath26.,scaledwidth=80.0% ] we use a sudakov decomposition of the momenta of the two incoming partons : @xmath27 where @xmath28 or @xmath29 .
we work in the centre - of - mass frame of the colliding hadrons and neglect the hadron masses so that the squared centre - of - mass energy is @xmath30 .
then , are @xmath31 . ] @xmath32 the penultimate propagators in the evolution ladder have momenta @xmath33 , so that the partons emitted in the last step have momenta @xmath34 the on - shell condition for the emitted partons , @xmath35 , determines @xmath36 where @xmath37 , so that the two incoming partons have virtuality @xmath38 .
the total momentum going into the subprocess labelled @xmath24 in fig .
[ fig : zkt ] is @xmath39 where @xmath40 .
the kinematic variables obey the ordering @xmath41 if @xmath42 are gluon momenta , then we must additionally impose angular - ordering constraints due to colour coherence @xcite : @xmath43 where @xmath44 and @xmath45 .
that is , the subprocess separates gluons emitted from each of the two hadrons .
this condition leads to a suppression of soft gluon emission : @xmath46 with @xmath47 and @xmath48 .
the partonic cross sections , represented by the rectangles in fig .
[ fig : zkt ] , depend on @xmath49 , @xmath50 , and @xmath51 through the momenta of the incoming partons @xmath52 .
therefore , the partonic cross sections must be convoluted with parton distributions depending on the same variables .
we can define dupdfs @xmath53 which satisfy the normalisation conditions @xmath54 for fixed @xmath15 and @xmath21 , independent of @xmath55 and @xmath13 . here , @xmath56 or @xmath57 are the conventional integrated pdfs
. explicit formulae for the dupdfs are given in @xcite .
the doubly - unintegrated quark distribution is @xmath58,\ ] ] where @xmath59 and @xmath60 are the unregulated lo dglap splitting kernels , and the quark sudakov form factor is @xmath61 the doubly - unintegrated gluon distribution is @xmath62,\ ] ] where the gluon sudakov form factor is @xmath63,\ ] ] with @xmath64 the number of active flavours . in terms of the dupdfs
, the hadronic cross section @xmath65 is related to the partonic cross sections @xmath66 by the @xmath0-factorisation formula @xmath67 this formula is represented schematically in fig .
[ fig : zkt](b ) for the case where the partons @xmath68 and @xmath69 are both quarks .
the partonic cross sections in are given by @xmath70 where the flux factor @xmath71 .
the last evolution steps in fig .
[ fig : zkt](a ) factorise from the rest of the diagram , to give the lo dglap splitting kernels , in the leading logarithmic approximation ( lla ) where only the @xmath72 term is kept .
therefore , @xmath73 is calculated with the replacement @xmath74 in the numerator in order to keep only this collinear divergent term .
however , any propagator virtualities appearing in the denominator of @xmath73 may be evaluated with the full kinematics , as may the phase space element @xmath75 . for this approach to work , it is vital that the @xmath72 term is obtained _ only _ from ladder - type diagrams like that in fig .
[ fig : zkt](a ) , and not from interference ( non - ladder ) diagrams .
this is true if we use a physical gauge for the gluon , where only the two transverse polarisations propagate . for hadron - hadron collisions ,
the natural choice is the planar gauge where the sum over gluon polarisations is performed using @xmath76 where we take the gauge - fixing vector @xmath77 .
such a gauge choice ensures that the @xmath72 term is obtained from ladder - type diagrams on _ both _ sides of the subprocess represented by the rectangle in fig .
[ fig : zkt](a ) @xcite .
a related requirement is that terms beyond the leading @xmath72 term , coming from non - ladder diagrams , for example , give a negligible contribution .
such terms are proportional to the sudakov variable @xmath78 and hence vanish in the limit that @xmath79 or @xmath80 ( unless @xmath81 ) . away from these limits
it is not obvious that these `` beyond lla '' terms will be small , a necessary condition for the factorisation to hold . for the case of inclusive jet production in dis and working in an axial gluon gauge
, it was observed in @xcite that the main effect of the extra terms was to suppress soft gluon emission .
when the angular - ordering constraint was applied , the extra terms were found to make a negligible difference to the cross section . for hadron - hadron collisions ,
although the number of possible non - ladder diagrams is larger , it is therefore reasonable to expect that the extra terms will have little numerical effect , at least for @xmath51 less than the hard scale of the subprocess .
a similar argument is made to justify the approximation made in the dglap - based parton showers used in monte carlo simulations , where only the collinear divergent part of the squared matrix element for each parton branching is kept and angular ordering is imposed in all evolution steps to account for some of the missing terms . here , we are more conservative and apply this approximation to the _ last _ evolution step only .
the dupdfs in are only defined for @xmath82 , where @xmath83 gev is the minimum scale for which dglap evolution of the integrated pdfs is valid .
the approximation of the @xmath84 contribution made in @xcite was to take the limit @xmath85 in the kinematic variables ( and in @xmath66 ) , then to make the replacement @xmath86 where @xmath87 are the sudakov form factors or .
this replacement ensures that the normalisation conditions are satisfied ( see @xcite ) .
a better approximation , which retains the @xmath51 dependence , is to take the limit @xmath79 in the kinematic variables , then to make the replacement @xmath88 where we have used ( 23 ) of @xcite .
the requirement that @xmath89 as @xmath90 is a consequence of gauge invariance @xcite .
a more complicated extrapolation of the dupdfs for @xmath91 , which allows both the @xmath13 and @xmath55 dependence to be retained in the kinematic variables , is to assume the polynomial form @xmath92.\ ] ] the two coefficients @xmath93 and @xmath94 can be determined to ensure continuity at @xmath95 and the correct normalisation , leading to @xmath96 if becomes negative , which is possible for @xmath97 , then @xmath53 is simply set to zero .
this extrapolation of the perturbative formulae accounts for some non - perturbative `` intrinsic '' @xmath13 of the initial partons , which is often parameterised by a gaussian distribution with @xmath98 .
numerical results are insensitive to the precise form , , or used for the @xmath91 contribution .
however , we will use the form which ensures continuity at @xmath99 .
distribution of @xmath2 or @xmath3 bosons in the @xmath0-factorisation approach.,scaledwidth=50.0% ] distribution of higgs bosons in the @xmath0-factorisation approach .
the cross represents the effective vertex in the limit @xmath100.,scaledwidth=50.0% ] perhaps the simplest application of @xmath0-factorisation at hadron - hadron colliders is electroweak boson production , where at lo the subprocess is simply @xmath101 , illustrated in fig .
[ fig : qqv ] , or @xmath102 , illustrated in fig .
[ fig : ggh ] . here
, the ` @xmath103 ' indicates that the incoming partons , with momenta @xmath52 given by , are off - shell with virtuality @xmath104 . in the collinear approximation , these diagrams give the born level estimate of the total cross section @xmath65 .
when each parton carries finite transverse momentum @xmath105 , the final electroweak boson has transverse momentum @xmath106 , so we can calculate the @xmath1 distribution @xmath107 with @xmath65 given by the @xmath0-factorisation formula .
the partonic differential cross sections for @xmath2 or @xmath3 production are @xmath108 where @xmath109 is the number of colours , @xmath110 is the fermi coupling constant , @xmath111 is the ckm matrix element squared , and @xmath112 is the sum of the vector and axial vector couplings squared .
the dominant mechanism for sm higgs production in hadron - hadron collisions is by gluon - gluon fusion via a top quark loop . for the case
where @xmath113 , the well - known effective @xmath114 vertex can be derived from the lagrangian @xcite @xmath115 where @xmath116 , @xmath117 is the gluon field strength tensor , and @xmath118 is the higgs field .
the partonic differential cross section for sm higgs production is then @xmath119 for @xmath120 , these expressions and are exactly as in the collinear approximation .
the difference arises when we consider the precise kinematics @xmath121,\qquad
r\equiv q_t^2/s , \\
q_t^2 = \lvert{\boldsymbol{k_{1,t}}}+{\boldsymbol{k_{2,t}}}\rvert^2 = k_{1,t}^2 + k_{2,t}^2 + 2\,k_{1,t}\,k_{2,t}\,\cos\phi.\end{gathered}\ ] ] applying the @xmath0-factorisation formula , the first delta function in and can be used to do the @xmath122 integration in , while the second delta function can be used to do the @xmath123 integration .
in addition , we need to average over the azimuthal angle @xmath124 between @xmath125 and @xmath126 .
the final hadronic differential cross sections for @xmath2 or @xmath3 production are @xmath127 with @xmath128 , @xmath129 , and @xmath130 in practice , the kinematic constraints mean that the @xmath131 solution does not contribute . the corresponding result for sm
higgs production is @xmath132 where @xmath133 and @xmath134 and @xmath135 are as above .
note that we have taken the @xmath114 vertex in the @xmath100 limit . for @xmath136 , the correction to the total cross section due to the top quark mass
can be approximated @xcite by a factor @xmath137 ^ 2.\ ] ] the @xmath84 contributions of and are accounted for using the approximation .
, such as the inclusion of additional intrinsic partonic transverse momentum with @xmath19 gev , would only affect the @xmath1 distributions at very low @xmath138 gev . ]
in the collinear approximation , higher order qcd corrections to the lo diagrams , @xmath8 or @xmath10 , are known to be significant when calculating the total cross section .
the ratio of the corrected result to the leading order result is the so - called @xmath139-factor .
a part of these higher order corrections is kinematic in nature , arising from real parton emission , which we have already accounted for at lo in the @xmath0-factorisation approach ( see fig . [
fig : zkt ] ) .
another part comes from the logarithmic loop corrections which have already been included in the sudakov form factors and .
however , we need to include the non - logarithmic loop corrections arising , for example , from the gluon vertex correction to figs . [
fig : qqv ] and [ fig : ggh ] .
a large part of these non - logarithmic corrections have a semi - classical nature and may be obtained from the analytic continuation of the double logarithm in the sudakov form factors in going from spacelike ( dis ) to timelike ( drell - yan ) electroweak boson momenta @xcite .
accounting for the running of the strong coupling @xmath140 , the sudakov form factors in the dlla are @xmath141 and @xmath142 where @xmath143 , @xmath144 , and the colour factors are @xmath145 and @xmath146 . replacing @xmath147 by @xmath148
, we obtain the @xmath149-enhanced part of the @xmath139-factors : @xmath150 and @xmath151 a particular scale choice @xmath152 has been found @xcite to eliminate certain sub - leading logarithms in the sudakov form factors .
therefore , we choose this scale to evaluate @xmath153 in and .
the @xmath1 distributions of produced @xmath2 and @xmath3 bosons were measured by the cdf @xcite and d @xcite collaborations during the tevatron run 1 , in @xmath154 collisions at a centre - of - mass energy of @xmath155 tev .
measurements were made of @xmath156 and @xmath157 decays ; therefore , we must multiply the theoretical predictions for @xmath2 or @xmath3 production by the appropriate leptonic branching ratios . and br@xmath158 @xcite . ]
we use the mrst2001 lo pdfs @xcite as input , with @xmath159 gev@xmath160 .
the lo predictions for the @xmath1 distributions , integrated over bins of 1 gev , are shown by the dashed lines in figs .
[ fig : cdfz ] , [ fig : d0z ] and [ fig : d0w ] . the integrated luminosity uncertainty ( 3.9% for cdf or 4.4% for d ) is not included in the error bars for the plotted data .
the lo predictions multiplied by the @xmath139-factor are shown by the solid lines .
although the @xmath139-factor makes up the major part of the discrepancy , we see that the solid lines in figs .
[ fig : cdfz ] , [ fig : d0z ] , and [ fig : d0w ] still underestimate the precise measurements at small @xmath1 to some extent . normalising to the total measured cross sections ( 248 pb , 221 pb , and 2310 pb ) by factors 1.23 , 1.10 , and 1.08 respectively , as shown by the dotted lines in figs .
[ fig : cdfz ] , [ fig : d0z ] , and [ fig : d0w ] , gives a very good description of the data over the entire @xmath1 range .
although it makes sense to take @xmath161 as low as possible , the insensitivity of the @xmath1 distributions to the @xmath91 treatment can be demonstrated by taking @xmath162 gev@xmath160 .
the @xmath1 distributions obtained are practically identical to those with @xmath159 gev@xmath160 , with less than a 0.5% change in the total cross sections .
notice that the predicted @xmath1 distribution of @xmath3 bosons peaks about 0.51.0 gev below the cdf data ( fig .
[ fig : cdfz ] ) .
one possible explanation for this is provided by non - perturbative power corrections , part of which may be interpreted as a negative correction of about @xmath163 gev@xmath160 to the factorisation scale at which the integrated pdfs are evaluated @xcite .
such a shift in the factorisation scale is found to move the peak of the @xmath1 distribution about 0.2 gev in the direction of the cdf data , with a slightly larger normalisation factor of 1.24 .
distribution of @xmath3 bosons compared to cdf data @xcite . ]
distribution of @xmath3 bosons compared to d data @xcite . ]
distribution of @xmath2 bosons compared to d data @xcite . ]
the small residual discrepancy between the solid lines in figs .
[ fig : cdfz ] , [ fig : d0z ] , and [ fig : d0w ] and the data is easily understood .
note that the mrst2001 lo pdfs @xcite have been determined by a global fit to data using the conventional collinear approximation .
a more precise treatment would fit the integrated pdfs , used as input to the last evolution step , to the proton structure function @xmath17 , for example , using the @xmath0-factorisation formalism at lo .
we would expect this treatment to give slightly larger integrated pdfs than the conventional sets by a factor of @xmath164 and so eliminate the small discrepancy between the @xmath0-factorisation predictions and the data .
alternatively , it was found in @xcite that the major higher order corrections to the inclusive jet cross section in dis could be accounted for by adding extra parton emissions to the lo diagram , @xmath165 . in an axial gluon gauge
, ladder - type diagrams gave the dominant contributions .
thus , it is likely that calculating the @xmath166 subprocesses @xmath167 and @xmath168 using the @xmath0-factorisation prescription would account for any significant higher order corrections not already included and so reduce the observed discrepancy without refitting the input integrated pdfs .
these reasons for the small discrepancies discussed in the previous two paragraphs can be illustrated by considering the proton structure function @xmath169 , which is plotted for two @xmath170 values in fig .
[ fig : f2 ] . in the collinear approximation ,
the lo prediction for this observable comes from @xmath171 @xmath172 indicated by the solid lines in fig .
[ fig : f2 ] , which gives a good description of the data points since this data set was included in the mrst2001 lo global fit .
the lo @xmath0-factorisation prediction comes from @xmath173 and may be obtained from ( 85 ) of @xcite : @xmath174 where the sudakov variables @xmath175 and @xmath176 are given in ( 54 ) of @xcite and the factorisation scale @xmath21 is given in ( 25 ) of @xcite .
the predictions of this formula are shown as the dashed lines in fig .
[ fig : f2 ] , while the first term of , representing the non - perturbative contributions from @xmath177 , is also shown separately as the dotted lines .
there is a clear difference between the predictions of and , which increases as @xmath178 decreases , due to the extra kinematic factor in the second term of .
this difference would be eliminated by fitting the input integrated pdfs using .
alternatively , a `` nlo '' prediction for @xmath17 may be calculated from the subprocesses @xmath179 and @xmath180 , and can be obtained from ( 94 ) of @xcite . here , a lower limit of @xmath181 is taken in the @xmath182 integration , and the @xmath183 contribution is instead accounted for using the first term of .
it is seen that these `` nlo '' predictions , shown as the dot - dashed lines in fig .
[ fig : f2 ] , give almost the same results as . at @xmath170 values of 22 gev@xmath160 ( top ) and 200 gev@xmath160 ( bottom ) in the collinear approximation , where the 1996/97 zeus data @xcite has been included in the mrst2001 lo
fit @xcite , and in the @xmath0-factorisation approach using the same pdfs as input .
the discrepancy between the data and the lo @xmath0-factorisation prediction can be eliminated by either refitting the input integrated pdfs or by adding some `` nlo '' contribution.,title="fig:",scaledwidth=80.0% ] + at @xmath170 values of 22 gev@xmath160 ( top ) and 200 gev@xmath160 ( bottom ) in the collinear approximation , where the 1996/97 zeus data @xcite has been included in the mrst2001 lo fit @xcite , and in the @xmath0-factorisation approach using the same pdfs as input . the discrepancy between the data and the lo @xmath0-factorisation prediction can be eliminated by either refitting the input integrated pdfs or by adding some `` nlo '' contribution.,title="fig:",scaledwidth=80.0% ] the @xmath1 distribution for sm higgs bosons of mass 125 gev produced at the lhc ( @xmath184 tev ) is shown in fig .
[ fig : hprod ] . to allow direct comparison with the results of @xcite
, we do _ not _ account for top quark mass effects .
note that the peak in the higgs @xmath1 distribution is broader and occurs at a higher @xmath1 than for vector boson production .
this is primarily due to the enhanced @xmath185 colour factor ( @xmath146 ) compared to the @xmath186 colour factor ( @xmath145 ) , resulting in a larger sudakov suppression at low @xmath1 . by the same reason the @xmath139-factor is larger . for @xmath187 ,
the @xmath1 distribution is in good agreement with recent , more sophisticated , resummation predictions ( see , for example , @xcite ) , bearing in mind the spread in the various predictions available due to the different approaches and pdfs used .
however , the peak occurs at a @xmath1 about 12 gev lower than the majority of the resummation predictions ( cf .
fig . 1 of @xcite ) .
evaluating the total cross section by integrating over all @xmath1 gives 38.6 pb , close to the next - to - next - to - leading order ( nnlo ) calculation which gives 39.4 pb @xcite .
note that matrix - element corrections are necessary in parton shower simulations at large @xmath1 . without such corrections ,
the herwig parton shower prediction falls off dramatically at large @xmath188 @xcite .
the fact that our prediction is much closer to the fixed - order results than parton shower predictions suggests that we have successfully accounted for a large part of the sub - leading terms .
distribution of sm higgs bosons produced at the lhc with mass 125 gev . ]
we have extended the method of @xmath0-factorisation using doubly - unintegrated parton distributions @xcite to hadron - hadron collisions .
the key idea is that the incoming partons to the subprocess have finite transverse momenta , which can be observed in the particles produced in the final state .
this transverse momentum is generated perturbatively in the last evolution step , with a suitable extrapolation for the non - perturbative contribution .
virtual terms in the dglap equation are resummed into sudakov form factors and angular - ordering constraints are applied which regulate soft gluon emission . by accounting for the precise kinematics in the subprocess , together with these sudakov form factors and angular - ordering constraints , we are able to include the main part of conventional higher order calculations .
we used this framework to calculate the @xmath1 distributions of @xmath2 and @xmath3 bosons produced at the tevatron run 1 .
the predictions gave a very good description of cdf and d data over the whole @xmath1 range , after multiplying by an overall factor of 1.11.2 , corresponding to multiplying each dupdf by a factor @xmath189 .
we explained the origin of the need for this extra factor , which should not be regarded as a deficiency of our approach , but rather reflects the fact that the input integrated pdfs should themselves be determined from data using @xmath0-factorisation .
we also used the framework to calculate the @xmath1 distribution for sm higgs bosons of mass 125 gev produced at the lhc . for @xmath187 ,
our simple prescription was found to reproduce , to a fair degree , the predictions of more elaborate theoretical studies @xcite .
adm thanks the leverhulme trust for an emeritus fellowship .
this work was supported by the uk particle physics and astronomy research council and by the federal program of the russian ministry of industry , science and technology ( grant ss-1124.2003.2 ) .
r. k. ellis and s. veseli , nucl .
b * 511 * ( 1998 ) 649 [ arxiv : hep - ph/9706526 ] ; and references therein . c. balazs and c. p. yuan , phys .
d * 56 * ( 1997 ) 5558 [ arxiv : hep - ph/9704258 ] ; and references therein . b. andersson _ et al .
_ [ small @xmath15 collaboration ] , eur . phys .
j. c * 25 * ( 2002 ) 77 [ arxiv : hep - ph/0204115 ] ; h. jung , mod .
a * 19 * ( 2004 ) 1 [ arxiv : hep - ph/0311249 ] ; j. r. andersen _ et al . _
[ small @xmath15 collaboration ] , eur .
j. c * 35 * ( 2004 ) 67 [ arxiv : hep - ph/0312333 ] .
a. gawron and j. kwiecinski , phys
. rev .
d * 70 * ( 2004 ) 014003 [ arxiv : hep - ph/0309303 ] ; j. kwiecinski and a. szczurek , nucl .
b * 680 * ( 2004 ) 164 [ arxiv : hep - ph/0311290 ] .
m. ciafaloni , nucl .
b * 296 * ( 1988 ) 49 ; s. catani , f. fiorani and g. marchesini , phys .
b * 234 * ( 1990 ) 339 ; s. catani , f. fiorani and g. marchesini , nucl .
b * 336 * ( 1990 ) 18 ; g. marchesini , nucl .
b * 445 * ( 1995 ) 49 [ arxiv : hep - ph/9412327 ] .
m. a. kimber , j. kwiecinski , a. d. martin and a. m. stasto , phys .
d * 62 * ( 2000 ) 094006 [ arxiv : hep - ph/0006184 ] ; m. a. kimber , a. d. martin and m. g. ryskin , eur .
j. c * 12 * ( 2000 ) 655 [ arxiv : hep - ph/9911379 ] .
m. a. kimber , a. d. martin and m. g. ryskin , phys .
d * 63 * ( 2001 ) 114027 [ arxiv : hep - ph/0101348 ] .
j. kwiecinski , a. d. martin and a. m. stasto , phys .
d * 56 * ( 1997 ) 3991 [ arxiv : hep - ph/9703445 ] . g. watt , a. d. martin and m. g. ryskin , eur .
j. c * 31 * ( 2003 ) 73 [ arxiv : hep - ph/0306169 ] .
y. l. dokshitzer , d. diakonov and s. i. troian , phys .
* 58 * ( 1980 ) 269 .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , phys .
b * 531 * ( 2002 ) 216 [ arxiv : hep - ph/0201127 ] .
m. a. dobbs _ et al .
_ , arxiv : hep - ph/0403045 . g. corcella _ et al .
_ , arxiv : hep - ph/0210213 . t. sjostrand , l. lonnblad , s. mrenna and p. skands , arxiv : hep - ph/0308153 .
h. jung , comput .
commun .
* 143 * ( 2002 ) 100 [ arxiv : hep - ph/0109102 ] . c. balazs , j. huston and i. puljak , phys .
d * 63 * ( 2001 ) 014021 [ arxiv : hep - ph/0002032 ] ; e. thome , arxiv : hep - ph/0401121 ; j. huston , i. puljak , t. sjostrand and e. thome , arxiv : hep - ph/0401145 .
v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. c * 23 * ( 2002 ) 311 [ arxiv : hep - ph/0111078 ] .
a. shuvaev , phys .
d * 60 * ( 1999 ) 116005 [ arxiv : hep - ph/9902318 ] ; a. g. shuvaev , k. j. golec - biernat , a. d. martin and m. g. ryskin , phys .
d * 60 * ( 1999 ) 014015 [ arxiv : hep - ph/9902410 ] ; a. d. martin and m. g. ryskin , phys .
d * 64 * ( 2001 ) 094017 [ arxiv : hep - ph/0107149 ] .
l. v. gribov , e. m. levin and m. g. ryskin , phys .
* 100 * ( 1983 ) 1 ; a. j. askew , j. kwiecinski , a. d. martin and p. j. sutton , phys .
d * 49 * ( 1994 ) 4402 [ arxiv : hep - ph/9310261 ] .
j. r. ellis , m. k. gaillard and d. v. nanopoulos , nucl .
b * 106 * ( 1976 ) 292 ; m. a. shifman , a. i. vainshtein , m. b. voloshin and v. i. zakharov , sov .
j. nucl .
phys . * 30 * ( 1979 ) 711 [ yad .
* 30 * ( 1979 ) 1368 ] .
r. k. ellis , w. j. stirling and b. r. webber , cambridge monogr .
phys . cosmol .
* 8 * ( 1996 ) 1 .
g. parisi , phys .
b * 90 * ( 1980 ) 295 ; g. curci and m. greco , phys .
b * 92 * ( 1980 ) 175 ; l. magnea and g. sterman , phys .
d * 42 * ( 1990 ) 4222 .
a. kulesza and w. j. stirling , nucl .
b * 555 * ( 1999 ) 279 [ arxiv : hep - ph/9902234 ] .
t. affolder _ et al .
_ [ cdf collaboration ] , phys . rev .
* 84 * ( 2000 ) 845 [ arxiv : hep - ex/0001021 ] .
b. abbott _ et al . _
[ d collaboration ] , phys .
d * 61 * ( 2000 ) 032004 [ arxiv : hep - ex/9907009 ] .
b. abbott _ et al . _
[ d collaboration ] , phys .
b * 513 * ( 2001 ) 292 [ arxiv : hep - ex/0010026 ] .
k. hagiwara _ et al .
_ [ particle data group collaboration ] , phys .
d * 66 * ( 2002 ) 010001 .
a. guffanti and g. e. smye , jhep * 0010 * ( 2000 ) 025 [ arxiv : hep - ph/0007190 ]
. s. chekanov _ et al . _
[ zeus collaboration ] , eur .
j. c * 21 * ( 2001 ) 443 [ arxiv : hep - ex/0105090 ] . c. balazs , m. grazzini , j. huston , a. kulesza and i. puljak , arxiv : hep - ph/0403052 ; and references therein .
g. corcella and s. moretti , phys .
b * 590 * ( 2004 ) 249 [ arxiv : hep - ph/0402146 ] ; g. corcella and s. moretti , arxiv : hep - ph/0402149 . | we describe the use of doubly - unintegrated parton distributions in hadron - hadron collisions , using the @xmath0-factorisation prescription where the transverse momentum of the incoming parton is generated in the last evolution step .
we apply this formalism to calculate the transverse momentum ( @xmath1 ) distributions of produced @xmath2 and @xmath3 bosons and compare the predictions to tevatron run 1 data .
we find that the observed @xmath1 distributions can be generated almost entirely by the leading order @xmath4 subprocesses , using known and universal doubly - unintegrated quark distributions .
we also calculate the @xmath1 distribution of the standard model higgs boson at the lhc , where the dominant production mechanism is by gluon - gluon fusion .
ippp/03/50 + dcpt/03/100 + 9th november 2004 + * unintegrated parton distributions and + electroweak boson production at hadron colliders * g. watt@xmath5 , a.d . martin@xmath5 and m.g .
ryskin@xmath6 + @xmath5 institute for particle physics phenomenology , university of durham , dh1 3le , uk + @xmath7 petersburg nuclear physics institute , gatchina , st .
petersburg , 188300 , russia |
computer science and physics , although different disciplines in essence , have been closely linked since the birth of the first one . more recently
, computer science has met together with statistical physics in the so called combinatorial problems and their relation to phase transitions and computational complexity ( see @xcite for a compendium of recent works ) .
more accurately , algorithmic phase transitions ( threshold property in the computer science language ) , i.e. sharp changes in the behavior of some computer algorithms , have attracted the attention of both communities @xcite .
it has been shown that phase transitions play an important role in the resource growing classification of random combinatorial problems @xcite .
the computational complexity theory is therefore nowadays experimenting a widespread growth , melting different ideas and approaches coming either from theoretical computation , discrete mathematics , and physics .
for instance , there exist striking similarities between optimization problems and the study of the ground states of disordered models @xcite .
+ problems related to random combinatorics appear typically in discrete mathematics ( graph theory ) , computer science ( search algorithms ) or physics ( disordered systems ) .
the concept of sudden change in the behavior of some variables of the system is intimately linked to this hallmark .
for instance , erd@xmath0s and renyi , in their pioneering work on graph theory @xcite , found the existence of _ zero - one _ laws in their study of cluster generation .
these laws have a clear interpretation in terms of phase transitions , which appear extensively in many physical systems .
more recently , computer science community has detected this behavior in the context of algorithmic problems .
the so called _ threshold phenomenon _
@xcite distinguishes zones in the phase space of an algorithm where the problem is , computationally speaking , either tractable or intractable .
it is straightforward that these three phenomena can be understood as a unique concept , in such a way that building bridges between each other is an appealing idea .
+ related to the concept of a phase transition is the task of classifying combinatorial problems .
the theory of computational complexity distinguishes problems which are tractable , that is to say , solvable in polynomial time by an efficient algorithm , from those which are not .
the so - called @xmath1 class gathers problems that can be solved in polynomial time by a non - deterministic turing machine @xcite .
this class generally includes many hard or eventually intractable problems , although this classification is denoted _ worst - case _ , that is to say , a rather pessimistic one , since the situations that involve long computations can be eventually rare . in the last years , numerical evidences suggest the presence of the threshold phenomenon in @xmath1 problems .
these phase transitions may in turn characterize the _ average - case _ complexity of the associated problems , as pointed out recently @xcite .
+ + in this paper we discuss a stochastic algorithm inspired on artificial chemistry models @xcite that has already been studied from a statistical physics point of view @xcite .
this algorithm generates prime numbers by means of a stochastic integer decomposition .
a continuous phase transition has been detected and described in a previous work @xcite : we can distinguish a stationary phase where the ability for producing primes is practically null and a stationary phase where the algorithm is able to reduce the whole system into primes .
it is straightforward to reinterpret the model as a search problem @xcite which undergoes an algorithmic phase transition related to a change in its computational complexity . in this paper
we firstly pretend to make a broader characterization of the system ; in this sense this work is a continuation of a previous one @xcite .
further on , we will situate the model in the context of the computational complexity theory , in order to relate its computational complexity with the phase transition present in the system . the paper
thus goes as follows : in section ii we will describe the model , which stands as a stochastic prime number generator . in section iii
we will characterize the phase transition present in the system , following the steps depicted in a previous work @xcite and providing some new data and additional information . concretely
, we will firstly outline the breaking of symmetry responsible for the order - disorder transition . after defining proper order and control parameters
, critical exponents will be calculated numerically from extensive simulations and finite size scaling analysis .
an analytical approach to the system will also be considered at this point , in terms of an annealed approximation . in section iv
, we will reinterpret the model as a search problem .
we will then show that the system belongs to the @xmath1 class in a worst - case classification .
we will find an easy - hard - easy pattern in the algorithm , as common in many @xmath1 problems , related in turn to the critical slowing down near the the transition point . according to @xcite
, we will finally relate the nature of the phase transition with the average - case complexity of the problem .
we will show that while the problem is in @xmath1 , the resource growing is only polynomial . in section v we will conclude .
the fundamental theorem of arithmetic @xcite states that every integer greater than one can be expressed uniquely as a product of primes or powers of primes . in a certain manner ,
prime numbers act as atoms in chemistry , both are irreducible . this way , composite numbers can be understood as molecules .
following this resemblance , the next algorithm has been introduced @xcite : suppose that we got a pool of positive integers @xmath2 , from which we randomly extract a certain number @xmath3 of them ( this will constitute the system under study ) .
note that the chosen numbers can be repeated , and that the integer @xmath4 is not taken into account .
now , given two numbers @xmath5 and @xmath6 taken from the system of @xmath3 numbers , the algorithm collision rules are the following : * rule 1 : if @xmath7 there is no reaction ( elastic shock ) , and the numbers are not modified .
* rule 2 : if the numbers are different ( say @xmath8 ) , a reaction will take place only if @xmath6 is a divisor of @xmath5 , i.e. if @xmath5 mod @xmath9 .
the reaction will then stand for @xmath10 where @xmath11 stands for the usual notation of a chemical reaction and @xmath12 .
* rule 3 : if @xmath8 but @xmath5 mod @xmath13 , no reaction takes place ( elastic shock ) .
the result of a reaction will be the extinction of the composed number @xmath5 and the introduction of a simpler one , @xmath14 .
+ the algorithm goes as follows : after randomly extracting from the pool @xmath2 a set of @xmath3 numbers , we pick at random two numbers @xmath5 and @xmath6 from the set : this is equivalent to the random encounter of two molecules .
we then apply the reaction rules . in order to have a parallel updating
, we will establish @xmath3 repetitions of this process ( @xmath3 monte carlo steps ) as a time step .
note that the reactions tend to separate numbers into its irreducible elements , as molecules can be separated into atoms .
hence , this dynamic when iterated may generate prime numbers in the system .
we say that the system has reached stationarity when every collision is elastic ( no more reaction can be achieved ) , whether because every number has become a prime or because rule 2 can not be satisfied in any case -frozen state- .
the algorithm then stops .
versus @xmath3 , for a pool size @xmath15 .
each run is averaged over @xmath16 realizations in order to avoid fluctuations .
note that the system exhibits a phase transition which distinguishes a phase where every element of the system becomes a prime in the steady state and a phase with low prime density.,scaledwidth=45.0% ] as stated in the previous section , this algorithm clearly tends to generate primes as far as possible : when the algorithm stops , one may expect the system to have a large number of primes or at least have a frozen state of non - divisible pairs . a first indicator that can evaluate properly this feature is the unit percentage or ratio of primes @xmath17 , that a given system of @xmath3 numbers reaches at stationarity @xcite . in figure [ r_y_v ]
we present the results of monte carlo simulations calculating , as a function of @xmath3 and for a concrete pool size @xmath15 , the steady values of @xmath17 . every simulation is averaged over @xmath16 realizations in order to avoid fluctuations .
we can clearly distinguish in figure [ r_y_v ] two phases , a first one where @xmath17 is small and a second one where the prime number concentration reaches the unity .
realizations ) of the @xmath3 elements , for @xmath18 and @xmath19 ( phase with low @xmath17 ) : this one is a uniform distribution u(2,m ) ( note that the distribution is not normalized ) .
the right figure stands for the same plot for @xmath20 and @xmath19 ( phase where @xmath17 reaches the unity ) : this one is a power law @xmath21.,scaledwidth=75.0% ] this is the portrait of a phase transition , where @xmath3 would stand as the control parameter and @xmath17 as the order parameter . in the phase with small @xmath17 , the average steady state distribution of the @xmath3 elements is plotted in the left side of figure ( [ distris ] ) : the distribution is uniform ( note that the vertical scale is zoomed in such a way that if we scale it between @xmath22 $ ] we would see a horizontal line ) , which is related to an homogeneous state . in this regime
, every number has the same probability to appear in the steady state . in the other hand ,
the average steady state distribution of the @xmath3 numbers in the phase of high @xmath17 is plotted in the right side of figure ( [ distris ] ) : the distribution is now a power law , which is related to a biased , inhomogeneous state . in this regime ,
the probability of having -in the steady state- a composite number is practically null , and the probability of having the prime @xmath23 is in turn proportional to @xmath24 @xcite . versus @xmath3 , for different pool sizes @xmath25 ( each simulation
is averaged over @xmath16 realizations).,scaledwidth=45.0% ] the breaking of this symmetry between steady distributions leads us to assume an order - disorder phase transition , the phase with small proportion of primes being the disordered phase and the ordered phase being the one where @xmath17 tends to one .
+ a second feature worth investigating is the size dependence of the transition . in figure [ r2 ]
we plot @xmath17 versus @xmath3 , for a set of different pool sizes @xmath25 .
note that the qualitative behavior is size invariant , however the transition point increases with @xmath25 .
this size dependence will be considered in a later section .
+ , ( middle ) critical phase @xmath26 , ( bottom ) ordered phase @xmath27.,scaledwidth=45.0% ] as a function of the time steps for the same configurations as for figure [ series].,scaledwidth=45.0% ] as a third preliminary insight , we shall study the temporal evolution of the system . in figure
[ series ] we plot , for a given pool size @xmath19 , the cumulated number of reactions that a system of @xmath3 numbers needs to make in order to reach stationarity . according to this , in figure [ concentraciones ]
we plot , for the same @xmath28 , the evolving value @xmath29 . in the disordered phase
we can see that the system is rapidly frozen : the algorithm is not efficient in producing primes , and @xmath17 is asymptotically small .
in the ordered phase the system needs more time to reach stationarity : this is due to the fact that the algorithm is producing many primes , as the evolving value of @xmath17 reaches the unity .
it is however in a neighborhood of the transition where the system takes the higher times to reach the steady state : the system is producing many reactions , but not that many primes .
this fact can be related to a critical slowing down phenomenon , and is studied further in the text . + it is worth notting in figures [ r_y_v ] and [ r2 ] that in the disordered phase the order parameter does nt vanish , as it should .
this is due to the fact that in a pool of @xmath25 numbers , following the prime number theorem , one finds on average @xmath30 primes @xcite .
thus , there is always a residual contribution to the ratio @xmath31 not related to the system s dynamics which only becomes relevant for small values of @xmath3 , when the algorithm is not able to produce primes .
this feature can be understood according to figures ( [ series ] , [ concentraciones ] ) , where in the disordered phase the system rapidly reaches a frozen state . note that the ratio in this case increases from @xmath32 -residual ratio- to an asymptotic value of @xmath33 .
it is clear that the system in this phase is not able to produce primes , and that the residual ratio takes practically the whole contribution .
+ since an order parameter should vanish in the disordered phase and be non - null in the ordered phase , a new order parameter that should avoid these problems has to be defined .
the preliminary study suggests the presence of a phase transition that distinguishes a phase where the algorithm is able to reduce every element of the system into a prime from another where the system is frozen in a state of low prime concentration .
it is straightforward that the algorithm searches stochastically the suitable combinations through which initial numbers can react and reduce into simpler ones .
let us now see how this phase transition can be understood as a dynamical process embedded in a catalytic network having integer numbers as the nodes . consider two numbers of that network , say @xmath34 and @xmath35 ( @xmath36 ) .
these numbers are connected ( @xmath37 ) if they are exactly divisible , that is to say , if @xmath38 with c being an integer .
the topology of similar networks has been studied in @xcite , concretely in @xcite it is shown that this network exhibits scale - free topology @xcite : the degree distribution is @xmath39 with @xmath40 . in our system , fixing @xmath3 is equivalent to selecting a random subset of nodes in this network . if @xmath34 and @xmath35 are selected they may react giving @xmath38 ; in terms of the network this means that the path between nodes @xmath34 and @xmath35 is travelled thanks to the catalytic presence of @xmath41 .
we may say that our network is indeed a catalytic one @xcite where there are no cycles as attractors but two different stationary phases : ( i ) for large values of @xmath3 all resulting paths sink into primes numbers , and ( ii ) if @xmath3 is small only a few paths are travelled and no primes are reached .
notice that in this network representation , primes are the only nodes that have input links but no output links ( by definition , a prime number is only divisible by the unit and by itself , acting as an absorbing node of the dynamics ) . when the temporal evolution of this algorithm is explored for small values of @xmath3 , we have observed in figures [ series],[concentraciones ] that the steady state is reached very fast . as a consequence ,
there are only few travelled paths over the network and since @xmath3 is small the probability of catalysis is small as well , hence the paths ending in prime nodes are not travelled .
we say in this case that the system freezes in a disordered state .
in contrast when @xmath3 is large enough , many reactions take place and the network is travelled at large . under these circumstances ,
an arbitrary node may be catalyzed by a large @xmath42 quantity of numbers , its probability of reaction being high .
thus , in average all numbers can follow network paths towards the prime nodes : we say that the system reaches an ordered state .
+ in the light of the preceding arguments , it is meaningful to define a new order parameter @xmath43 as the probability that the system has for a given ( @xmath3,@xmath25 ) to reduce every number from @xmath3 into primes , that is to say , to reach an ordered state .
in practice , @xmath43 is calculated in the following way : given ( @xmath3,@xmath25 ) , for each realization we check , once stationarity has been reached , whether the whole set of elements are primes or not . in the first case , @xmath43 sums up one , in the second , it sums up zero .
we then average @xmath43 over every realization .
+ in figure [ p ] we plot @xmath43 versus @xmath3 , for different pool sizes @xmath25 .
the phase transition that the system exhibits has now a clear meaning ; when @xmath44 , the probability that the system has to be able to reduce the whole system into primes is null ( disordered state ) , and viceversa when @xmath45 .
+ in each case , @xmath46 , the critical value separating the phases @xmath44 and @xmath45 , can now be defined .
observe in figure [ p ] that @xmath47 increases with the pool size @xmath25 . in order to describe this size dependence ,
we need to find some analytical argument by means of which define a system s characteristic size . as we will see in a few lines , this
one wo nt be @xmath25 as one would expect in a first moment .
+ versus @xmath3 , for the same pool sizes as figure ( [ r2 ] ) ( averaged over @xmath16 realizations ) . note that @xmath43 is now a well defined order parameter , as long as @xmath48 $ ] .
again , @xmath47 depends on the pool size @xmath25.,scaledwidth=45.0% ] ( as explained in the text ) versus @xmath3 , for different values of pool size @xmath25 , in the annealed approximation.,scaledwidth=45.0% ] the system under hands shows highly complex dynamics : correlations take place between the @xmath3 numbers of the system at each time step in a non trivial way .
find an analytical solution to the former problem is thus completely out of the focus of this paper .
however , an annealed approximation can still be performed .
the main idea is to obviate these two - time correlations , assuming that at each time step , the @xmath3 elements are randomly generated .
this way , we can calculate , given @xmath3 and @xmath25 , the probability @xmath49 that at a single time step , no pair of molecules between @xmath3 are divisible .
thus , @xmath50 will be the probability that there exist at least one reacting pair .
note that @xmath50 will somehow play the role of the order parameter @xmath43 , in this oversimplified system .
+ in a first step , we can calculate the probability @xmath51 that two molecules randomly chosen from the pool @xmath25 are divisible : @xmath52 where the floor brackets stand for the integer part function . obviously , @xmath53 is the probability that two molecules randomly chosen are not divisible in any case .
now , in a system composed by @xmath3 molecules , we can make @xmath54 distinct pairs . however , these pairs are not independent in the present case , so that probability @xmath55 is nt simply @xmath56 . correlations between pairs must be somehow taken into account . at this point , we can make the following ansatz : @xmath57 where @xmath58 characterizes the degree of independence of the pairs .
the relation @xmath59 versus @xmath3 is plotted in figure [ campo ] for different values of the pool size @xmath25 .
note that for a given @xmath25 , the behavior of @xmath59 is qualitatively similar to @xmath43 , the order parameter in the real system .
+ for convenience , in this annealed approximation we will define a threshold @xmath47 as the one for which @xmath60 .
this value is the one for which half of the configurations reach an ordered state .
this procedure is usual for instance in percolation processes , since the choice of the percolation threshold , related to the definition of a spanning cluster , is somewhat arbitrary in finite size systems @xcite . taking logarithms in equation ( [ ansatz ] ) and expanding up to first order
, we easily find an scaling relation between @xmath47 and @xmath25 , that reads @xmath61^{\alpha } .\label{escala1}\ ] ] this relation suggests that the system s characteristic size is not @xmath25 , as one would expect in a first moment , but @xmath30 . in figure [ scaling1 ]
we plot , in log - log , the scaling between @xmath47 and the characteristic size @xmath30 that can be extracted from figure [ campo ] .
the best fitting provides a relation of the shape ( [ escala1 ] ) where @xmath62 ( note that the scaling is quite good , what gives consistency to the leading order approximations assumed in equation [ escala1 ] ) . versus the system s characteristic size in the annealed approximation .
the plot is log - log : the slope of the straight line provides the exponent @xmath63 of equation ( [ escala1]).,scaledwidth=45.0% ] versus the characteristic system s size @xmath30 in the prime number generator , for pool size @xmath64 .
the plot is log - log : the slope of the curve provides an exponent @xmath65.,scaledwidth=45.0% ] the annealed approximation introduced in the preceding section suggests that the characteristic size of the system is not @xmath25 as one would expect but rather @xmath30 .
this is quite reasonable if we have in mind that the number of primes that a pool of @xmath25 integers has is on average @xmath30 @xcite : the quantity of primes does nt grow linearly with @xmath25 .
this is the so called prime number theorem , and states that the number of primes in the set of integers @xmath66 is asymptotic with @xmath30 when @xmath25 is large enough .
+ in order to test if this conjecture also applies to the prime number generator , in figure [ scaling2 ] we represent ( in log - log ) the values of @xmath47 ( obtained numerically from the values where @xmath67 becomes non - null for the first time ) as a function of @xmath30 .
we find the same scaling relation as for the annealed system ( equation [ escala1 ] ) , but with a different value for @xmath68 .
this little disagreement is logical and comes from the fact that in the annealed approximation correlations are obviated .
+ let us apply generic techniques of finite size scaling in order to calculate the critical exponents of this phase transition . reducing the control parameter as @xmath69
, the correlation exponent @xmath70 is defined as : @xmath71 where we find @xmath72 and @xmath73 .
note that the transition tends to zero in the thermodynamical limit because its value increases more slowly than the system s size . ) for different values of @xmath25 , assuming the scaling relation [ escala1 ] .
the collapse is very good , the scaling relation seems to be consistent .
, scaledwidth=35.0% ] the critical exponent of the order parameter @xmath74 can be deduced from the calculation of the correlation exponent @xmath70 and from the finite size scaling relation @xmath75 the best fitting provides a value of @xmath76 .
+ in figure [ colapso ] we have collapsed all curves @xmath67 according to the preceding developments .
note that the collapse is excellent , something which provides consistency to the full development .
hitherto , we have seen that the dynamical process that the prime number generator shows gives rise to a continuous phase transition embedded in a direct catalytic network . as pointed out in @xcite , phase transitions quite similar to the former one as percolation processes for instance
can be easily related to search problems . in the case under study
we can redefine the percolation process as a decision problem in the following terms : one could ask when does the clause _ every number of the system is prime when the algorithm reaches stationarity _ is satisfied .
it is clear that through this focus , the prime number generator can be understood as a sat - like problem , as long as there is an evident parallelism between the satisfiability of the preceding clause and our order parameter p. thereby , in order to study the system from the focus of computational complexity theory , we must tackle the following questions : what is the algorithmic complexity of the system ? and
how is related the observed phase transition to the problem s tractability ?
the algorithm under study , which generates primes by stochastic decomposition of integers , is related to both primality test and integer decomposition problems .
although primality was believed to belong to the so - called @xmath1 problems @xcite ( solvable in non - deterministic polynomial time ) , it has recently been shown to be in @xmath43 @xcite : there exists at least an efficient deterministic algorithm that tests if a number is prime in polynomial time .
the integer decomposition problem is in turn a harder problem , and to find an algorithm that would factorize numbers in polynomial time is an unsolved problem of computer science .
furthermore , exploring the computational complexity of the problem under hands could eventually shed light into these aspects .
+ for that task , let us determine how does the search space grows when we increase @xmath3 . in a given time step , the search space corresponds to the set of configurations that must be checked in order to solve the decision problem : this is nothing but the number of different pairs that can be formed using @xmath3 numbers . applying basic combinatorics ,
the set of different configurations @xmath77 for @xmath3 elements and @xmath78 pairs is : @xmath79 we get that the search space increases with @xmath3 as @xmath80 . on the other hand , note that the decision problem is rapidly checked ( in polynomial time )
if we provide a candidate set of @xmath3 numbers to the algorithm .
these two features lead us to assume that the problem under hands belongs , in a worst - case classification @xcite , to the @xmath1 complexity class .
+ note that this is not surprising : the preceding sections led us to the conclusion that the process is embedded in a ( dynamical ) scale - free catalytic network .
as a matter of fact , the phase transition is related to a dynamical process embedded in a high dimensional catalytic network . in this hallmark , it is straightforward that this underlying network is non - planar @xcite .
now , it has been shown that non - planarity in this kind of problems usually leaves to np - completeness @xcite ( for instance , the ising model in two - dimensions is , when the underlying network topology is non - planar , in @xmath1 ) . as defined in the text versus @xmath3 , for different pool sizes , from left to right : @xmath81 .
every simulation is averaged over @xmath16 realizations .
note that for each curve and within the finite size effects @xmath82 reaches a maximum in a neighborhood of its transition point ( this can be easily explored in figure [ p]).,scaledwidth=45.0% ] for the curves of figure [ t_sin_colapso ] .
the goodness of the collapse validates the scaling relations.,scaledwidth=45.0% ] an ingredient which is quite universal in the algorithmic phase transitions is the so called _ easy - hard - easy pattern _
@xcite : in both phases , the computational cost of the algorithm ( the time that the algorithm requires to find a solution , that is , to reach stationarity ) is relatively small
. however , in a neighborhood of the transition , this computational time reaches a peaked maximum . in terms of search or decision problems , this fact has a clear interpretation : the problem is relatively easy to solve as long as the input is clearly in one phase or the other , but not in between . in the system under study , the algorithm is relatively fast in reaching an absorbing state of low concentration of primes for small @xmath3 because the probability of having reactions is small . in the other hand ,
the algorithm is also fast in reaching an absorbing state of high concentration of primes for high @xmath3 , because the system has enough `` catalytic candidates '' at each time step to be able to reduce them , the probability of having reactions is high . in the transition s vicinity ,
the system is critical .
reactions can be achieved , however , the system needs to make an exhaustive search of the configuration space in order to find these reactions : the algorithm requires in this region much more time to reach stationarity . + note that this easy - hard - easy pattern is related , in second order phase transitions , to the the phenomenon of critical slowing down , where the relaxation time in the critical region diverges @xcite .
+ we have already seen in figure [ series ] that the system reaches the steady state in a different manner , depending on which phase is located the process . more properly ,
when @xmath83 ( disordered phase ) , the system rapidly frozens , without practically achieving any reaction .
when @xmath27 ( ordered phase ) , the system takes more time to reach the steady state , but it is in the regime @xmath26 where this time is maximal .
in order to be able to properly compare these three regimes , let us define a characteristic time in the system @xmath84 as the number of average time steps that the algorithm needs to take in order to reach stationarity
. remember that we defined a time step @xmath85 as @xmath3 monte carlo steps ( @xmath3 operations ) .
thus , in order to normalize over the number of molecules , it is straightforward to define a characteristic time as : @xmath86 note that @xmath84 can be understood as a measure of the algorithm s time complexity @xcite . in figure [ t_sin_colapso ]
we plot @xmath84 versus @xmath3 for a set of different pools @xmath87 ( simulations are averaged over @xmath16 realizations ) .
note that given a pool size @xmath25 , @xmath84 reaches a maximum in a neighborhood of its transition point @xmath46 , as can be checked according to figure [ p ] .
as expected , the system exhibits an easy - hard - easy pattern , as long as the characteristic time @xmath84 required by the algorithm to solve the problem has a clear maximum in a neighborhood of the phase transition .
moreover , the location of the maximum shifts with the system s size according to the critical point scaling found in equation [ escala1 ] . in the other hand , this maximum also scales as : @xmath88 where the best fitting provides @xmath89 .
note that in the thermodynamic limit , the characteristic time would diverge in the neighborhood of the transition .
it is straightforward to relate this parameter with the relaxation time of a physical phase transition .
according to these relations , we can collapse the curves @xmath90 of figure [ t_sin_colapso ] into a single universal one . in figure [ colapse_t ] this collapse
is provided : the goodness of the former one supports the validity of the scaling relations .
the system under study is interpreted in terms of a search problem , belonging to the @xmath1 class in a worst - case classification .
now , an average - case behavior , which is likely to be more useful in order to classify combinatorial problems , turns out to be tough to describe . in @xcite , monasson et al .
showed that there where @xmath1 problems exhibit phase transitions ( related to dramatic changes in the computational hardness of the problem ) , the order of the phase transition is in turn related to the average - case complexity of the problem . more specifically , that second order phase transitions are related to a polynomial growing of the resource requirements , instead of exponential growing , associated to first order phase transitions .
+ it has been shown that the system actually exhibits a second order phase transition and an easy - hard - easy pattern . following monasson et al .
@xcite , while our prime generator is likely to belong to the @xmath1 class , it shows however only a polynomial growing in the resource requirements , in the average case .
one may argue that one of the reasons of this hardness reduction is that the algorithm does nt realize a direct search but on the contrary this search is stochastic : the search space is not exhaustively explored .
thereby , the average behavior of the system and thus the average decision problem can be easily solved by the algorithm , in detriment of the probable character of this solution .
in this paper a ( stochastic ) algorithmic model which stands for a prime number generator has been studied .
this model exhibits a phase transition which distinguishes a phase where the algorithm has the ability to reduce every element into a prime , and a phase where the system is rapidly frozen .
analytical and numerical evidences suggest that the transition is continuous . on a second part
, the model has been reinterpreted as a search problem .
as long as the model searches paths to reduce integers into primes , the combinatorial problem is related to primality test and decomposition problem .
it has been shown that this model belongs to the @xmath1 class in a worst - case classification , moreover , an easy - hard - easy pattern has been found , as common in many algorithmic phase transitions . according to the fact that the transition is continuous , and based on previous works
, it has been put into relevance that the average - case complexity may be only polynomial .
this hardness reduction is in turn related to the fact that the algorithm only yields probable states .
the authors wish to thank the instituto de fisica at unam for support and hospitality .
the research was supported by unam - dgapa grant in-118306 ( om ) , grants fis2006 - 26382-e and fis2006 - 08607 ( bl and ll ) and the _
tomas brody lectureship 2007 _ awarded to bl .
30 this is something quite intuitive if we take into account that there are typically @xmath78 multiples of two , @xmath91 multiples of three , and so on : the probability that the prime @xmath23 appears in a random composite number is in average @xmath24 . | we introduce a prime number generator in the form of a stochastic algorithm .
the character of such algorithm gives rise to a continuous phase transition which distinguishes a phase where the algorithm is able to reduce the whole system of numbers into primes and a phase where the system reaches a frozen state with low prime density . in this paper
we firstly pretend to give a broad characterization of this phase transition , both in terms of analytical and numerical analysis .
critical exponents are calculated , and data collapse is provided .
further on we redefine the model as a search problem , fitting it in the hallmark of computational complexity theory .
we suggest that the system belongs to the class _ np_. the computational cost is maximal around the threshold , as common in many algorithmic phase transitions , revealing the presence of an easy - hard - easy pattern .
we finally relate the nature of the phase transition to an average - case classification of the problem . |
the interplay between the magnetism and transport properties is one of the main issues in the physics of the strongly correlated electronic systems . doped manganites r@xmath11a@xmath12mno@xmath13 ( r = la , pr , nd , sm ; a= ca
, sr , ba ) have recently attracted considerable interests from this viewpoint in connection to the colossal magnetoresistance ( cmr ) .
it has been believed that the most fundamental mechanism for this cmr is the double exchange interaction , which has been studied long time ago @xcite .
however recent systematic experimental studies have revealed the vital role of the orbital degrees of freedom , which is the subject of this paper . in the parent compound rmno@xmath13 , the valence of the mn ion is mn@xmath14 with the configuration @xmath15 . due to the crystal field
the five @xmath10-orbitals split into two @xmath0-orbitals @xmath16 and three @xmath17-orbitals @xmath18 . according to hund s rule
the four electrons occupy three @xmath17-orbitals and one of the @xmath0-orbitals with parallel spins
. then the @xmath0-electron has the orbital degrees of freedom in addition to the spin ones .
when the holes with concentration @xmath1 are doped , the vacancies in the @xmath0-orbitals are introduced which induces the ferromagnetism due to the double exchange interaction and the charge transport .
however there are several experimtental facts which can not be explained in the simple double exchange models .
one is the rich phase diagrams which is studied theoretically in section ii and the other is the anomalous features in the ferromagnetic metallic phase described in section iii .
all these issues are colsely related to the orbital degrees of freedom , which control the dimensioanlity of the electronic systems and whose quantum fluctuation gives rise to the orbital liquid and orbital kondo effect .
section iv is devoted to the conclusions where the possibility and implications of the dynamical phase separation are mentioned .
we set up the three - dimensional cubic lattice consisting of the manganese ions .
two kinds of the @xmath0 orbital ( @xmath19 ) are introduced in each site , and the @xmath17 electrons are treated as the localized spin with @xmath20 .
the hamiltonian without the jt coupling is given as follows @xcite , @xmath21 @xmath22 is the operator which creates an electron with spin @xmath23 in orbital @xmath24 at site @xmath25 , and @xmath26 is the spin operator for the @xmath0 electron defined by @xmath27 .
the first term describes the electron transfer between @xmath28 orbital at site @xmath25 and @xmath29 orbital at the nearest - neighbor site @xmath30 .
the value of @xmath31 is estimated by considering the oxygen @xmath32 orbitals between the nearest mn - mn pair , and is represented by @xmath33 , where @xmath34 is the numerical factor depending on the orbitals @xcite .
@xmath35 is estimated to be @xmath36 , which we choose the unit of energy below ( @xmath37 ) @xcite .
the second line shows the electron - electron interaction terms where @xmath8 , @xmath38 and @xmath39 is the intra- , inter - orbital coulomb interactions , and inter - orbital exchange interaction , respectively .
this interaction can be rewritten as @xmath40 @xcite . here
the spin operator @xmath41 and the iso - spin operator @xmath42 for the orbital degrees of freedom are introduced , and the two positive coefficients @xmath43 and @xmath44 , which are defined by @xmath45 and @xmath46 , represent the interaction to induce the spin and iso - spin moments , respectively . the last line is the sum of the hund coupling and the af interaction between the nearest neighboring @xmath17 spins .
here we adopt the mean field approximation by introducing the order parameters @xmath47 , @xmath48 , and @xmath49 .
these order parameters are determined to optimize the mean field energy at zero temperature . for both spin and orbital ,
the four types of the ordering are considered , that is , the ferromagnetic ( f - type ) ordering , where the order parameters are uniform , and the three af - like orderings , i.e. , the layer - type ( a - type ) , the rod - type ( c - type ) and the nacl - type ( g - type ) af orderings .
hereafter , types of the orderings are termed as , for example , ( spin : c ) , and so on . in fig . 1 , we present the calculated phase diagram in the case of @xmath50 , @xmath51 , which is relevant to the actual manganese oxides . in each region
, the spin structure is specified by the letter f , a , c , g and the orbital configuration by picture .
the important results are the followings .
[ 1 ] there occurs the crossover from the superexchange dominated region for small @xmath1 ( @xmath52 ) to that of the double exchange interaction for large @xmath1 ( @xmath53 ) .
the former region could not be reproduced by the analysis without coulombic repulsion . in this respect ,
the spin f state up to @xmath54 in the phase diagram is mostly due to the superexchange interaction . [ 2 ] because of the anisotropy of the transfer integral between the @xmath0-orbitals , the most natural orbital configuration expected from the double exchange interaction is @xmath55 with the two dimensional electronic dispersion .
this corresponds to the a phase for @xmath56 . in this case
the interlayer transfer is forbidden , and the spin a - type sturucture is stabilized by @xmath57 .
no spin canting is expected in this case . [ 3 ] similarly the spin c structure is accompanied by the @xmath58 orbital which gives the 1d - like dispersion along the @xmath59-axis .
hence the spin and orbital structure is closely correlated through the dimensionality of the electronic dispersion . from this viewpoint ,
the orbitals have more degrees of freedom in the ferromagnetic state .
this suggests the large orbital fluctuation in the spin f state as discussed in the next section .
0.3 cm we now discuss the comparison between the mean field phase diagram in fig . 1 and the experiments . at the moment
, a value of @xmath60 can not be estimated accurately , but there are two rough estimates .
one is from the nel temperature @xmath61 for camno@xmath13 @xmath62 @xcite , which suggest @xmath63 in the mean field approximation .
the fluctuations lower @xmath64 , and hence increase the estimate for @xmath60 .
another estimate is obtained from the numerical calculations for lamno@xmath13 @xmath65 , which suggests @xmath66 @xcite .
although @xmath60 might depend on @xmath1 in real materials , we tentatively fix @xmath60 to be @xmath67 represented by the broken line in fig .
then the spin structure changes as a @xmath68 f @xmath68 a@xmath69 c @xmath69 g , as @xmath1 increases , which is in good agreement with the experiments @xcite .
another implication to the experimental results is about spin : a phase appearing around @xmath70 . in @xmath71 , the ferromagnetic metallic phase is realized up to about @xmath72 and the ce - type af structure with the charge ordering tunes up @xcite . with further increasing of @xmath1 , the metallic state with spin
: a again appears at about @xmath73 , and the large anisotropy in the electrical resistivity is observed in this phase .
the similar metallic phase accompanied with spin : a is also reported in @xmath74 @xcite .
the spin canting is not observed experimentally @xcite as predicted above in this spin : a metal , in contrast to the spin : a in the small @xmath1 region which has been discussed by de gennes @xcite . summarizing this section
we have studied the phase diagram of la@xmath11sr@xmath12mno@xmath13 in the plane of @xmath1 ( hole concentration ) and @xmath57 ( af exchange interaction between the @xmath17 spins ) in the mean field approximation .
the global features can be understood in terms of the interplay between the superexchange interaction and the double exchange one which is considerably modified with taking the orbital degrees of freedom into account .
the dimensionality of the energy band attributed to the orbital structure plays an essential role to determine the phase boundary .
the orbital structure in the ferromagnetic phase is sensitive to changes of the carrier concentration and the energy parameters .
this is because the spin structure does not forbid the kinetic energy gain for any direction , and the orbitals have more freedom .
this enhances the fluctuation of the orbital degrees of freedom .
actually the orbital fluctuation has been studied in the slave - fermion formalism , and it has been found that the orbitals remain a liquid down to the low temperature due to the fluctuation among the three directions of the @xmath75-type ordering @xcite . what quenches the entropy of the system is the quantum tunneling between the different orbital configurations , which could be regarded as the orbital - kondo effect .
this will be studied in the next section .
experimentally there are several anomalous features in the low temperature ferromagnetic phase below @xmath76 . [
a ] in the neutron scattering experiment no temperature dependent phonon modes have been observed @xcite .
the recent raman scattering experiment also shows that the jahn - teller phonons ( especially their frequencies ) are temperature independent and insensitive to the ferromagnetic transition at @xmath76 @xcite .
[ b ] the photo - emission spectra show a small discontinuity at the fermi edge even at @xmath77 , which suggests some interactions still remain strong there @xcite .
[ c ] the optical conductivity @xmath78 at @xmath77 is composed of two components , i.e. , the narrow drude peak ( @xmath79 ) and the broad incoherent component extending up to @xmath80 @xcite .
the drude weight is very small , which seems to be consistent with the photo - emission spectra .
[ d ] the low temperature resistivity @xmath81 can be fitted by @xmath82 where @xmath4 is a large constant of the order of 500@xmath83 @xcite , again suggesting the strong electron correlation .
[ e ] contrary to the case of resistivity [ d ] , the coefficient of @xmath84-linear specific heat is very small with @xmath85 , which violates the kadowaki - woods law for these compounds @xcite .
although [ e ] is difficult to reconcile with [ a]-[d ] , we consider the latter as the evidence for the strong correlation and orbital fluctuation even at @xmath86 . because the spins are perfectly aligned at @xmath86 , the only remaining degrees of freedom are the orbital ones . in this section ,
we study a large @xmath10 model @xcite for the ferromagnetic state including both the electron - electron interaction @xmath8 and the jahn - teller coupling @xmath87 @xcite .
this model takes care of the following two respects , ( i ) including the electron - electron interation , ( ii ) including the quantum fluctuations .
especially the latter is essential to describe the low temperature fermi liquid state , which is described by the kondo peak in the large @xmath10 limit @xcite .
the strong electron - electron interaction with the reasonable magnitude of the jahn - teller coupling explains both the large isotope effect and [ a ] .
moreover , the features of strong correlation [ b]-[d ] are at least consistent with the large @xmath8 picture although [ e ] still requires another physics , which we will discuss in the next section .
we start with the hubbard - holstein model for the ferromagnetic state .
@xmath88 where the spin indices @xmath89 correspond to the orbital degrees of freedom as @xmath90 and @xmath91 , and the real spins are assumed to be perfectly aligned .
we consider only one jahn - teller displacement mode @xmath92 for each site , while there are two modes @xmath93 in the real perovskite structure @xcite .
however this does not change the qualitative features obtained below .
the jahn - teller mode is assumed to be an einstein phonon with a frequency @xmath94 .
the transfer integral @xmath95 depends on a pair of orbitals @xmath96 and the hopping direction .
these dependences lead to the various low lying orbital configurations , and thus they suppress the orbital orderings in the ferromagnetic state .
actually there is no experimental evidence for the orbital orderings , e.g. , the anisotropies of the lattice constants and/or transport properties in the ferromagnetic state .
then we assume that no orbital ordering occurs down to the zero temperature due to the quantum fluctuations , and the transfer integral is assumed to be diagonal for simplicity , i.e. , @xmath97 .
this means that the ground state is a fermi liquid with two degenerate bands at the fermi energy . in order to describe this fermi liquid state
, we employ the large-@xmath10 approach where the model eq.(3 ) is mapped to the impurity anderson model with the self - consistent condition @xcite .
this approach has been applied to the manganese oxides to study hund s coupling by furukawa @xcite and to study the additional jahn - teller coupling by millis _
the action for the effective impurity model is given by @xmath98 + \int_0^\beta d \tau { m \over 2 } \biggl [ ( \partial_\tau q(\tau ) ) ^2 + \omega^2 q(\tau)^2 \biggr],\end{aligned}\ ] ] where @xmath99 is the dynamical weiss field representing the influence from the surrounding sites .
the self - consistency condition is that the on - site green s function @xmath100 should be equal to the hilbert transform of the density of states @xmath101 as @xmath102 we take the lorentzian density of states @xmath103 because there is no need to solve the self - consistency equation in this case , and the weiss field is given by @xmath104 the only unknown quantity is the chemical potential @xmath105 , which is determined by the electron number .
the determination of chemical potential requires some numerical calculations , but our discussion below is not sensitive to the location of the chemical potential
. then we take @xmath105 as the parameter of our model , and the problem is now completely reduced to that of a single impurity .
now we introduce a stratonovich - hubbard ( sh ) field @xmath106 to represent the coulomb interaction @xmath8 .
then the action is given by @xmath107 + \zeta(\tau ) q(\tau ) - ( \delta \xi(\tau ) + g q(\tau ) ) ( n _ { \uparrow}(\tau ) - n _ { \downarrow}(\tau ) ) \biggr],\end{aligned}\ ] ] where @xmath108 and @xmath109 is the test field to measure the phonon correlation function . at this stage the electron is coupled with the linear combination of the sh field @xmath110 and the phonon @xmath111 as @xmath112 . then the effective action can be derived as @xmath113 \nonumber \\ & + & \sum_{\omega_n } { 1 \over { 2m(\omega_n^2 + { \tilde \omega}^2 ) } } \biggl [ - \zeta(i \omega_n ) \zeta(- i \omega_n ) + { { 2 \delta^2 } \over u } m(\omega_n^2 + \omega^2 ) \eta(i \omega_n ) \eta(- i \omega_n ) \nonumber \\ & + & { { 2 g \delta } \over u } ( \zeta(i \omega_n ) \eta(- i \omega_n ) + \eta(i \omega_n ) \zeta(- i \omega_n ) ) \biggr],\end{aligned}\ ] ] where @xmath114 .
the effective interaction @xmath115 is the sum of the coulomb repulsion @xmath8 and the lattice relaxation energy @xmath116 as @xmath117 .
the phonon green s function @xmath118 is given by @xmath119 where @xmath120 is the orbital susceptibility of the electronic system .
the first term is the usual phonon green s function with the renormalized frequency @xmath121 , which is higher than the bare @xmath94 .
this @xmath121 does not depend on the electron response function and hence temperature independent .
when @xmath122 , @xmath121 is much larger than @xmath94 and the first term of eq.(9 ) is irrelevant in the phonon frequency region @xmath123 . in the opposite limit @xmath124
, the renormalization of the phonon frequency is small .
the second term depends on the temperature ( kohn anomaly ) and shows the characteristic lineshape with the broadening . from the above consideration together with the experimental fact that no temperture dependence of the phonon frequency has been observed , we conclude that the coulomb interaction @xmath8 is larger than the lattice relaxation energy @xmath9 .
now let us study the electronic system in detail .
following the analysis in section iii c of ref.@xcite , the quantum fluctuation between two orbital configurations gives rise to the orbital kondo effect with the kondo temperature @xmath125 .
we obtain an estimate for the kondo temperature as @xmath126.\ ] ] where @xmath127 here we have introduced the dimensionless variables @xmath128 and @xmath129 .
the main conclusion here is that when @xmath130 , i.e. , @xmath131 , the reduction factor @xmath132 is proportional to @xmath1 even in the limit @xmath133 .
this is because the overlap of the phonon wavefunction enters the tunneling matrix element even when the coulomb interaction is dominating .
this kondo temperature @xmath125 gives the effective bandwidth @xmath134 in the large-@xmath10 model , and the effective mass enhancement is estimated as @xmath135 $ ] .
this strong mass enhancement manifests itself in the physical quantities as follows .
the orbital susceptibility @xmath136 has a peak at @xmath137 with the peak height of the order of @xmath138 and with the width of the order of @xmath125 .
the specific heat coefficient @xmath28 should be proportional to @xmath139 , while the coefficient @xmath4 of @xmath140 in eq.(2 ) scales as @xmath141 .
the ferromagnetic transition temperature @xmath76 is expected to be proportional to @xmath134 , and hence we expect a large isotope effect when @xmath142 . in other words ,
the large isotope effect does not rule out the large coulomb interaction @xmath8 .
the edge of the photoemission spectrum at the fermi energy and the drude weight are reduced by the factor @xmath143 . as mentioned above ,
many experiments except for those of specific heat seem to suggest that the manganese oxides in their ferromagnetic state belong to the strongly correlated region .
then it is natural that the large isotope effect is observed .
the band width is rather difficult to pin down accurately , but a rough estimate from @xmath4 in eq.(2 ) is @xmath144 , which is comparable to @xmath94 .
this suggests that the suppression is not extremely large .
another experimental fact is that the temperature independent jahn - teller phonon is observed at the frequency near the original one @xcite , which means that @xmath8 is larger or at least of the order of @xmath9 .
then we conclude that the relative magnitude of the energy scales is @xmath145 for the ferromagnetic phase in the manganese oxides .
the small specific heat coefficient necessitates the physical mechanisms which have not been included in our model , which will be discussed in the next section . summarizing this section ,
we have studied a model of manganese oxides in the ferromagnetic state taking into account both the coulomb repulsion @xmath8 and the jahn - teller interaction @xmath9 .
these two interactions collaborate to induce the local orbital moments . in the strong coupling case ,
i.e. , @xmath146 ( @xmath147 transfer integral ) , the overlap of the phonon wavefunctions ( @xmath148 ) enters the tunneling amplitude between the two minima for the orbital moment .
the phonon spectral function consists of two parts , i.e. , the sharp peak at the renormalized frequency @xmath149 and the broad peak with the width of the order of the band width .
these results can reconcile the large isotope effect on @xmath76 and the apparent temperature independent phonon spectrum assuming @xmath150 .
we have studied the vital role of the orbitls in manganese oxides .
the most important ingredient is the strong coulombic interaction , which induces the orbital polarization .
this polarization controls the dimensionality of the electronic system when ordered . in the isotropic ferromagnetic state , large orbital fluctuation
is expected which gives rise to the quantum liquid state below the orbital kondo temperature @xmath125 .
> from this viewpoint , the double exchange interaction is switched on below the orbital kondo temperature @xmath125 , and the ferromagnetic transition temperature @xmath76 is roughly identified with @xmath125 .
this scenario might explain the photoemission experiment which shows that the discontinuity at the fermi level grows only below @xmath76 .
however there still remain several puzzles as follows .
( i ) the small specific heat , ( ii ) violation of the kadowaki - woods relation , ( iii ) absence of the quasi - particle peak in the angle - resolved photoemission spectra in double layered manganese oxides , ( iv ) the residual resistivity larger than the mott limit .
one possible explanation for these is the ( dynamical ) phase separation .
the phase separation occurs when the total energy @xmath151 as a function of the hole concentration @xmath1 has more than two minima and it is possible to draw a common tangential line .
this is found to occur for several models .
here we want to point out that when the bandwidth strongly depends on the orbital configuration , the @xmath1-dependence of it naturally gives rise to the phase separation . in reality
the long range coulomb interaction will suppress the phase separation , but which still remain as a fluctuation , i.e. , dynamical one .
suppose phase separation occurs , and assume that @xmath152 is the metallic fraction while @xmath153 is the insulating fraction .
now we relate the observed physical quantities to the metallic ones and the fraction @xmath152 .
the @xmath84-linear specific heat @xmath154 is reduced by the factor @xmath152 , and hence @xmath155 . concerning the resistivity
, the percolating path will lower the capacitance energy when the voltage is applied , and the observed resistivity is enhanced by the factor @xmath156 , i.e. , @xmath157 .
therefore the kadowaki - woods ratio is related to @xmath158 the observed ratio is @xmath159 times larger than the universal ratio , and if we attribute this discrepancy to the factor @xmath152 , we obtain @xmath160 .
this also suggests @xmath161 , which means that the mass enhancement is @xmath162 . furthermore the drude wieght is reduced by @xmath152 in addition to the mass enhancement , and is expected to be about @xmath163 of the noninteracting value in agreement with experiment .
the residual resistivity is enhanced by @xmath156 and can be larger than the mott limit . roughly speaking , the hall constant , on the other hand ,
is not modified by the phase separation .
the extremely large noise might be also related to the dynamical fluctuation of the phase separation .
thus the phase separation seems to explain many puzzles in the manganese oxides .
however the detailed studies including the time / length scale of the fluctuation are left for future investigations .
the authours would like to thank y.tokura , a. millis , j. ye , g. kotliar , q. si , h. kuwahara , s. maekawa , h.fukuyama , k. miyake , t.okuda , k.yamamoto and y.endoh for valuable discussions .
this work was supported by coe and priority areas grants from the ministry of education , science , culture and sports of japan .
part of this work has been done in apctp workshop and aspen center for physics , and we acknowledge their hospitalities . 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 , j. inoue , and s. maekawa , physica c * 263 * , 130 ( 1996 ) , and phys .
b * 55 * , 8280 ( 1997 ) . s. ishihara , m. yamanaka , n. nagaosa , phys . rev .
b * 56 * , 686 ( 1997 ) .
r. maezono , s. ishihara and n. nagaosa , phys .
b * 57 * , r21822 ( 1998 ) .
r. maezono , s. ishihara and n. nagaosa , to be published in phys .
b * 58 * , november ( 1998 ) .
e. o. wollan , and w. c. koehler , phys
* 100 * , 545 ( 1955 ) .
h. kawano , r. kajimoto , h. yoshizawa , y. tomioka , h. kuwahara , and y. tokura , phys .
. lett . * 78 * , 4253 ( 1997 ) .
h. kuwahara , t. okuda , y. tomioka , t. kimura , a. asamitsu and y. tokura , mat .
. proc . * 494 * , 83 ( 1998 ) .
y. tomioka ( unpublished ) .
y. moritomo , t. akimoto , a. nakamura , k. ohoyama , and m. ohashi , phys .
b * 58 * , 5544 ( 1998 ) .
m.c.martin , g.shirane , y.endoh , k.hirota , y.moritomo , y.tokura , phys . rev .
* 53 * , 14285 ( 1996 ) .
k.yamamoto and y.tokura , private communications . d.d.sarma _
et al . _ ,
rev . b*53 * , 6874 ( 1996 ) .
et al . _ ,
57 * , 109 ( 1995 ) .
et al . _ ,
b*51 * , 11103 ( 1995 ) .
b.woodfield , m.l.wilson , and j.m.byers , phys . rev
. lett . * 78 * , 3201 ( 1997 ) ; t.okuda and y.tokura , private communications . for
a review see a.georges _ et al .
* 68 * , 13 ( 1996 ) .
n. nagaosa , s. murakami , and h.c.lee , phys .
b * 57 * , r6767 ( 1998 ) .
j.kanamori , j. appl .
phys.*31 * , 14s ( 1960 ) .
n.furukawa , j. phys .
jpn * 63 * , 3214 ( 1994 ) .
a.j.millis , r.mueller , and b.i.shraimain , phys .
* , 5389 ; 5405 ( 1996 ) .
d.r.hamann , phys . rev . * 2 * , 1373 ( 1970 ) .
+ figure 1 .
mean field phase diagram as a function of the carrier concentration ( @xmath1 ) and the antiferromagnetic interaction between @xmath17 spins ( @xmath60 ) .
the energy parameters are chosen to be @xmath164 and @xmath165 .
schematic orbital structure in the each phase is also shown .
dotted line ( @xmath60 = 0.009 ) well reproduces the change of the spin structure experimentally observed . | we study the manganese oxides from the viewpoint of the strongly correlated doped mott insulator .
the magnetic ordering and the charge transport are governed by the orbital degrees of freedom , and their dimensionality is controlled by the anisotropic transfer integrals between the @xmath0 orbitals . as @xmath1 increases the magnetic structure is predicted to change as @xmath2 ( @xmath3 : ferromagnet , @xmath4 : layered antiferromagnet , @xmath5 : rod - type antiferromagnet , @xmath6 : usual antiferromagnet ) , in agreement with experiments .
especially the orbital is aligned as @xmath7 in the metallic @xmath4 state , which explains the quasi 2d transport and no canting of the spin observed experimentally .
next we discuss the ferromagnetic state without the orbital ordering due to the quantum fluctuation . here the interplay between the electron repulsion @xmath8 and the jahn - teller electron - phonon interation @xmath9 is studied with a large @xmath10 model .
in addition to this strong correlation , we propose that the dynamical phase separation could explain the specific heat as well as the various anomalous physical properties , e.g. , resistivity , photo - emission , etc . |
the crumpling transition of membranes is an interesting topic in the softmatter physics as well as in the biological physics @xcite .
a well - known model for such transition is the surface model of helfrich , polyakov , and kleinert @xcite .
a considerable number of theoretical and numerical studies have been devoted to reveal the phase structure of the model @xcite .
recently , it was shown by monte carlo ( mc ) simulations that the model undergoes a first - order transition on spherical and fixed connectivity surfaces @xcite , and the transition is universal @xcite .
the vertices can move only locally on the surface because of the fixed connectivity nature in those surface models .
however , the crumpling transition is not yet clearly understood in biological membranes .
if we consider the possibility of the transition in the cell membranes , we should take account of the fluid nature such as the lateral diffusion of lipids . conventionally , the free diffusion of lipids has been realized by the dynamical triangulation technique in the surface model@xcite .
the diffusion of lipids corresponds to the reparametrization invariance of the model , because the diffusion has no cost in energy on fluid surfaces . in the cell membranes ,
however , the free diffusion of lipids is suffered from heterogeneous structures .
lipid molecules of the cell membranes are known to undergo the so - called hop diffusion over the surface , which was recently found experimentally @xcite .
the free diffusion of lipids is prohibited due to the cytoskeleton .
the diffusion rate is , therefore , 10 - 100 times lower than that of artificial membranes , which is usually homogeneous and has no such domain structure .
moreover , some artificial membranes are considered to have skeletons , because they are partly polymerized @xcite .
motivated by this fact observed in the cell membranes , we study in this letter a dynamically triangulated surface model with compartmentalized domains whose boundaries are consisting of triangle edges ( or bonds ) that are not to be flipped .
the diffusion is constrained so that vertices can diffuse only inside each domain , and hence the vertices never jump across from one domain to the other domains .
nevertheless , we consider that such constrained lateral diffusion can simulate the hop diffusion in the cell membranes as a first approximation .
by dividing every edge of the icosahedron into @xmath0-pieces of the uniform length , we have a triangulated surface of size @xmath1 (= the total number of vertices ) .
the starting configurations are thus characterized by @xmath2 and @xmath3 , where @xmath4 is the total number of vertices with the co - ordination number @xmath5 .
the compartmentalized structure is built on the surface by keeping the boundary bonds unflipped in the mc simulations with dynamical triangulation .
the boundary of the compartment is constructed from a sequence of bonds that remain unflipped .
the total number @xmath6 of the compartments depends on the surface size @xmath7 .
we fix @xmath8 the total number of vertices inside a compartment to the following three values : @xmath9 as a consequence , @xmath6 is increased with the increasing @xmath7 .
the reason why we fix @xmath8 is that the size of compartment is considered to be finite , and then it is expected that total number of lipids in the compartment also remains finite in the cell membranes .
we must emphasize that the finiteness of @xmath8 is physically meaningful rather than the value of @xmath8 , because we do not always have one to one correspondence between the vertices and the lipid molecules .
figures [ fig-1](a),(b ) show surfaces of @xmath10 and @xmath11 for the starting configurations of mc simulations .
thick lines denote the compartment boundary consisting of the bonds that are not to be flipped .
vertices on the boundary of compartments can locally fluctuate , and they are prohibited from the diffusion .
the other vertices freely diffuse only inside the compartment . 0.1 in ( 0,0 )
( 0,0 ) ( 6,-1.5)(0,0)(a ) @xmath10 ( 28,-1.5)(0,0)(b ) @xmath11 and ( b ) @xmath11 , where @xmath8 is the total number of vertices inside one compartment .
thick lines denote the compartment boundary consisting of the bonds not to be flipped.,width=396 ] we note that the tethered surface model is obtained from the compartmentalized model in the limit @xmath12 , where the vertices are prohibited from the free diffusion . on the contrary , we obtain the fluid surface model in the limit of @xmath13 , where all the vertices freely diffuse over the surface .
the surface model is defined by the partition function @xmath14,\\ s(x , { t})=s_1 + b s_2 , \nonumber\end{aligned}\ ] ] where @xmath15 is the bending rigidity , @xmath16 denotes that the center of the surface is fixed in the integration .
@xmath17 denotes that the hamiltonian @xmath18 depends on the position variables @xmath19 of the vertices and the triangulation @xmath20 .
@xmath21 denotes the sum over all possible triangulations @xmath20 , which keeps the compartments unflipped .
the gaussian term @xmath22 and the bending energy term @xmath23 are defined by @xmath24 where @xmath25 in @xmath22 is the sum over bonds @xmath26 connecting the vertices @xmath27 and @xmath28 , and @xmath25 in @xmath23 is also the sum over bonds @xmath26 , which are edges of the triangles @xmath27 and @xmath28 .
@xmath29 in eq .
( [ disc - eneg ] ) is the unit normal vector of the triangle @xmath27 .
the vertices @xmath19 are shifted so that @xmath30 , where @xmath31 is randomly chosen in a small sphere .
the new position @xmath32 is accepted with the probability @xmath33 $ ] , where @xmath34 .
the summation over @xmath20 in @xmath35 of eq.([part - func ] ) is performed by the standard technique for bond flip .
the bonds are labeled with sequential numbers .
the total number of bonds @xmath36 is given by @xmath37 , which includes the bonds making the boundary of compartments .
firstly , the odd - numbered bonds are sequentially chosen to be flipped , and secondly the remaining even - numbered bonds are chosen .
the flip is accepted with the probability @xmath38 $ ] . in this procedure
, the compartment boundary remains unflipped .
@xmath7-updates for @xmath19 and @xmath39-updates for @xmath20 are consecutively performed and make one mcs ( monte carlo sweep ) .
the radius of the small sphere for @xmath31 is chosen so that the rate of acceptance for @xmath19 is about @xmath40 .
we introduce the lower bound @xmath41 for the area of triangles .
no lower bound is imposed on the bond length .
we use surfaces of size @xmath42 , @xmath43 , and @xmath44 for @xmath45 , and @xmath46 , @xmath47 , and @xmath48 for @xmath49 , and @xmath50 , @xmath51 , and @xmath52 for @xmath53 .
vs. @xmath15 obtained on the surfaces of ( a ) @xmath45 , ( b ) @xmath49 , and ( c ) @xmath53 .
solid lines were obtained by multihistogram reweighting technique.,width=396 ] the mean square size @xmath54 is defined by @xmath55 where @xmath56 is the center of the surface .
figures [ fig-2](a),(b ) , and ( c ) show @xmath54 obtained on the surface of @xmath45 , @xmath49 , and @xmath53 , respectively .
solid lines connecting data were obtained by multihistogram reweighting technique @xcite . discontinuous change of @xmath54 can be seen in each case of @xmath8 when the size @xmath7 increases , and indicates a first - order transition .
we also find that the transition point @xmath57 moves left on the @xmath15 axis as @xmath8 decreases .
it is expected in the limit of @xmath58 that @xmath57 reduces to the value , which corresponds to the transition point of the tethered surface model @xcite . on the contrary
, @xmath57 moves right on the @xmath15-axis as @xmath8 increases .
we expect that @xmath57 disappears in the limit of @xmath13 at sufficiently large @xmath7 , which will be confirmed later in this letter . vs. @xmath15 obtained on the surface of ( a ) @xmath45 , ( b ) @xmath49 , and ( c ) @xmath53 .
@xmath59 is the total number of bonds .
, width=396 ] the bending energy @xmath60 defined in eq.([disc - eneg ] ) is shown in figs .
[ fig-3](a),(b ) , and ( c ) , which correspond to @xmath45 , @xmath49 , and @xmath53 , respectively . discontinuous change of @xmath60 can also be seen at @xmath15 where @xmath54 discontinuously changes , although the discontinuity of @xmath60 is not sufficiently clear in the figures .
we find also that the value of @xmath60 in the smooth phase ( or in the crumpled phase ) increases as @xmath8 decreases at the transition point .
the value of @xmath60 at the transition point becomes identical to the one of the tethered surface model in the limit of @xmath58 @xcite . on the contrary ,
the discontinuous change of @xmath60 is expected to disappear at sufficiently large @xmath8 , because the first - order transition at @xmath61 disappear when @xmath62 at sufficiently large @xmath7 . against @xmath8 , which are obtained on the @xmath63 , @xmath64 , @xmath11 surfaces .
the solid circles ( @xmath65 ) and the dashed lines denote the value of @xmath60 in the smooth ( the crumpled ) phase at the first - order transition point .
the symbol ( @xmath66 ) is unknown critical value of @xmath8 denoted by @xmath67 , where the discontinuous transition is expected to disappear .
the position of @xmath67 is also unknown on the @xmath15-axis .
, width=377 ] this implies that there exists a finite @xmath67 , where the first - order transition turns to a continuous or a higher - order one .
figure [ fig-4 ] shows the variation of the gap of @xmath60 against @xmath8 .
the gap of @xmath60 at @xmath45 reduces as @xmath8 increases and eventually goes to zero at @xmath68 , which will be clarified in the future .
we note that the maximum co - ordination number @xmath69 is as follows : @xmath70 on the @xmath71 surface at @xmath72 , @xmath73 on the @xmath74 surface at @xmath75 , and @xmath76 on the @xmath77 surface at @xmath78 .
these were obtained throughout the mc simulations . and the corresponding variation of @xmath60 against mcs obtained on the surfaces of ( a),(b ) @xmath79 , ( c),(d ) @xmath80 , and ( e),(f ) @xmath81 . , width=396 ] in order to show the discontinuity in @xmath60 more clearly
, we plot in figs . [ fig-5](a)[fig-5](f ) the distribution ( or histogram ) @xmath82 of @xmath60 and the corresponding variation of @xmath60 .
they were obtained on the surfaces of @xmath79 , @xmath83 , and @xmath84 .
the discontinuity of @xmath60 can be seen in the histogram on the @xmath85 surfaces .
because of the size effect , the transition appears to be continuous on the @xmath47 surface .
the double peak at @xmath79 is more clear than that at @xmath81 , because the gap of @xmath60 reduces as @xmath8 increases .
we should note that a double peak structure is very hard to see in @xmath82 on the @xmath86 surfaces .
when the configuration is once trapped in the smooth ( the crumpled ) state it hardly changes to the crumpled ( the smooth ) state at the transition point on such large surfaces .
this problem may be resolved with more sophisticated mc techniques than the current one in this letter . vs. @xmath15 , ( b ) @xmath60 vs. @xmath15 , and ( c ) @xmath87 vs. @xmath15 obtained on the fluid surface model without the compartmentalized structure .
the symbol @xmath88 drawn on the figures denotes that the surfaces are those without the compartment.,width=396 ] the phase transition is expected to disappear from the fluid surface model defined on the surfaces that have no compartment . in order to show this , we performed mc simulations on the surfaces without the compartments up to the size @xmath89 .
figure [ fig-6](a ) shows @xmath54 against @xmath15 .
we can see no abrupt growing of @xmath54 in the figure .
the bending energy @xmath60 shown in fig . [ fig-6](b ) , where the variation of @xmath23 against @xmath15 seems almost independent of @xmath7 .
the specific heat @xmath87 , which is defined by @xmath90 , is expected to reflect the phase transition .
however , we can see no anomalous behavior in @xmath87 shown in fig .
[ fig-6](c ) ; there can be seen no peak in @xmath87 .
thus we confirmed that the phase transition disappears from the model if @xmath91 at sufficiently large @xmath7 . 0.1 in ( 0,0 )
( 0,0 ) ( 6,-1.5)(0,0)(a ) @xmath92 , @xmath53 ( 28,-1.5)(0,0)(b ) the surface section obtained in the smooth phase at ( a ) @xmath93 , which is close to the transition point , and ( b ) the corresponding surface section .
, width=396 ] finally , we show in fig . [
fig-7](a ) a snapshots of the @xmath11 surface , which is obtained in the smooth phase close to the transition point .
figure [ fig-7](b ) shows the surface section .
thus we confirmed that the smooth phase can be seen close to the transition point at finite @xmath15 in contrast to the case @xmath13 , where it can be seen only at @xmath94 .
to the summary , we have studied a compartmentalized surface model for the cell membranes and found that the model undergoes a first - order transition between the smooth phase and the crumpled phase .
the model is classified as a fluid surface model , although the long - range order and the phase transition can be seen at finite @xmath15 .
the compartmentalized structure is an origin of a broken symmetry for the reparametrizaition invariance , which is considered to be a reason for the existence of the phase transition .
consequently , the result is not in contradiction with the standard argument for the non - existence of long - range order in fluid membranes .
moreover , the critical point of the transition is strongly expected at finite @xmath95 . the number @xmath67 and the corresponding critical bending rigidity @xmath96 remained to be clarified .
however , it is possible that the compartmentalized fluid surface model undergoes a continuous transition , because the strength of the transition is expected to be continuously softened with the increasing @xmath8 .
we observed that the phase transition remains first - order at @xmath53 and disappear at @xmath13 in the limit of @xmath97 .
although the demonstration of the continuous transition is computationally a hard task , it will be performed and the results including those in this letter will be reported elsewhere .
wolfhard janke , _ histograms and all that _ , in : computer simulations of surfaces and interfaces , nato science series , ii .
mathematics , physics and chemistry - vol .
114 , proceedings of the nato advanced study institute , albena , bulgaria , 9 - 20
september 2002 , edited by b. dunweg , d.p .
landau , and a.i .
milchev ( kluwer , dordrecht , 2003 ) , pp . | dynamically triangulated surface model is found to undergo a first - order crumpling transition between the smooth phase and a crumpled phase .
triangulated spheres are divided into compartmentalized domains , whose boundary bonds remain unflipped in the monte carlo simulations
. the vertices can freely diffuse only inside the compartments , and they are prohibited from the free diffusion over the surface due to the domain boundaries .
the violation of reparametrization invariance of the hamiltonian is a reason for the existence of the long - range order in the fluid surface model .
phase transition , extrinsic curvature , elastic membranes 64.60.-i , 68.60.-p , 87.16.dg |
one of the main goals of intermediate energy heavy ion collision ( hic ) is to study properties of nuclear matter , especially to determine the nuclear equation - of - state ( eos ) .
hic provides a unique means to compress nuclear matter to hot and dense phase within a laboratory environment . the pressures that result from the high densities achieved during such collisions strongly influence on the motion of ejected matter and provide the sensitivity to the eos .
in comparison with the conventional hadronic probes , photons interacting only weakly through the electromagnetic force with the nuclear medium are not subjected to distortions by the final state ( neither coulomb nor strong ) interactions and therefore photon delivers an undistorted picture of the emitting source @xcite .
since the last two decades , many model calculations and experimental facts @xcite have indicated that in intermediate energy heavy - ion collisions hard photons defined as photons with energies above the giant dipole resonance domain , above 30 mev in this paper , mainly originate from incoherent proton - neutron bremsstrahlung collisions , namely @xmath1 .
a nice review paper on hard photon was available @xcite .
these hard photons are emitted from two distinct sources .
the first and dominant component denoted as direct hard photon ( called as `` direct photon '' for short ) is associated with the first - chance proton - neutron collisions in the initial phase of the heavy - ion reaction .
the second one originates from the secondary proton - neutron collisions in the later stage of the reactions when the di - nuclear system tends to be thermalized , accordingly called as thermal hard photon ( called as `` thermal photon '' for short ) . because of their distinct emission sources , direct photons and thermal photons can deliver thermodynamic and dynamical information of hot and dense nuclear matter formed during the various stages of the heavy - ion collisions . in relativistic heavy ion collisions ,
photons are also of very interesting since they can be served as one of the potential signals of quark - gluon plasma ( qgp ) formation , eg . , see ref .
a hot qgp radiates a large amount of thermal photons , which dominate the spectra at small transverse momenta , whereas hard processes in nucleon - nucleon scatterings produce large momenta photons .
therefore photon enhancement at low transverse momenta could be seen as a qgp signal , which has been observed at the bnl - relativistic heavy - ion collider @xcite .
similar to the feature of photon production in intermediate energy hic , photons emitted from the interior of the hot matter no longer interact with the hadronic medium , in contrast to hadronic observables .
the paper is organized as follows . in sec .
ii , the simulation tool which we used is briefly introduced and the calculation method of photon production is outlined .
iii describes the classification of hard photons and definition of anisotropic flow , and presents the results of the azimuthal asymmetry for direct photons and free protons . in sec .
iv , we discuss the different variables ( impact parameter , beam energy and eos ) dependences of the hard photon production and/or anisotropic flow parameters .
v gives the results and discussions on two - photon correlation functions , namely momentum correlation function , which is also called hanbury - brown and twiss ( hbt ) correlation function , as well as azimuthal correlation function .
finally the summary is given in sec .
the transport model is very useful for treating heavy - ion collision dynamics and obtaining important information of nuclear matter . in intermediate energy heavy - ion collisions ,
the blotzmann - uehling - ulenbeck ( buu ) model is an extensively useful tool @xcite .
the buu equation takes both pauli blocking and the mean field into consideration , reads @xmath2 \nonumber \\ & \delta^3(p+p_2-p_3-p_4 ) .
\label{buu } \end{aligned}\ ] ] it is solved with the method of bertsch and das gupta @xcite . in eq .
( [ buu ] ) , @xmath3 and @xmath4 are in - medium nucleon - nucleon cross section and relative velocity for the colliding nucleons , respectively , and @xmath5 is the mean field potential including the isospin - dependent term : @xmath6 where @xmath7 is the normal nuclear matter density ; @xmath8 , @xmath9 , and @xmath10 are the nucleon , neutron and proton densities , respectively ; @xmath11 equals 1 or -1 for neutrons and protons , respectively ; the coefficients @xmath12 , @xmath13 and @xmath14 are parameters for nuclear equation of state .
three sets of mean field parameters are used , namely the soft eos with the compressibility @xmath15 of 200 mev ( @xmath12= -356 mev , @xmath13 = 303 mev , @xmath14 = 7/6 ) , and the semi - soft eos with @xmath15 of 235 mev ( @xmath12 = -218 mev , @xmath13 = 164 mev , @xmath14 = 4/3 ) , and the hard eos with @xmath15 of 380 mev ( @xmath12 = -124 mev , @xmath13 = 70.5 mev , @xmath14 = 2 ) .
@xmath16 is the symmetry energy strength due to the density difference of neutrons and protons in nuclear medium , which is important for asymmetry nuclear matter ( here @xmath17 mev is used ) , but it is trivial for the symmetric system studied in the present work .
the buu model was shown to be very successful in describing the bulk properties of the reaction and nucleon emission in intermediate - energy heavy - ion collisions .
in addition , the proton - neutron bremsstrahlung photon can be simulated as well in the model . for determining the elementary double - differential hard photon production cross sections on the basis of individual proton - neutron bremsstrahlung
, the hard - sphere collision was adopted from ref .
@xcite , and modified as in ref .
@xcite to allow for energy conservation .
the double differential probability is given by @xmath18 here @xmath19 is the radius of the sphere , @xmath20 is the fine structure constant , @xmath21 and @xmath22 are the initial and final velocity of the proton in the proton - neutron center of mass system , and @xmath23 is the angle between incident proton direction and photon emitting direction .
more details for the model can be found in ref .
in the present work , we simulate the reaction of @xmath24ca + @xmath24ca collisions in most cases .
sometimes @xmath25ca + @xmath25ca and kr + ni are also simulated .
[ time](a ) and ( b ) show the time evolutions of the production rate of bremsstrahlung hard photons and of system densities , including both maximum density ( closed circles ) and average density ( open circles ) , respectively , for @xmath24ca + @xmath24ca collisions at 60@xmath26 mev in the centrality of 40%60% .
it is found that the hard - photon production rate is sensitive to the density oscillations during the whole reaction evolution . with the increase in density
when the collision system is in the compression stage , the system produces more hard photons .
in contrast , when the system starts to expand , the hard photon production rate decreases .
actually , the density oscillation of the colliding heavy - ion systems can be observed in the experiments via hard - photon interferometry measurements @xcite as well as dynamical dipole @xmath27 radiation @xcite .
apparently , hard photons are mostly produced at the early stage of the reaction .
thereafter we call these photons , emitted before the time of the first maximum expansion of the system ( @xmath28 fm / c in ca + ca at 60@xmath26 mev ) , as direct hard photons ( on the left side of the blue dashed line in fig .
[ time](a ) ) .
it is also coincident with the definition in the sec.i of direct hard photons .
we call the residual hard photons producing in the later stage as thermal hard photons ( on the right side of blue dashed line in fig . [ time ] ) . in this way
, we can identify the producing photon as either direct or thermal by the emission time of photons in the present simulation .
we notice that the production rate of thermal photon tends to decrease for later time , after 300 fm / c for ca + ca at 60@xmath26 mev , which can be understood by the decrease of n - p collisions in an expanding system . however , the hard photon yields in the later stage after 300 fm / c is relative small fraction of the total thermal hard photon emission .
also , in the following calculations , we mainly focus on direct hard photons which dominte the yield and do not depend on the time evolution for the later stage .
ca + @xmath24ca at 60@xmath26 mev in 40%60% centrality
. the eos parameters with the compressibility @xmath15 of 235 mev are used ; ( b ) time evolution of reduced maximum density @xmath29 ( closed circles ) and reduced average density @xmath30 ( open circles ) of the whole reaction system in the same reaction .
the blue dashed line represents the time when till the first expansion stage , and in the panel ( a ) it separates direct hard photons ( on the left side ) and thermal hard photons ( on the right side ) . ] for an example , fig .
[ photon - spectra ] demonstrates that the direct and thermal hard photons for 60@xmath26 mev @xmath24ca + @xmath24ca collisions with a compressibility @xmath15 of 235 exhibit different spectuml shapes : thermal photons give rise to a softer energy spectral than direct ones . in the nucleon - nucleon cente - of - mass system
, the photon spectrum can be described by the function : @xmath31 where @xmath32 is a normalization constant and @xmath33 is the slope parameter , which reflects the apparent source temperature of the photon emission source .
the slope parameter @xmath33 depends on the bombarding energy and on the average intrinsic momentum of the participant nucleons .
as less energy is available on average in second - chance n - p collisions than in first - chance collisions because most of the projectile kinetic energy is damped , the thermal - photon spectrum becomes much softer .
if one adds the two sources together in the photon spectrum , one obtains an empirical photon spectrum of the form @xcite : @xmath34 with the constant @xmath35 defined by : @xmath36 @xmath37 represents the intensity of each photon source , @xmath38 stands for the direct photon and @xmath39 for stands for the thermal photon .
however , we notice that the change in slope of the photon yield could be also affected by the 1/@xmath40 factor which enters the elementary @xmath41 bremsstrahlung probability @xcite . in this case , eq . [ fit ] should be changed into @xmath42 in the next paragraph , we will compare the difference for the fitted slope with eq .
[ fit ] and eq .
[ fit2 ] .
as shown in the inset of the figure , the slope parametesr of direct photons and thermal photons are 16.7 and 8.1 mev , respectively , for inclusive events of @xmath24ca + @xmath24ca collisions at 60@xmath26 mev with an incompressibility @xmath15 of 235 mev if we fit the spectrum with the eq .
however , eq . [ fit2 ] will give a larger slope value , namely 22.5 and 9.8 mev , respectively , which is 35% and 20% higher than the fits without the 1/@xmath40 factor ( not plotted in the figure ) .
with two different fit formulas , we put the impact parameter dependence of the slope parameter together in fig .
[ slope - b ] .
a remarkable difference is seen between the slope values with and without the 1/@xmath40 factor , i.e. eq . [ fit2 ] gives a higher apparent photon temperature than eq .
in some previous experimental analysis @xcite , eq .
[ fit ] was mostly used to extract the apparent slope , but this is probably not correct since a factor 1/@xmath40 in front of the exponential form can be derived from their elementary cross section [ see eq .
[ ddcs ] ] . on the other hand
, the slope of hard photons displays a slight decreasing behavior with increasing impact parameter , indicating that the emission source becomes cooler in peripheral collisions . moreover , direct photons are hotter than thermal photons as we expected .
ca + @xmath24ca at 60@xmath26 mev with the compressibility @xmath15 of 235 mev .
black dots and blue dots represent the direct photons and thermal photons , respectively .
lines represent the exponential fits with th eq .
[ fit ] . ]
ca+@xmath24ca at 60@xmath26 mev with the incompressibility @xmath15 of 235 mev .
open symbols represent the results from the eq .
[ fit2 ] and solid symbols are from eq .
[ fit ] . ]
it is well known that collective flow is an important observable in heavy ion collisions and it can carry some essential information on the nuclear matter , such as the nuclear equation of state @xcite .
anisotropic flows are defined as different @xmath43th harmonic coefficients @xmath44 of an azimuthal fourier expansion of the particle invariant distribution @xcite @xmath45 where @xmath46 is the azimuthal angle between the transverse momentum of the particle and the reaction plane .
note that in the coordinate system the @xmath47-axis is along the beam axis , and the impact parameter axis is labeled as the @xmath48-axis . the first harmonic coefficient @xmath49 represents the directed flow , @xmath50 where @xmath51 is the transverse momentum .
the second harmonic coefficient @xmath52 represents the elliptic flow which characterizes the eccentricity of the particle distribution in momentum space , @xmath53 in relativistic heavy - ion collisions directed and elliptic flows of hard photons have been recently reported in the experiments and through theoretical calculations @xcite , demonstrating a very useful tool for exploring the properties of hot dense matter
. however , there are no experimental data available so far on the flow of hard photons in intermediate energy heavy ion collisions . ) of hard photons ( a ) and directed flow ( @xmath49 ) ( b ) for 60@xmath26 mev ca + ca collisions at centrality of 40 - 60@xmath54 . ] in this context , here we calculate directed transverse or elliptic flows in intermediate energy heavy ion collisions .
in addition , given that hard photons mostly originate from bremsstrahlung by individual proton - neutron collision and that free nucleons are also emitted from nucleon - nucleon collisions , therefore it will be interesting to compare flows of protons and photons . fig .
[ flow_time ] shows the time evolution of the average values of directed flow and elliptic flow of the photons .
considering the nearly symmetric behavior for directed flow versus rapidity , here we calculate the average @xmath49 over only the positive rapidity range , which can be taken as a measure of the directed flow . from fig .
4 , the directed flow of photons show rapid rising with positive values during the compression stage and later on it decreases to even negative value . afterwards , the directed flow remains negative since the system is never compressed .
for elliptic flow , the behavior shows contrary trend to the @xmath49 and later on it shows oscillation for thermal photon emission .
the times corresponding to the peak or valley values of flows roughly keep synchronized with the compression or expansion oscillation on the system evolution as shown in fig .
1(b ) .
rapidity for 60@xmath26 mev ca + ca collisions in the semi - central centrality of @xmath55 .
the dashed line segment is a fit over the mid - rapidity region @xmath56 .
( b ) same as the left panel but for free protons .
the dashed line segment is a fit over the mid - rapidity region @xmath57 . ] mev ca + ca collisions in the semi - central centrality of @xmath55 .
both of them are fitted to the 4-th order fourier expansion . ] from the above calculations , we learn that direct photons are preferentially emitted in the out - of - plane ( negative @xmath52 ) direction and thermal photons are emitted from a thermalizing system which makes their emission more isotropic ( i.e. the oscillated elliptic flow ) than the direct ones produced in the pre - equilibrium stage .
in addition , thermal photons contribute less than @xmath58 of the total yield of hard photon in the present model . therefor we will focus on direct hard photons to discuss the flow results .
the quantity of directed transverse flow at mid - rapidity can be also defined by the slope : @xmath59 , where @xmath60 is the rapidity of particles in the center of mass and @xmath61 is the mean in - plane transverse momentum of photons or protons in a given rapidity region . in fig .
[ flow](a ) and ( b ) , we show @xmath61 plotted versus the _ c.m . _
rapidity @xmath62 for direct hard photons ( a ) as well as @xmath61 plotted versus the reduced _ c.m .
_ rapidity @xmath63 for free protons ( b ) for 60@xmath26 mev ca + ca collisions in the semi - central centrality of @xmath55 .
the eos parameters with the compressibility @xmath15 of 235 mev are used .
the errors shown are only statistical . for emitted proton ( free proton ) ,
we identify it in the buu calculation as its local density @xmath64 .
we take the values of flows when the system has been already in the freeze - out time at 180 fm / c .
a good linearity is seen in the mid - rapidity region @xmath65 and the slope of a linear fit can be defined as the magnitude of directed transverse flow .
the extracted value for the directed transverse flow of direct hard photons is about @xmath66 mev / c , and that of free protons is about @xmath67 mev / c .
thus direct hard photons do exist the directed transverse flow even though the absolute value is smaller than the proton s flow , and the sign of its flow is just opposite to that of free protons . ) for direct hard photons ( a ) and free protons ( b ) as a function of transverse momentum ( @xmath68 ) for 60@xmath26 mev ca + ca collisions in the semi - central centrality of @xmath55 . ] in order to extract the value of elliptic flow and reduce the error of fits , we fit the azimuthal distribution to the @xmath69 order fourier expansion .
fig.[azimuthal ] shows an example for such fits with the elliptic flow parameter @xmath52= -2@xmath54 for photons and @xmath52= 5.5 @xmath54 for protons .
[ v2pt ] shows the differential elliptic flow of direct hard photons ( a ) and of free protons ( b ) as a function of transverse momentum @xmath68 for @xmath24ca + @xmath24ca at 60@xmath26 mev .
similar to the directed transverse flow , the values of elliptic flow of direct hard photons and of free protons also have an opposite signs at this reaction energy , i.e. , reflecting a different preferential transverse emission in the direction of out - of - plane or in - plane , respectively
. meanwhile , absolute flow values for photons are smaller than the proton s like the behavior of transverse flow .
except of the opposite sign , we see that both elliptic flows have similar tendency with the increasing of @xmath68 , i.e. , their absolute values increase at lower @xmath68 , and becomes gradually saturated , especially for direct hard photons . to explain the correlation of the collective flow between direct hard photons and free protons indicating above , we should note that flows of direct hard photons originate from the individual proton - neutron collisions . as eq .
[ ddcs ] shows , we can roughly consider that in the individual proton - neutron center of mass system , in directions perpendicular to incident proton velocity , i.e. @xmath70 , the possibility of hard photon production is much larger than the parallel direction , actually it is also in agreement with the theoretical calculations and the experiments @xcite . as a whole , the flow of hard photons should be correlated with the collective movement of the nucleons , and presents an opposite behavior .
consequently , flows of hard photons and of free nucleons exhibit an opposite character .
it is well known that anisotropic flow mainly originates from the initial asymmetric overlap zone of colliding nuclei which induces different pressures or rotational collective motion of the participant region and leads to anisotropic emission of the particles .
peripheral collisions have more initial asymmetry in overlap zone than central collisions , thus more anisotropic emission of the particles is expected .
therefore anisotropic emission shall be sensitive to impact parameter .
here we simulate the reaction @xmath24ca + @xmath24ca collisions at 60@xmath26 mev , and use the eos parameters with the compressibility @xmath15 of @xmath71 mev for the nuclear mean field @xmath5 .
[ fig8 ] ( a ) and ( b ) shows the directed transverse flow parameter @xmath72 of direct photons and free protons , respectively , as a function of the reduced impact parameter ( i.e. normalized by the maximum impact parameter ) . fig .
[ fig8 ] ( c ) and ( d ) are the same as ( a ) and ( b ) but for elliptic asymmetry coefficient @xmath52 .
we can see that both @xmath72 and @xmath52 of direct photons and free protons have the similar tendency with impact parameter , i.e. their absolute values increase with the impact parameter except of slightly decreasing in very peripheral collisions for @xmath52 .
we also see that at all impact parameters , in contrast to to the negative directed transverse flow and positive elliptic flow of free protons , direct photons show the positive @xmath72 and the negative @xmath52 , i.e. the anisotropy is shifted by a phase @xmath73/2 .
it agrees with the previous conclusion that the azimuthal asymmetry of direct photons is anti - correlated with the corresponding free proton s flow .
( a ) and ( b ) are the directed transverse flow parameter @xmath72 of direct photons and free protons , respectively , as a function of the reduced impact parameter for the reaction @xmath24ca + @xmath24ca collisions at 60@xmath26 mev .
( c ) and ( d ) same as above but for elliptic asymmetry coefficient @xmath52 . ] in fig .
[ v2-e ] we show the directed transverse flow parameter @xmath72 and elliptic asymmetry coefficient @xmath52 of direct photons and free protons , respectively , as a function of beam energy for the ca + ca reaction at @xmath74 centrality . in the beam energy range
studied here , same as the impact parameter dependence , the opposite signs of @xmath72 and @xmath52 between direct photons and free protons are also anticorrelated .
moreover , except for the opposite sign , directed transverse flow parameter @xmath72 of direct photons and free protons have similar structures with the increasing of beam energy .
the value of direct photon elliptic asymmetry coefficient @xmath52 increases with beam energy from negative to positive , a tendency similar to available experimental results on hard photons in relativistic heavy - ion collisions . of direct photons and free protons , respectively , as a function of beam energy for the reaction which is the same as fig .
[ fig8 ] but only semi - central events @xmath74 .
( c ) and ( d ) same as above but for elliptic asymmetry coefficient @xmath52 . ] in this section we will discuss the nuclear equation of state and its symmetry energy term depdences of hard photon production . in our calculations , the single - particle potential taken as the skyrme parameterized mean field potential including symmetry potential is shown below @xmath75 where the coefficients @xmath12 , @xmath13 and @xmath14 are parameters for nuclear equation of state which is determined by the nuclear saturation property and the incompressibility @xmath15 of symmetric nuclear matter ; @xmath76 is the isospin - asymmetry parameter , @xmath77 , and @xmath78 is the symmetry potential of neutrons ( protons ) . first , we only investigate the effects of the incompressibility @xmath15 on the hard photon production in symmetric reaction system . in the skyrme potential ,
we take the incompressibility @xmath15 as 200 mev , 235 mev and 380 mev , in which the first two correspond to soft and semi - soft potential and the last one corresponds to hard potential as we introduced in sec.ii a. because of the sensitivity to the density oscillations of colliding system , hard photon should be also dependent of the eos of nuclear matter , especially for thermal hard photons @xcite . actually as shown in fig .
[ product - rate ] , photon production shows its sensitivity to the compressibility , especially for thermal photon which are produced after @xmath79 fm / c .
it shows that the stiffer the eos , the higher multiplicity the thermal photons .
in contrast , direct photons are produced by the first channel neutron - proton bremsstrahlung , their production rate only weakly depends on eos since direct hard photons are emitted by the first channel n - p scattering when the system is in a highly nonequilibrium state during the compression stage , and they don not have enough times to feel the eos .
ca + @xmath24ca at 30@xmath26 mev .
different eos are used , namely the hard eos , the semi - soft eos and the soft eos , respectively . ] . shown in fig .
[ fig : m_t_eos ] is the time evolution of inclusive hard photon multiplicity from the symmetric reaction system @xmath24ca + @xmath24ca at 60@xmath26 mev .
we can see that at this energy , for different incompressibility @xmath15 , multiplicities of hard photons produced in the early stage of collisions are nearly equivalent and later they show a clear correlation with incompressibility @xmath15 . corresponding to direct and thermal photons , this indicates that direct photon production is not sensitive to incompressibility @xmath15 , because at this reaction energy , comparable with two - body interactions , mean field can be neglected in producing hard photons in the early stage of collisions .
however , there is a correlation between @xmath15 and thermal photon production , with larger @xmath15 producing more thermal photons . for the reaction @xmath24ca + @xmath24ca at 60@xmath26 mev . ] for the magnitude of flow parameters @xmath72 and @xmath52 , comparisons are made for different equation of state .
[ flow - eos ] shows the @xmath72 and @xmath52 as a function of impact parameter with different eos .
generally , the directed flow becomes larger with the increasing of stiffness of the eos , whereas the elliptic flow becomes smaller . ) and
elliptic flow ( @xmath52 ) as a function of impact parameter with different eos parameters for ca + ca at 30@xmath26 mev .
] in order to study the effects of symmetry potential , we now set incompressibility @xmath15 as 380 mev .
we mainly consider three kinds of symmetry potential : ( 1 ) the symmetry potential is neglected , that is @xmath80 ( 2 ) the symmetry potential is a linear function of isospin - asymmetry parameter @xmath76 , that is @xmath81 where @xmath16=@xmath82 mev , @xmath83=1 , @xmath84=-1 ; + ( 3 ) the li et al .
single - particle potential from the parametrization of nuclear symmetry energy used in ref .
@xcite for studying the properties of neutron stars , that is @xmath85 , li _ et al .
_ derived a single - particle symmetry potential as @xcite : @xmath86\times \delta^2\\ & + \left(e_0(\rho_0)(\frac{\rho}{\rho_0})^{\gamma}-12.7(\frac{\rho}{\rho_0})^{2/3}\right)\delta\tau^{n(p ) } , % \nonumber \label{vasylba } \end{aligned}\ ] ] where @xmath83=1 , @xmath84=-1 , @xmath87 sets as 30 mev in this paper , and parameter @xmath27 represents the stiffness of the symmetry energy . in the following , we consider two cases of @xmath27 = 0.5 and 2 , respectively , corresponding to soft and stiff symmetry energy , to explore the effects of different symmetry potential .
the symmetry potentials for neutron and proton with @xmath76 = 0.20 is plotted in fig .
[ potential ] , where the red line represents @xmath88 , and the blue and green lines represent @xmath89 for @xmath27 parameters of 0.5 and 2.0 for neutrons ( upper branch ) and protons ( lower branch ) , respectively . for small isospin asymmetry and density near @xmath7
the above symmetry potentials reduce to the well known lane potential which varies linearly with @xmath76 @xcite .
generally , the repulsive / attractive symmetry potential for neutrons/ protons increases with density .
= 0.2 and the @xmath27 parameters of 0.5 ( red line ) and 2.0 ( green line ) , respectively .
solid lines are for neutrons and dashed lines for protons . ]
( color online ) time evolution of hard photon multiplicity with different symmetry potentials for the reaction @xmath25ca + @xmath25ca at 60@xmath26 mev . ] fig .
[ fig : vasyca48lba ] shows the time evolution of inclusive hard photon multiplicity in the reaction @xmath25ca + @xmath25ca at 60@xmath26 mev . in order to separate the effects of symmetry potential from the incompressibility
, we set @xmath15 as a constant 380 mev .
[ fig : vasyca48lba ] presents the effects of different symmetry potentials on hard photon productions .
same as the incompressibility @xmath15 , direct photon production is insensitive to symmetry potential at the studied reaction energy . for thermal photon
, the curves of @xmath90 and @xmath88 have the nearly same multiplicities with time evolution , and they are intermediate between @xmath91 and @xmath92 .
we note that thermal photon production is sensitive to @xmath27-parameter in symmetry potential @xmath93 , i.e. less @xmath27 , corresponds to softer symmetry energy , induces more thermal photons .
and we also note that the sensitivity of thermal photon production to @xmath27 is not remarkable in comparison with its sensitivity to the incompressibility @xmath15 , which is of course understandable since the symmetry energy is relative small term in comparison with the incompressibility @xmath15 . from the effects of the incompressibility and symmetry potential on hard photon production
, it appears that the yield of thermal photons has rather strong dependence on nuclear compressibility as well as symmetry energy , whereas the yield of direct photons is rather insensitive to nuclear eos .
therefore thermal hard photon could serve as a porble of nuclear eos in intermediate energy hic .
in above sections , we mainly focus on the properties of inclusive hard photons , from which we find that production distribution and anistropic flows of hard photons reveal rich information of nuclear dynamics .
also , the sensitivity to the nuclear equation of state and symmetry energy for hard photons has been discussed . in this section
, we will discuss the correlation properties of photons in terms of two - particle correlation technique that can provide us a very powerful tool to characterise the properties of a particle source .
in particular , two - boson relative momentum distributions enable one to determine the space - time structure according to the formalism of bose - einstein correlations . the magnitude of the correlation can be related to ther space - time distribution of the boson source . in the follwoing calculations
, we will construct two - photon momentum correlation function as well as azimuthal correlation function , from which the space - time structure of the photon souce and anistropy property can be indicated .
intensity interferometry ( also called hbt correlation ) is used as a universal tool to study the properties of any boson sources such as stars @xcite , or photon and meson sources in the early phase of heavy - ion collisions @xcite .
the formalism was developed starting from optics and quantum statistics and was finally adapted to the dynamics of heavy - ion collisions @xcite .
we have performed the calculation which evaluates the correlation function directly from the photon source distribution given by the buu calculation .
we store for each @xmath94th @xmath95 collision its position @xmath96 and the associated photon probability distribution @xmath97 . after the completion of the calculation
, we analyze this output data to construct plane waves with four - momentum @xmath98 at @xmath99 and calculate the two - photon probability for @xmath100 as @xcite : @xmath101\right ] , \label{two_photon_pro}\end{aligned}\ ] ] where @xmath102 represents the probability to produce a pair without correlation , and @xmath26 and @xmath103 are the amplitudes related to the normalized probabilities of the direct @xmath104 $ ] and cross terms @xmath105 $ ] .
this corresponds to fourier transforming the photon source event by event .
we set the weight of the interference in eq .
[ two_photon_pro ] as @xmath106 , that is @xmath107 , which was predicted between 0.5 and 1.0 depending on the anisotropy of hard photon emission .
the exact experimental filter was finally applied to the projection onto the lorentz - invariant relative four momentum @xmath108 of the resulting distribution @xmath109 and @xmath102 of eq .
[ two_photon_pro ] , then the two - photon correlation function was calculated as @xmath110 = f(q)$ ] . for convenience ,
@xmath106 was set to 0.75 for our calculations in order to take into account the established anisotropic component in the angular distribution of hard photons .
the two - photon correlation function has been fitted with a gaussian parameterization with correlation strength @xmath111 and radius parameter @xmath112 : @xmath113 in the above relation @xmath112 is the space - time parameter conjugate to @xmath114 , which measures an invariant length depending on the source - size parameters @xmath19 and @xmath115 : @xmath116 one can easily know that , since @xmath117 , @xmath112 is a first order measure of the spatial extent of the source , that is @xmath118 .
we can then calculate from @xmath119 the rms radius of the source as the one of a static three - dimensional gaussian source : @xmath120 .
kr + @xmath121ni collisions at 60@xmath26 mev in the laboratory frame , @xmath122 mev .
the green squares represent for direct photons , blue triangles for thermal photons , and red open circles for inclusive photons .
the black closed circles represent for the experimental data taken from ref . @xcite . ]
ca + @xmath24ca collisions at 60@xmath26 mev .
the symbols are the same as fig .
[ c_hbt ] and the dashed lines are fitting functions as eq . [ hc_fit ] . ] for the comparisons with experimental data , we calculated the reaction @xmath123kr + @xmath121ni at 60@xmath26 mev , employing here the following filter which is very similar to the experiment : @xmath124 mev , detector positions between polar angles of @xmath125 and @xmath126 ( orienting downstream ) , azimuthal opening angles of @xmath127 and @xmath128 , and @xmath129 for the minimum opening angle .
[ c_hbt ] presents hbt correlation functions of direct photons ( green squres ) , thermal photons ( blue triangles ) and inclusive photons ( red open circles ) , respectively .
we find that the correlation function of direct photons is much larger than thermal one , and the correlation function of inclusive photons is intermediate .
this result is reasonable , because direct photons are emitted at the early compressed stage of collisions , so they have stronger interference than thermal photons which are produced in later thermalizing stage .
we also see that our result can well reproduce the experimental data @xcite , especially for the correlation function of thermal photons , which agrees with the experimental function even the oscillation structure . therefore , we can successfully reproduce the hbt correlation function of hard photons by the buu simulation . in order to further study the hbt correlation of hard photon and
extract photon source information , we calculated the symmetric reaction system @xmath24ca + @xmath24ca at 60@xmath26 mev in the c.m .
frame , and only the events of central collision ( @xmath130 ) were taken for simplification . in fig .
[ c_hbt2 ] , we find that the correlation function of direct photons is the largest , the thermal one is the least , and the correlation function of inclusive photons are between them . moreover , in the conditions of above reaction , all of them can be well fitted by eq .
[ hc_fit ] . after fitting correlation function ,
we can obtain two useful fitting parameters : correlation strength @xmath111 and radius parameter @xmath112 . as the results on top right corner show , direct photons have the largest correlation strength @xmath111 , the second is inclusive photons and
the least one is thermal photons .
thus @xmath111 is a parameter which is sensitive to the intensity of interference .
and we also get three radius parameters @xmath112 .
@xmath112 of thermal photons is nearly twice than direct photons , and inclusive photon is equivalent to the later . to explain this , we know that direct photon is mostly emitted in the early stage when the reaction system is strongly compressed , so the emission source of photons is small . and
later on , the thermalizing system extends very much in company with the production of thermal photons . in this reaction energy at @xmath131 = 60@xmath26 mev , most of hard photons are produced in the early stage of the collision , so the spatial source extent of inclusive photons should be approximate to the direct photons . as a result , they have equivalent radius parameter @xmath112 .
then we can get the rms radius of the photon source by the equation @xmath120 .
therefore , two - photon correlation function provides the information of hard photon source , which is available to investigate the emission source during the collisions . from individual - photon azimuthal distribution
, we find that direct hard photons exhibit the azimuthal asymmetric emission in intermediate energy heavy - ion collisions , especially negative elliptic flow parameter @xmath0 .
actually it is difficult to extract the elliptic flow parameter @xmath0 by the method of reconstructing reaction plane in the experiment , so as below , we will discuss to use the method of two - photon azimuthal correlation to extract the elliptic flow of direct photons directly .
for particles in the same class , we defined the particle azimuthal correlation function following the multi - fragment azimuthal correlation method @xcite .
the multi - fragment azimuthal correlation function is defined as follows : @xmath132 where @xmath133 is the @xmath134 distribution for fragment pairs from the same event and @xmath135 is the @xmath134 distribution by randomly selecting each member of a fragment pair to form mixed events .
the @xmath134 between all selected fragments in an event are used to construct the correlation function , that is , @xmath136 angles for @xmath43 fragments . in our calculations for two - photon azimuthal correlations , we get @xmath133 from the @xmath134 distribution of photon pairs in the same event and @xmath135 from the mixed events . for these correlations , we may make a fit of the fourier series with the expression @xmath137 , \label{cazifit}\ ] ] where @xmath138 and @xmath139 are treated as fit parameters .
ca + @xmath24ca at 60@xmath26 mev for semi - central events ( 40%60% ) , and the dashed line is the fitting function as eq . [ cazifit ] ; bottom panel : individual - photon azimuthal distribution for the same reaction , and the dashed line is a fit of fourier expansion.,title="fig : " ] ca + @xmath24ca at 60@xmath26 mev for semi - central events ( 40%60% ) , and the dashed line is the fitting function as eq . [ cazifit ] ; bottom panel : individual - photon azimuthal distribution for the same reaction , and the dashed line is a fit of fourier expansion.,title="fig : " ] under the assumption of statistically independent emission of particles with the same azimuthal distribution @xmath140 in an event , the azimuthal correlation function is simply related to @xmath140 via the convolution @xmath141 one the other hand , @xmath140 can be described by the fourier expansion : @xmath142 where the different @xmath43th harmonic fourier expansion coefficient @xmath143 is defined as the @xmath43th anisotropic flow , of which @xmath144 is directed flow parameter and @xmath0 is so called elliptic flow parameter . if we only take @xmath49 and @xmath52 terms in eq .
[ azi_dist ] and substitute eq .
[ azi_dist ] into eq .
[ fourier ] , we can derive the form of @xmath145 as follows : @xmath146 \label{c_derive}\ ] ] from eq.[fourier ] and eq.[azi_dist ] , we then get the relationship between the fitted parameter @xmath139 and the elliptic flow parameter @xmath0 @xcite : @xmath147 for reaction @xmath24ca + @xmath24ca at 60@xmath26 mev with centrality 40%60% , open circles represent the amplitude of @xmath52 extracted from two - photon azimuthal correlation and red ones from the fit of fourier expansion to individual - photon azimuthal distribution . ]
now we make the comparison of elliptic flow parameter @xmath0 extracted from two - photon azimuthal correlation with fitting value from individual - photon azimuthal distribution . in fig .
[ c_azi ] , we calculated the reaction @xmath24ca + @xmath24ca at 60@xmath26 mev for semi - central events ( 40%60% ) with the compressibility @xmath15 = 235 mev , and we extracted the elliptic flow parameters @xmath0 by the two distinct methods above .
then we can see that the absolute values of @xmath0 are nearly equivalent , except that we can only get the amplitude of elliptic flow from two - photon azimuthal correlation , and the fitting value of @xmath52 from individual - photon azimuthal distribution is negative .
further , as shown in fig .
[ hcazi_v2e ] , we get the excitation function of @xmath52 by the above two methods which shows nearly same amplitude of @xmath52 in various reaction energy but different signs . therefore , in view of the difficulty to reconstruct reaction plane in experiment , two - photon azimuthal correlation provides us an alternative method to extract the amplitude of elliptic flow parameter @xmath0 of hard photons .
in summary , we have systematically investigated hard photon production by the process of proton - neutron bremsstrahlung as well as the behavior of azimuthal asymmetry of hard photons and free protons produced in intermediate energy heavy - ion collisions in the framework of the buu model .
as discussed in previous studies , hard photons can be separated into direct hard photon which is produced by the first channel neutron - proton bremsstrahlung and thermal hard photon which is produced by the later stage neutron - proton bremsstrahlung during the system s evolution towards thermalization .
the kinetic energy spectra of two kinds of photons show the exponential form which provides us with information on the apparent temperature in different stage of heavy - ion collision .
the azimuthal asymmetry parameters have been investigated in detail .
the time evolution of directed flow and elliptic flow of hard photons exhibits rich structure as the density of the system oscillates during the pre - equilibrium and thermalization stage of reaction system .
this structure indicates that the azimuthal asymmetry evolves with time .
a nonzero directed transverse flow and elliptic flow parameter have been predicted for the direct hard photons produced by the first channel proton - neutron bremsstrahlung process in the intermediate energy heavy - ion collisions .
the asymmetry parameters of hard photons are plotted as a function of rapidity and transverse momentum , which show contrary signs in comparison with the flow of free protons , i.e. the azimuthal asymmetry of direct hard photons seems to be anticorrelated to that of the corresponding free proton s .
therefore we can expect the direct hard photon can server as a good probe to nuclear matter properties in the early stage of hic .
different variables dependences are investigated for the hard photon production and anisotropic flow .
we can see that the absolute value of both directed flow and elliptic flow of direct photons and free protons increase with the impact parameter , indicating that the flow mainly reflects the initial geometric asymmetry of the collision zone in a given beam energy and revealing a rich structure of flows with increasing beam energy . for directed flow , direct photons reach a minimum of around 40@xmath26 mev where the free protons approach the maximum negative value , while for elliptic flow , the absolute values of direct photons show decreasing trends with increasing beam energy .
the eos dependence for hard photons at different time indicates that the direct photon is not sensitive to the nuclear incompressibility nor to the symmetry energy .
however , for the thermal photon , its multiplicity increases with nuclear incompressibility . for a given nuclear incompressibility ,
the soft symmetry energy favors thermal photon production .
finally , we calculated the two - photon correlations , including hbt momentum correlations and azimuthal correlations . from two - photon hbt correlations , we can extract photon source information , such as intensity interference and the spatial extent of the emission source .
we also find that two - photon azimuthal correlations can provide us with a good method for extracting the amplitude of the elliptic flow parameter @xmath0 of hard photons in the experiment . in light of the present study
, we expect that direct photons would be a very useful probe for exploring nuclear reaction dynamics in intermediate energy heavy - ion collisions , while the thermal hard photon can give us some hints on the nuclear eos , including the symmetry energy .
this work is supported partially by the knowledge innovation project of chinese academy of sciences under grant no .
kjcx2-ew - n01 , the national natural science foundation of china under contract no.s 11005140 , 11035009 , 10979074 , 10875160 , 10805067 and 10975174 . 0 y. schutz , g. martnez , f.m .
marqus , a. marn , t. matulewicz , r.w .
ostendorf , p. boek , h. delagrange , j. daz , m. franke , k.k .
gudima , s. hlav , r. holzmann , p. lautridou , f. lefvre , h. lhner , w. mittig m. ploszajczak , j.h.g .
van pol , j. qubert , p. roussel - chomaz , a. schubert , r.h .
siemssen , r.s .
simon , z. sujkowski , v.d .
toneev , v. wagner , h.w .
wilschut , gy .
wolf , nucl .
a * 622 * , 404 ( 1997 ) .
r. wada , d. fabris , k. hagel , g. nebbia , y. lou , m. gonin , j. b. natowitz , r. billerey , b. cheynis , a. demeyer , d. drain , d. guinet , c. pastor , l. vagneron , k. zaid , j. alarja , a. giorni , d. heuer , c. morand , b. viano , c. mazur , c. ng , s. leray , r. lucas , m. ribrag , and e. tomasi phys .
c * 39 * , 497 ( 1989 ) .
g. martinez , f.m .
marqus , y. schutz , gy . wolf , j. daz , m. franke , s. hlav , r. holzmann , p. lautridou , f. lef + evre , h. lhner , a. marn , t. matulewicz , w. mittig , r.w .
ostendorf , j.h.g .
van pol , j. qubertf , p. roussel - chomaz , a. schubert , r.h .
siemssen , r.s .
simon , z. sujkowski , v. wagner , h.w .
wilschut , phys .
b * 349 * , 23 ( 1995 ) .
d. g. denterria , l. aphecetche , a. chbihi , h. delagrange , j. daz , m. j. van goethem , m. hoefman , a. kugler , h. lhner , g. martnez , m. j. mora , r. ortega , r. w. ostendorf , s. schadmand , y. schutz , r. h. siemssen , d. stracener , p. tlusty , r. turrisi , m. volkerts , v. wagner , h. w. wilschut , and n. yahlali , phys .
. lett . * 87 * , 022701 ( 2001 ) .
f. m. marqus , g. martnez , y. schutz , j. daz , m. franke , s. hlav , r. holzmann , p. lautridou , f. lefvre , h. lhner , a. marn , t. matulewicz , w. mittig , r.w .
ostendorf , j.h.g .
van pol , j. qubert , p. roussel - chomaz , a. schubert , r.h .
siemssen , r.s .
simon , z. sujkowski , v. wagner , h.w .
wilschut , gy .
wolf , phys .
b * 349 * , 30 ( 1995 ) .
a. corsi , o. wieland , v.l .
kravchuk , a. bracco , f. camera , , g. benzoni , n. blasi , s. brambilla , f.c.l .
crespi , a. giussani , s. leoni , b. million , d. montanari , a. moroni , f. gramegna , a. lanchais , p. mastinu , m. brekiesz , m. kmiecik , a. maj , m. bruno , m. dagostino , e. geraci , j , g. vannini , s. barlini , g. casini , m. chiari , a. nannini , a. ordine , m. di toro , c. rizzo , m. colonna , v. barank , phys .
b * 679 * , 197 ( 2009 ) .
y. g. ma , w. q. shen , j. feng , y. q. ma , phys .
c * 48 * , r1492 ( 1993 ) ; z. phys .
a * 344 * , 469 ( 1993 ) ; y. g. ma , w. q. shen , z. y. zhu , phys .
c * 51 * , 1029 ( 1995 ) ; y.g .
yan , x.z .
cai , j.g .
chen , d.q .
fang , w. guo , g.h .
liu , c.w .
shen , y. shi , q.m .
tian , h.w .
wang , k. wang , nucl phys .
a * 787 * , 611c ( 2007 ) .
w. g. lynch , c. b. chitwood , m. b. tsang , d. j. fields , d. r. klesch , c. k. gelbke , g. r. young , t. c. awes , r. l. ferguson , f. e. obenshain , f. plasil , r. l. robinson , a. d. panagiotou , phys .
lett . * 51 * , 1850 ( 1983 ) . y. b. wei , y. g. ma , w. q. shen , g. l. ma , k. wang , x. z. cai , c. zhong , w. guo , j. g. chen , phys . lett .
b * 586 * , 225 ( 2004 ) .
y. g. ma , y. b. wei , w. q. shen , x. z. cai , j. g. chen , j. h. chen , d. q. fang , w. guo , c. w. ma , g. l. ma , q. m. su , w. d. tian , k. wang , t. z. yan , c. zhong , and j. x. zuo , phys .
c * 73 * , 014604 ( 2006 ) ; y. g. ma , x. z. cai , j. g. chen , d. q. fang , w. guo , g. h. liu , c. w. ma , e. j. ma , w. q. shen , y. shi , q. m. su , w. d. tian , h. w. wang , k. wang , y. b. wei , t. z. yan , nucl .
a * 790 * , 299c ( 2007 ) .
m. marqus , r. w. ostendorf , p. lautridou , f. lefevre , t. matulewicz , w. mittig , p. roussel - chomaz , y. schutz , j. qubert , j. daz , a. marn , g. martnez , r. holzmann , s. hlva , a. schubert , r. s. simon , v. wagner , h. lhner , j. h. g. van pol , r. h. siemssen , h. w. wilschut , m. franke , z. sujkowski , phys .
lett . * 73 * , 34 ( 1994 ) ; f. m. marqus , g. martinez , t. matulewicz , r. w. ostendorf , y. schutz , phys .
c * 54 * , 2783 ( 1996 ) .
r. a. lacey , a. elmaani , j. lauret , t. li , w. bauer , d. craig , m. cronqvist , e. gualtieri , s. hannuschke , t. reposeur , a. vander molen , g. d. westfall , w. k. wilson , j. s. winfield , j. yee , s. yennello , a. nadasen , r. s. tickle , and e. norbeck , phys .
* 70 * , 1224 ( 1993 ) . | hard photons emitted from energetic heavy ion collisions are very interesting since they do not experience nuclear interaction , and therefore they are useful to explore properties of nuclear matter .
we investigated hard photon production and its properties in intermediate energy heavy - ion collisions with the help of the blotzmann - uehling - ulenbeck model .
two components of hard photons are discussed : direct and thermal .
the positive directed flow parameter and negative elliptic flow parameter of direct photons are demonstrated and they are anti - correlated to the flows of free protons .
the dependencies of hard photon production and anisotropic parameters on impact parameter , beam energy , nuclear equation of state and symmetry energy are also discussed .
furthermore , we investigated the two - photon momentum correlation function from which the space - time structure information of the photon source could be extracted as well as the two - photon azimuthal correlation which could provide another good method to determine the elliptic flow parameter @xmath0 of direct hard photons . |
the most sophisticated approach to high energy hadronic interactions is the so - called gribov - regge theory @xcite .
this is an effective field theory , which allows multiple interactions to happen `` in parallel '' , with phenomenological objects called _ pomerons _ representing elementary interactions @xcite .
using the general rules of field theory , one may express cross sections in terms of a couple of parameters characterizing the pomeron .
a big disadvantage of grt implementations was so far the fact that cross sections and particle production are not calculated consistently : the fact that energy needs to be shared between many pomerons in case of multiple scattering is well taken into account when considering particle production ( in particular in monte carlo applications ) , but not for cross sections @xcite .
another problem is that at high energies , one also needs a consistent approach to include both soft and hard processes .
the latter are usually treated in the framework of the parton model , which only allows the calculation of inclusive cross sections .
we consider now a new approach called parton - based gribov - regge theory , in order to solve the above - mentioned problems .
we use both the language of pomerons ( as in gribov - regge theory ) in order to calculate probabilities ( and related to this : cross sections ) and the language of strings ( to treat particle production ) .
but we treat these both aspects in a consistent fashion .
this is the really new and attractive feature of our approach .
the price to be paid are great technical difficulties which took us a couple of years to solve them .
for all the details and applications of the corresponding monte carlo program nexus 3 see @xcite .
we will discus the basic features of the new approach in a qualitative fashion .
it is an effective theory based on effective elementary interactions .
multiple interactions happen in parallel in proton - proton collisions .
an elementary interaction is referred to as pomeron , and can be either elastic ( uncut pomeron ) or inelastic ( cut pomeron ) .
the spectators of each proton form a remnant , see fig .
[ allin1]a . -0.5 cm
-1.5 cm -.5 cm since a pomeron is finally identified with two strings , the pomeron aspect ( to obtain probabilities ) and the string aspect ( to obtain particles ) are treated in a completely consistent way .
_ in both cases energy sharing is considered in a rigorous way , and in both cases all pomerons are identical . _ this theory provides also a consistent treatment for hard and soft processes : each pomeron can be expressed in terms of contributions of different types , soft , hard and semihard , cf .
[ allin1]b .
a hard pomeron stands for a hard interaction between valence quarks of initial hadrons .
a semihard one stands for an interaction between sea quarks but in which a perturbative process involves in the middle .
no perturbative process occures at all in soft pomeron .
a pomeron is an elementary interaction . but
those pomerons may interact with each other at high energy @xcite , then they give another type of interaction called _
enhanced diagram_. there are many types of enhanced diagrams depending on the number of pomerons for each vertices and on the number of vertices . in our model , effective first order of triple and 4-pomeron vertices ( y and x diagrams see fig .
[ allin1]c ) are enough to cure unitarity problem which occure at high energy without this kind of diagram @xcite .
indeed , y - type diagrams are screening corrections which are negative contributions to the cross - section .
x diagram is anti - screening .
the inelastic contributions ( cut enhanced diagrams on fig .
[ allin1]d ) of this diagrams contribute to the increase of the fluctuations in particle production .
thanks to a monte carlo , first the collision configuration is determined : i.e. the number of each type of pomerons exchanged between the projectile and target is fixed and the initial energy is shared between the pomerons and the two remnants . then particle production is accounted from two kinds of sources , remnant decay and cut pomeron . a pomeron may be regarded as a two - layer ( soft ) parton ladder attached to projectile and target remnants through its two legs .
each leg is a color singlet , of type q@xmath0 , qqq or @xmath0@xmath0@xmath0 from the sea , and then each cut pomeron is regarded as two strings , cf .
[ nexus1]a .
-1.cm -.5
cm it is a natural idea to take quarks and antiquarks from the sea as string ends for soft pomeron in nexus , because an arbitary number of pomerons may be involved .
in addition to this soft pomerons , hard and semihard pomerons are treated differently . to give a proper description of deep inelastic scattering data , hard and
some of semihard pomerons are connected to the valance quarks of the hadron .
thus , besides the three valence quarks , each remnant has additionally quarks and antiquarks to compensate the flavours of the string ends , as shown in fig .
[ nexus1]c . according to its number of quarks and antiquarks and to the phase space
, a remnant decays into mesons , antibaryons and baryons @xcite .
therefore , from remnant decay , baryon production is favored due to the initial valence quarks and in particular produce strong leading particle effects ( two wings in the rapidity spectra ) for proton and @xmath1 .
considering energy sharing , enhanced diagram and identical elementary interactions , a great number of particle distributions can be calculated within nexus 3 for any kind of hadronic and nuclear interaction @xcite . as an example , fig .
[ proton ] depicts the rapidity spectra for proton and anti - proton production for 2 different energies .
the different contributions for the particle production ( cut pomeron : dashed - dotted line , remnants : dashed line and dotted line ) show that the assimetry between baryon and anti - baryon production comes from the remnant contributions .
proton and anti - proton rapidity spectra for na49 @xcite ( 158 gev lab ) ( top ) and rhic ( 200 gev cms ) ( bottom ) energies .
the remnant contributions are the dashed and the dotted lines , the dashed - dotted line is the pomeron contribution and the full line is the sum . ] at low energy ( 158 gev lab ) , nexus is in good agreement with na49 data @xcite . for rhic energy ( 200 gev cms ) , the prediction shows that the cut pomeron distribution becomes dominant at midrapidity even for the proton , which leads to reduce the assymetry in this region ( ratio closer to 1 ) .
v. n. gribov , sov .
jetp * 26 * , 414 ( 1968 ) [ zh . eksp .
53 * , 654 ( 1968 ) ] .
m. baker and k. a. ter - martirosian , phys .
* 28 * , 1 ( 1976 ) . v. a. abramovsky and g. g. leptoukh , sov .
j. nucl .
phys . * 55 * , 903 ( 1992 ) .
h. j. drescher _ et al .
_ , phys .
rept . * 350 * , 93 ( 2001 ) .
a. b. kaidalov , l. a. pomomarev , and k. a. ter - martirosian , sov .
j. nucl.phys * 44 * , 468 ( 1986 ) . | we recently proposed a new approach to high energy nuclear scattering , which treats hadronic collisions in a sophisticated way . demanding theoretical consistency as a minimal requirement for a realistic model
, we provide a solution for the energy conservation , screening problems and identical elementary interactions , the so - called `` parton - based gribov - regge theory '' including enhanced diagrams .
we can now present some of our results for sps and rhic energies . |
observations indicate that the earth was at least warm enough for liquid water to exist as far back as 4 gyr ago , namely , as early as half a billion years after the formation of the earth @xcite ; in fact , there is evidence suggesting that earth may have been even warmer then than it is now @xcite .
these relatively warm temperatures required on early earth are in apparent contradiction to the dimness of the early sun predicted by the standard solar models .
this problem has generally been explained by assuming that earth s early atmosphere contained huge amounts of carbon dioxide ( co@xmath7 ) , resulting in a large enough greenhouse effect to counteract the effect of a dimmer sun .
however , the recent work of @xcite places an upper limit of 0.04 bar on the partial pressure of co@xmath7 in the period from 2.75 to 2.2 gyr ago , based on the absence of siderite in paleosols ; this casts doubt on the viability of a strong co@xmath7 greenhouse effect on early earth .
the existence of liquid water on early mars has been even more of a puzzle ; even the maximum possible co@xmath7 greenhouse effect can not yield warm enough martian surface temperatures @xcite .
these problems can simultaneously be resolved , for both earth and mars , if the early sun were brighter than predicted by the standard solar models .
this could be accomplished if the early sun were slightly more massive than it is now .
helioseismic observations provide revolutionary precision for probing the solar interior .
helioseismic frequencies are observed with an accuracy of a few parts in @xmath8 , allowing measurement of the sound speed profile throughout most of the sun s interior to an accuracy of a few parts in @xmath9 @xcite .
this high precision permits one to search for subtle effects in the interior structure of the present sun resulting from events in the distant past .
in particular , modest mass loss ( between 1% and 7% of the sun s mass ) early on the main sequence might have left enough of a fingerprint on the interior structure of the present sun to be detectable by helioseismological observations .
@xcite first presented the hypothesis that stars like the sun might lose significant amounts of mass on the early main sequence .
@xcite were the first to compute solar models with such early main sequence mass loss , namely , an extreme case with an initial mass of @xmath10 .
such extreme mass loss ( of @xmath11 ) turns out to be unrealistic , as discussed below ; but small mass loss cases can not be ruled out at the present . @xcite considered an initial solar mass of @xmath12 , showing that this is the upper limit allowed by the observed solar lithium depletion .
some work has been carried out recently attempting to use helioseismology to constrain early solar mass loss @xcite ; however , the first of these used rather crude solar interior models , and the second yielded ambiguous results , as we discuss in detail below .
@xcite were the first to attempt to use helioseismic observations to constrain early solar mass loss ; they considered initial solar masses of 1.1 and @xmath10 , concluding that the @xmath10 could be ruled out by helioseismic observations of low - degree modes .
they also claimed that a mass loss timescale of 0.2 gyr was favored over a 0.45 gyr timescale for the @xmath12 case , based on `` frequencies for modes that probe just below the convection zone bottom ( @xmath13 , @xmath14hz ) , '' but this latter conclusion is rendered dubious by the fact that this is the particular region of the sun where one should expect the _
worst _ agreement in models that do not attempt to include rotation - induced mixing ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
furthermore , these models of @xcite were based on relatively crude approximations to the physical processes in the solar interior , particularly for the opacities .
they state that their solar interior opacities used the @xcite analytical fits to the opacity tables of @xcite , with the electron scattering ( esk ) term in the iben fit being multiplied either by a factor of 1.5 to approximate the laol tables , or by a factor of 2 to approximate the ( early ) opal opacities of @xcite ; in addition , they adjusted the opacities _ separately _ for each of their solar models , by adjusting the @xmath15 term in the iben fit ( which mainly affects opacities at @xmath16k ) in order to yield a position for the base of solar envelope convection at @xmath17 @xcite .
even if they had in fact used the early opal opacity tables of @xcite , the models of @xcite suggest that this would be expected to yield solar models with helioseismic disagreements 2 or 3 times worse than can be achieved by models using the more recent opal opacities @xcite ; we have shown that using the laol opacity tables would result in even larger errors @xcite .
@xcite used the recent mhd equation of state tables @xcite , albeit for a fixed value of @xmath18 ( the metallicity @xmath19 actually varied from @xmath20 to @xmath21 from the envelope to the center of their models ) , but the equation of state is not very sensitive to metallicity , and we estimate that this fixed-@xmath19 approximation introduced errors in their thermodynamic quantities of no more than a few parts in @xmath22 .
they considered diffusion of hydrogen , helium , and four of the cno isotopes , using diffusion constants from @xcite ; as they point out , the fact that this treatment probably overestimates the extent of gravitational settling of heavy elements @xcite may compensate for the fact that they considered diffusion for elements comprising only 62% of the metallicity @xmath19 .
finally , @xcite compared models which had not been converged to the same final surface @xmath23 value ; their @xmath23 differences of up to 4% would by themselves yield significant ( though not major ) differences between the models , largely through the effect on the opacities @xcite .
@xcite also used helioseismology to test an initial solar mass of @xmath12 , comparing sound speed and density profiles in their solar models to profiles obtained via an inversion from helioseismic frequencies .
their solar models incorporated much more up - to - date input physics ( though they also used the fixed-@xmath19 approximation in the equation of state , with @xmath24 ) , and they tested some different formulations of the input physics ; but their tests of the effects of early solar mass loss yielded results that were ambiguous at best .
they first considered models using the ceff equation of state @xcite , diffusion constants from @xcite , and the more recent opal opacities @xcite ; however , in these models @xcite also used the fixed-@xmath19 approximation when interpolating the _ opacities _ , which can lead to significant opacity errors and thus significant effects on the solar models , as they themselves showed ( see also * ? ? ?
@xcite compared mass - losing models with an initial mass of @xmath25 to a standard solar model ( @xmath26 ) , testing both a short ( 0.2 gyr ) and a long ( 0.45 gyr ) mass loss timescale .
they found that the long mass loss timescale yielded fractional shifts of up to @xmath27 in the sound speed profile relative to the standard solar model ( a significant but not major effect ) , while the short mass loss timescale had an insignificant effect ( of @xmath28 , roughly one fifth as large ) .
they next considered models with the opal equation of state @xcite , and where the variation of metallicity @xmath19 inside the star _ was _ considered when interpolating in the opal opacities ( each of these latter two changes yielded shifts of up to 0.001 in the sound speed profiles of their standard solar models ) .
they compared a model with an initial mass @xmath25 and their short ( 0.2 gyr ) mass loss timescale to a standard solar model , finding differences of up to @xmath27 in the sound speed profile between them 5 times as large as the effect from the same comparison with the first set of models ( and large enough to suggest the possibility of highly significant effects from a longer mass loss timescale ) .
such large disagreements between their two sets of results for the differential effects of early solar mass loss suggest that their mass - losing models may have had much larger numerical or convergence errors than they estimated for their standard solar models ; and indeed their second short - timescale mass - losing model shows sound speed disagreements near the surface that might be interpreted as being due to poor convergence to the present solar luminosity , radius , and/or surface @xmath23 though even very poor convergence would not expected to yield such large interior disagreements @xcite .
the goal of the present work was to resolve the above ambiguity as to what constraints might be placed on early solar mass loss by helioseismic observations , by computing models with both up - to - date input physics and high numerical precision .
we considered additional observational constraints on solar mass loss ( as discussed in
[ sssec : mdotobs ] below ) , and tested a number of different initial solar masses and mass loss timescales .
we also compared the effects of mass loss to the effects of other reasonable variations in the solar input parameters and input physics these latter effects were discussed in detail in our companion paper `` our sun iv '' @xcite .
finally , we present the effects of the various mass loss cases on the solar flux reaching earth and mars as a function of time .
presently , the sun is experiencing only a negligible amount of mass loss : the solar wind removes mass at a rate @xmath29 yr@xmath30 .
if this mass loss rate had been constant over the last 4.5 gyr , the young sun would have been more massive by only @xmath31 .
the contemporary solar wind has been observed only for a few decades , and has been found to be a highly variable phenomenon all properties , including flux , velocity , and composition vary significantly @xcite .
the lunar surface material carries the signature of the solar wind irradiation over the past several gyr ; measurements of noble gas isotopes in lunar samples suggest that the average solar wind flux over the past @xmath4 gyr was an order of magnitude higher than it is today @xcite .
this implies a total solar mass loss of @xmath32 over the past 3 to 4 gyr ( the age of the oldest available lunar material ) .
some older , solar - flare irradiated grains from meteorites imply an early solar flare activity about @xmath22 times that of the present sun @xcite ; the associated solar wind may have been enhanced by a similar factor of @xmath33 , most likely during the first @xmath34 gyr of the sun s life on the main sequence @xcite , implying a total mass loss during this first 1 gyr period of as much as @xmath35 ( if the average mass loss rate throughout that period was indeed @xmath33 times the present rate of @xmath36yr@xmath30 ) .
such a change in the solar mass would be sufficient to cause a significant increase in the luminosity of the young sun . since the sun is a typical main sequence star , it is reasonable to assume that mass loss rates in the young sun would be similar to those in other young solar - type main sequence stars .
there have been several attempts to measure mass loss in early main sequence stars .
it is observationally a very challenging task .
@xcite attempted to obtain mass loss rates for 17 young main sequence stars somewhat hotter and more massive than the sun ( a and f dwarfs ) , finding upper limits to the mass loss rates of @xmath37 to @xmath38/yr ; these limits are even less constraining than the highest solar mass loss rate suggested by the meteoritic and lunar data .
@xcite used 3.6 cm vla observations to place more stringent upper limits of @xmath39yr@xmath30 on mass loss rates of three young main sequence stars of roughly solar mass ( @xmath40 uma , @xmath41 cet , and @xmath42 com ) as discussed in
[ ssec : mdotused ] , the largest initial solar mass that we consider , namely @xmath1 , would require early solar mass loss rates that are marginally consistent with these more stringent limits .
@xcite recently obtained hst observations of ly@xmath43 absorption from the region where the stellar wind collides with the interstellar medium , using these to measure the stellar wind from the sun - like star @xmath43 cen .
they found a mass loss rate roughly twice as large as that of the sun ; note that @xmath43 cen is slightly _ older _ than the sun ( with an age of @xmath44 gyr ) , as well as being very slightly more massive ( @xmath45 ) .
they also found an upper limit roughly ten times lower for its cooler , less massive companion proxima cen .
a similar method had earlier been used by @xcite to look at four other main sequence stars cooler and less massive than the sun ( finding stellar winds of roughly the same order of magnitude as the solar wind ) .
such a method applied to young , sun - like stars holds promise for placing stringent limits on early main sequence mass loss .
very recently , @xcite demonstrated that mass loss rates in gk dwarfs ( measured using the above method ) are generally correlated with their x - ray flux ( measured by rosat , and using the hipparcos distances ) . for 7 of the 9 stars they looked at , there was a fairly tight correlation , with @xmath46 over a range of 2 orders of magnitude in @xmath47 and @xmath48 .
they then used the estimates from @xcite of the correlation of x - ray flux with rotational velocity @xmath49 and of rotational velocity with age @xmath50 to obtain the first empirically derived relation describing the mass - loss evolution of cool main - sequence stars like the sun , namely , @xmath51 .
they point out that the above correlations do not really apply to very young stars ( age@xmath52 gyr ) , and assume instead that the maximum mass loss rate at early times corresponds to the maximum x - ray flux observed in sun - like stars , namely , @xmath22 times the solar value .
given the sun s present mass loss rate of @xmath53yr@xmath30 , this would imply total solar mass loss of order @xmath54 ( albeit with an uncertainty of a factor of 5 or so ) , with most of this mass loss taking place in the first fraction of a gyr of the sun s lifetime .
the effects of such a mass loss law will be investigated in more detail in a future paper , but we make some rough estimates of the probable effects in [ sec : results ] and [ sec : concl ] .
the observed depletion of lithium in the sun provides a stringent upper limit to the total solar mass loss of @xmath55 ; i.e. , the initial solar mass @xmath56 ( 4.5 gyr ago ) is constrained to be @xmath57 @xcite . however , this is much too generous an upper limit .
there are additional mechanisms that can deplete solar lithium .
one mechanism , namely , pre - main - sequence lithium depletion ( during the sun s initial contraction phase ) , was taken into account in our mass - losing solar models ( for the standard sun , this depletion was a factor of @xmath58 , as discussed below and in our companion paper `` our sun iv '' [ @xcite ] ) .
another mechanism is rotation - induced turbulent mixing , which probably is the major cause of the main - sequence lithium depletion ; however , rotation models have free parameters , and can be fitted to _
any _ required amount of solar lithium depletion ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
also , it has been shown that mass loss can not be the major contributor to the observed lithium depletions in the young hyades cluster @xcite ; this constraint is discussed in more detail in [ sssec : hyrest ] below .
an even more stringent upper limit to the sun s initial mass is imposed by the requirement that the early earth not lose its water via a _ moist greenhouse effect _ , which would occur if the solar flux at earth were more than 10% higher than its present value @xcite a moist greenhouse occurs when the stratosphere becomes wet , and h@xmath7o is lost through uv dissociation and the subsequent loss of hydrogen to space .
this solar flux limit corresponds to an upper limit on the sun s initial mass of @xmath59 , which is the most stringent upper limit on the sun s initial mass .
the only strong lower limit on @xmath56 comes from the fact that the sun is converting matter into energy and radiating it away ; @xmath60 , where @xmath61 is the total energy radiated away , @xmath62 is the average solar luminosity ( including the neutrino luminosity ) @xmath63 is the @xmath64 gyr duration of the nuclear burning , @xmath65 is the amount of mass converted into energy , and @xmath66 is the speed of light ( note that elsewhere in the paper we use `` @xmath66 '' to denote the adiabatic sound speed ) . at present , mass is radiated away as photons and neutrinos at a rate slightly over twice the solar wind mass loss rate . for the standard solar model , the sun s average luminosity over the last 4.5 gyr was about 0.85 times its present luminosity .
it follows that @xmath67 from radiation losses alone ( i.e. , that @xmath68 ) .
such a minor amount of mass loss has a negligible effect on the early solar luminosity .
there are also considerations that put soft lower limits on the sun s initial mass @xmath56 .
if the present observed solar wind rate of @xmath29/yr had been constant over the sun s history , the total amount of solar wind mass loss would have been only @xmath69 ; including the @xmath67 from radiation losses would imply @xmath70 . however , measurements of the noble gases implanted in lunar samples suggest an average solar wind flux over the past @xmath4 gyr an order of magnitude higher than at present , @xcite , implying a total solar mass loss over that period of @xmath71 , i.e. , a solar mass 3 gyr ago of @xmath72 note that @xmath73 is used to refer to the solar mass at @xmath74 years before the present . the @xmath4 gyr age of these lunar rocks means that they place no limits on _ earlier _ solar mass loss , so that all one can say is that @xmath75 .
older , solar - flare irradiated grains from meteorites imply early solar flare activity about @xmath22 times that of the present sun @xcite , which might possibly correspond to similarly high mass loss rates during the first @xmath34 gyr period of the sun s life , but can not be used to provide any sort of limit . the semi - empirical mass loss formula recently presented by @xcite suggests total solar mass loss of order @xmath54 , but with an uncertainty of about a factor of 5 , as discussed above .
another limit on the sun s initial mass comes from the requirement that mars was warm enough for liquid water to exist 3.8 gyr ago ( at the end of the late heavy bombardment period ) . according to @xcite and @xcite , this requires a solar flux ( at mars ) 3.8 gyr ago at least 13% larger than that from the standard solar model , in order to make it possible for a co@xmath7 greenhouse effect on mars to be able raise the temperature to @xmath76c .
such an increase in flux would correspond to a mass of the sun at that time of @xmath77 .
since the lunar rock measurements constrain the sun s mass @xmath4 gyr ago to be @xmath72 , the sun s average mass loss rate between 3.8 and 3 gyr ago would be @xmath78/yr .
if this same mass loss rate also occurred throughout the period from the sun s birth @xmath79 gyr ago until 3.8 gyr ago , this would imply an initial solar mass of @xmath80 .
note that this lower limit assumes that the only greenhouse effect on early mars is due to co@xmath7 .
if a smog - shielded ammonia greenhouse could exist on early mars , such as that proposed for the early earth by @xcite , then this lower limit on @xmath56 might be softened or eliminated .
@xcite established that mass loss could not be the _ major _ cause of the main - sequence lithium depletion for stars in the hyades cluster .
their result has frequently been mis - quoted and misunderstood ; it has often been used to rule out the possibility of mass loss during the sun s early main sequence phase . however
, their result does _ not _ rule out relatively small amounts of mass loss for either the hyades or the sun . for the hyades cluster , which is 0.6 gyr old , lithium abundances in many stars
have been observed , exhibiting a fairly tight relationship between a star s lithium abundance and its surface temperature the observed lithium abundance drops off steeply with decreasing surface temperature , below @xmath81 k. @xcite considered lithium depletion due both to pre - main - sequence burning and to main sequence mass loss .
they found that the observed lithium - temperature relationship could not be accounted for by pre - main - sequence lithium depletion alone , but that it could be accounted for fairly well if one added main sequence mass loss .
however , they found that all the stars with surface temperatures below 5500 k would then have to have nearly identical initial masses ( with a wide range of mass loss rates ) .
such a distribution of initial stellar masses , with a high , narrow peak in the distribution near @xmath82 , is unrealistic .
this argument has been widely misquoted , to rule out early main sequence mass loss in stars ( including the sun ) .
the swenson - faulkner conclusion applies _ only _ if one is trying to match the hyades lithium depletions _ without including rotation - induced mixing_. as soon as one includes the latter as a major component , one can reproduce the observed lithium - temperature relation of hyades stars by choosing suitable values for the adjustable parameters in the rotational mixing formalism ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
stellar rotation is ubiquitous in young stars , and is commonly assumed to be the cause of all main sequence lithium depletion ; the presence of a relatively small amount of mass loss merely requires that the large lithium depletion due to rotation be decreased by a small amount ( by small changes in the adjustable parameters for rotational mixing ) . for the hyades , even a mass loss as large as @xmath83 in a star near @xmath84 would imply a lithium depletion factor due to mass loss alone of only @xmath44 , ( according to the models of * ? ? ?
* ) , and would still require a depletion factor due to rotational mixing of @xmath85 in order to reproduce the observed lithium - temperature relation . for the sun , pre - main sequence burning yields a lithium depletion factor of order 20 @xcite ; as discussed in [ sssec : lithium ] , combining pre - main sequence lithium burning with an early solar mass loss of @xmath83 increases this by less than a factor of 2 , i.e. , a combined lithium depletion factor of @xmath86 , still much smaller than the total observed lithium depletion of @xmath87 @xcite
rotation would be responsible for the remaining lithium depletion .
mass loss of order @xmath83 or less _ is consistent _ with the hyades lithium observations , i.e. , does not require an unrealistic initial stellar mass distribution . in their later work ,
@xcite found they were able to reproduce the hyades lithium - temperature relation by pre - main - sequence depletion alone , provided that the oxygen abundance was assumed to be at the upper limit of the observed range ( and using the most up - to - date opal and alexander opacities ) .
however , other clusters such as ngc 752 , m67 , or ngc 188 _ can not _ be explained via pre - main - sequence lithium depletion alone their observed lithium depletions are much _ larger _ than those of the hyades ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) , while their pre - main - sequence depletions would be significantly _ smaller _ ( due to their lower metallicities ) . the same is true of the sun . in other words , the observed lithium depletions demand main sequence depletion , which would be largely due to rotation - induced mixing , with possibly a small effect from mass loss as well .
we computed mass - losing solar models having initial masses @xmath0 , 1.02 , 1.03 , 1.04 , 1.05 , 1.06 , and @xmath1 ; as discussed in
[ sssec : mdotobs ] , an initial mass of @xmath1 is the upper limit consistent with the requirement that the young earth not lose its water via a moist greenhouse effect @xcite .
we considered three different forms for early solar mass loss , which we call `` exponential , '' `` step - function , '' and `` linear '' ; for each of these , two limiting cases are displayed in figure [ fig : mdot ] ( i.e. , cases with initial solar masses of 1.01 and @xmath88 ) . in the `` exponential '' mass loss case ,
the mass loss rate starts out high and declines exponentially , with an initial mass loss rate @xmath89 and decay time constant @xmath90 chosen such as to give the present observed solar mass loss rate of @xmath91yr@xmath30 at the sun s present age .
in other words , @xmath92 , with @xmath93 to @xmath94yr@xmath30 and @xmath95 to 0.551 gyr , for initial solar masses of 1.01 to @xmath88 , respectively .
this exponential mass loss case is our most conservative one : a simple mass loss scenario that is still reasonably consistent with the observed lunar mass loss constraint .
this scenario yields average mass loss rates over the past 3 gyr of 13 to 42 times the present value , for initial solar masses of 1.01 to @xmath88 , respectively .
these average mass loss values are reasonably consistent with measurements of noble gases in lunar rocks , which suggest a mass loss rate an order - of - magnitude higher than the present value .
since figure [ fig : mdot ] is a log - linear plot , these `` exponential '' mass loss cases appear as straight lines .
the `` step - function '' mass loss rate was chosen to have a constant high value during the period before the lunar rock observations apply , namely , the first 1.6 gyr of the sun s life ; over the remaining 3 gyr of the sun s life ( up to the present ) , a mass loss rate averaging ten times the present value was assumed , declining linearly over this period to reach the present solar mass loss rate at the present solar age .
this is the most extreme case which is still consistent with the observed lunar rock mass loss constraints : it keeps the solar flux as high as possible for as long as possible . for the first 1.6 gyr , this scenario has constant mass loss rates of @xmath96 to @xmath97yr@xmath30 , for initial solar masses of 1.01 to @xmath88 , respectively . in the `` linear '' mass loss case ,
the mass loss rate starts out high and declines slowly and linearly , to reach the present solar mass loss rate at the present solar age .
this was chosen as our most radical case , with maximum impact on the sun s internal structure .
due to the linear decline , the mass loss rate remains high throughout most of the sun s lifetime , remaining of the same order as the initial mass loss rate ( @xmath98 to @xmath99yr@xmath30 , for initial solar masses of 1.01 to @xmath88 , respectively ) . during the past 3 gyr ,
the mass loss rate for this `` linear '' case is much higher than for the other mass loss cases above , violating the observed lunar mass loss constraints ( the `` linear '' case has average mass loss rates over the past 3 gyr of 50 to 330 times the present rate , for initial solar masses of 1.01 to @xmath88 , respectively ) .
the `` linear '' mass loss cases appear as curved lines in figure [ fig : mdot ] .
the solar evolution program is discussed in detail in our companion paper `` our sun iv '' @xcite ; we provide only a brief summary here .
we used the opal equation of state @xcite , extended to lower temperatures by the mhd equation of state @xcite .
the 1995 opal opacities @xcite were used for @xmath100 ; since these opacities ( as well as both sets of equation of state tables ) were based on the heavy element composition of @xcite , this mixture was used in order to obtain self - consistent solar models ( along with their recommended value @xmath101 in the present solar envelope ) . at lower temperatures ( @xmath102 ) ,
the @xcite molecular opacities were used . both the equation of state and the opacities were interpolated in metallicity as well as in hydrogen abundance , temperature , and density , in order to take into account metallicity variations due to diffusion and nuclear burning . for the `` exponential '' mass loss cases
, we also computed an alternate set of solar models based on the recent solar composition observations of @xcite , which yield @xmath103 in the present solar envelope ; for this composition , we obtained appropriate opal opacities via the online opacity computation feature of the opal web page[multiblock footnote omitted ] .
also , in these models we performed the switchover between the opal and mhd equations of state at @xmath104 ( rather than at @xmath105 , which corresponds to @xmath106 in the sun ) .
as discussed in @xcite , the switchover region between the two equations of state was wide enough , and the equations of state were similar enough in the chosen switchover regions , that artifacts introduced by the switchover should not be significant compared to the size of the inconsistencies in the opal equation of state .
we used the nacre nuclear reaction rate compilation @xcite , supplemented by the 7be electron capture rates of @xcite .
weak screening @xcite was used note that it is a very good approximation to the exact quantum mechanical solution for solar conditions ( see , e.g. , * ? ? ?
* ; * ? ? ?
all of the stable isotopes up to and including o were considered in detail , except for deuterium ( which was assumed to have been burned to 3he ) .
the other isotopes up to were included in the code , but not in the nuclear reaction network , since there are no significant effects under solar conditions ( except for f , which was assumed to be in cno - cycle nuclear equilibrium for nuclear rate purposes ) .
neutrino capture cross sections were taken from @xcite , except for the 8b - neutrino cross section for capture on , where the more recent value ( 5% higher ) of @xcite was used .
a set of subroutines were kindly provided to us that take into account the diffusion ( gravitational settling ) of helium and heavy elements relative to hydrogen ( see also * ? ? ?
* ; * ? ? ?
a present solar mass of @xmath107 g @xcite was used , and a solar radius at the photosphere ( @xmath108 ) of @xmath109 mm @xcite .
our solar luminosity value of @xmath110 erg s@xmath30 @xcite is close ( less than 1-@xmath111 ) to the more recent value of @xcite ; as discussed in our companion `` our sun v '' paper @xcite , such a luminosity difference has negligible effect on the solar structure ( and only a minor effect on the neutrino rates ) .
we used a total solar age of @xmath112 gyr , measured from the sun s birth on the pre - main - sequence hayashi track ; this is only just outside the range 4.55 gyr@xmath113 gyr ( in effect , @xmath114 gyr ) allowed by meteoritic ages @xcite , sufficiently close that a more precise age value would have had very little effect on the solar structure and helioseismology ( see * ? ? ?
note that the earlier total solar ( and solar system ) age estimate @xmath115 gyr of @xcite is consistent with the limits of @xcite .
our models took about 40 myr to reach the zero age main sequence ( zams ) , the point at which nuclear reactions in the core provide essentially all the sun s luminosity , and the pre - main - sequence contraction stops ; this pre - main - sequence timescale implies that the _ total solar age _
@xmath116 used in this paper can be converted into a _ main sequence _ solar lifetime by subtracting about 0.04 gyr ( this was also pointed out by * ? ? ?
we investigated the effects of using two different zonings .
our coarse - zoned solar models had about 2000 spatial zones in the model , and about 200 time steps in the evolution from the zero - age main sequence to the present solar age ( plus about 800 time steps on the pre - main - sequence ) , comparable to other work in this field ( e.g. , a factor of 2 more than * ? ? ? * ) ; mass - losing solar models might require several times as many timesteps , due to the constraint that mass loss effects be kept small over one timestep .
typically , these models were converged to match the solar luminosity , radius , and surface @xmath23 value to within a few parts in @xmath8 ; a few cases where convergence was slow reached only about a part in @xmath9 .
because the inconsistencies in the mass - losing solar models of @xcite had suggested the possibility that such models might tend to have numerical inaccuracies larger than in standard solar models , as discussed in [ sssec : mdotmodelsun ] , we also computed a number of fine - zoned models , with @xmath117 spatial zones and 1500 main - sequence time steps ( plus 6000 pre - main - sequence time steps ) a factor of 5 increase in both spatial and temporal precision .
it was indeed more difficult to converge mass - losing solar models to precise solar luminosity , radius , and surface @xmath23 values the standard solar models were converged to these values with an accuracy nearly a factor of ten better than the typical mass - losing models but comparing the last few models in a convergence sequence revealed no untoward effects . considering the much larger amounts of cpu - time required , there seemed to be no point in attempting to match the solar luminosity , radius , and surface @xmath23 values any better for fine - zoned cases than for coarse - zoned ones .
( a fine - zoned converged mass - losing solar model took a few weeks of cpu - time on a fairly high - performance es40 computer , as compared to about half a day for a coarse - zoned case , while a standard solar model without mass loss took only about a fifth as long in each case ; these times were roughly tripled on a 450 mhz pentium iii pc . )
as discussed in our companion `` our sun iv '' paper @xcite , even the worst of the above convergence accuracies has a negligible effect on the solar sound speed profile : up to 1 or 2 parts in @xmath9 in the convective envelope , and a few parts in @xmath8 below it .
the fine - zoned mass - losing models differed from the corresponding coarse - zoned ones by the same negligibly small systematic shift ( less than a part in @xmath9 in the sound speed ) that had been found in standard solar models ( see * ? ? ? * ) ; we concluded that our properly converged coarse - zoned mass - losing models are perfectly adequate , and do not suffer from the problems encountered by the mass - losing solar models of @xcite .
even the largest of the mass loss rates we considered ( 3 orders of magnitude larger than the present solar wind ) is still quite small in absolute terms ( for example , asymptotic giant branch stars encounter mass loss rates 6 orders of magnitude larger still , and high - mass main sequence stars also have very high mass loss rates ) .
nonetheless , over a reasonable model timestep , such mass loss would cause a mass layer near the solar surface to move outwards significantly .
we therefore used an outer boundary condition for our models at a point @xmath118 inwards from the solar surface , computing sets of static envelopes for the region exterior to this point ( and requiring timesteps short enough that changes at this outer boundary point remained small from one timestep to the next ) .
since the mass loss timescales for these outer layers are always at least 4 orders of magnitude longer than their thermal timescales , this should be a good approximation ; and ignoring energy changes in this small outer region should introduce fractional errors in the solar luminosity at the level of @xmath119 or less .
test cases where this outer boundary point was chosen differently showed no effect , as expected .
we compared our solar models to profiles of the solar sound speed @xmath120 , density @xmath121 , and adiabatic index @xmath122 obtained from the helioseismic reference model of @xcite , which they obtained by inversion from the helioseismic frequency observations . in the inversion process
, a standard solar model is required , but @xcite demonstrated that the resulting @xmath120 and @xmath121 profiles of the helioseismic reference model are relatively insensitive to uncertainties in the standard solar model used for this purpose ( except for uncertainties in @xmath123 ) .
they estimated a net uncertainty of few parts in @xmath9 for the sound speed @xmath120 and adiabatic index @xmath122 , and a few parts in @xmath22 for the density @xmath121 .
however , in the sun s core ( @xmath124 ) , systematic uncertainties in the helioseismic sound profile are increased by a factor of @xmath44 ; this was demonstrated by @xcite , who compared helioseismic inversions of different helioseismic data sets .
we used their comparison to estimate the @xmath125-dependence of the systematic error in @xmath120 in the core and in the convective envelope ( namely , a fractional systematic error decreasing linearly from 0.0013 at @xmath126 to 0.0003 at @xmath127 , constant from there to @xmath128 , then increasing linearly to 0.00052 at @xmath129 ) . for @xmath120
, this systematic error can be significantly larger than the statistical errors quoted in table 2 of @xcite , and we combined the two in quadrature to get the total fractional error @xmath130 for the purpose of calculating weighted rms differences the rms fractional difference in @xmath66 is given by
@xmath131 ^ 2 \right\ } / \left\ { \sum \left [ 1 / ( \sigma_c / c ) \right]^2 \right\ } \right)^{1/2}$ ] . for @xmath122 and @xmath121 ,
the systematic errors are comparable to or smaller than the statistical ones , and the statistical errors sufficed for calculating weighted rms differences .
we present all our sound speed and density profiles in terms of differences relative to the observed helioseismic reference profiles .
this choice of presentation not only allows one to see the effects of the choice of initial mass and mass loss type , but also the extent to which the models agree with the helioseismic observations .
for comparison with our solar mass loss rates , we used the most recent observed upper limits on stellar mass loss rates from three young sun - like stars ( namely , @xmath39yr@xmath30 , from @xmath40 uma , @xmath41 cet , and @xmath42 com ) , as presented by @xcite .
even our highest mass loss cases are very close to being consistent with these limits , as is illustrated in figure [ fig : mdot ] .
we present in figure [ fig : helio ] profiles of the adiabatic sound speed differences @xmath132 ; profiles of the density differences @xmath133 are available online[multiblock footnote omitted ] .
note that we use `` @xmath134 '' to denote differences between the helioseismic profile and one of our models , and `` @xmath135 '' to denote differences between two of our models with different input parameters the `` @xmath134 '' values are the profiles plotted in figure [ fig : helio ] , while the `` @xmath135 '' values refer to the difference between one plotted curve and another .
solar masses as a function of time for the corresponding cases are presented in figure [ fig : mass ] , and solar fluxes at the planets ( relative to their present values ) are presented in figure [ fig : solarflux ] . since the prominent peak at @xmath136 results from the neglect of rotational mixing
, we did not require agreement in this region between profiles from our theoretical models and profiles inferred from the helioseismic observations . nor did we require agreement in core region , since the present helioseismic observations still result in large uncertainties in the inferred profiles there . on the other hand
, we aimed for agreement in the regions @xmath137 and @xmath138 , where disagreements are due to imperfections in the input physics or uncertainties in the observed solar parameters .
this is demonstrated by our variant models of our companion paper `` our sun iv '' @xcite and of @xcite and @xcite .
except around @xmath136 and near the surface , the sound speed profile in our standard solar models with the @xcite composition ( _ heavy solid lines _ in fig .
[ fig : helio]a , b , c ) differ by @xmath139 from the helioseismic sound speed , typical agreement for standard solar models with this composition ( see , e.g. , * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
however , if one updates to the more recent solar composition value of @xcite , there is a systematic shift in the sound speed profile in the interior that worsens agreement with helioseismology by a factor of @xmath5 ( _ heavy solid line _ in fig .
[ fig : helio]d differences near the solar surface are due to extending the use of the opal equation of state to lower temperatures in this model ) .
this composition effect was also demonstrated by solar models of @xcite .
even larger systematic effects are allowed by the quoted uncertainties in the solar composition , and ( to a lesser extent ) the uncertainties in the opal opacities , the @xmath6 nuclear reaction rate , and the diffusion constants for gravitational settling , as shown in detail in our companion paper `` our sun iv '' @xcite . in figure
[ fig : helio ] , the _ filled circles _ ( connected by _
light solid lines _ ) indicate the extremes of the range of @xmath140 profiles found in @xcite for solar models with `` reasonable '' variations in these input parameters ( i.e. , variations allowed by the quoted uncertainties ) . in the comvective envelope ,
the largest effects on the sound speed profile arises from the uncertainties in the observed solar radius and in the equation of state ; in the bulk of the solar interior , the largest effects arise from uncertainties in the observed solar surface composition ; in the core , the largest effect arises from the uncertainty in the @xmath6 nuclear reaction rate @xcite .
figure [ fig : helio]a , d demonstrates that all of our `` exponential '' mass loss models agree better with the helioseismic observations than the standard solar model , which has no mass loss ( i.e. , they lie closer to the zero line of perfect agreement ) .
however , this improvement is clearly not significant : even for our most extreme `` exponential '' mass loss case ( with @xmath141 ) , the effect is only about one third as large as the maximum effect allowed by variations in other input solar parameters ( also shown in fig .
[ fig : helio ] ) .
the rms sound speed differences provide a numerical measure of the extent of the above agreement between a given theoretical model and the profile inferred from helioseismic observations ; these rms values for each of our mass - losing cases are given in table [ tab : results ] ( for completeness , the fine - zoned cases are presented as well as the coarse - zoned ones , although the latter proved to be quite accurate enough , as discussed in
[ ssec : inputs ] ) .
the differences rms@xmath142 relative to the helioseismic sound speed profile are shown both for the entire sun and for the region @xmath143 ( where the latter region excludes the peak near @xmath136 that results from neglect of rotation - induced mixing , and the envelope which is sensitive mostly to the solar radius and the equation of state ) ; unlike the sound speed , the density profile does not respond in a localized way to local variations in the solar models , so only the rms@xmath144 over the entire sun is shown .
the differences rms@xmath145 ( and rms@xmath146 ) relative to the corresponding theoretical standard ( non - mass - losing ) solar model are shown for the entire sun , since they measure the differential effects of mass loss .
table [ tab : results ] shows that the `` exponential '' @xmath141 cases have rms@xmath142 errors about 30% smaller than the corresponding standard solar models , but again this difference is not significant compared to the large variations that can arise from uncertainties in the other input solar parameters ( shown in the last line of table [ tab : results ] ) .
this is perhaps most easily seen by comparing the relative rms measures : the `` exponential '' cases differ by rms@xmath147 from their corresponding standard solar model , while the shifts in the sound speed profile from uncertainties in other input parameters can be as large as rms@xmath148 , nearly three times as large . in summary , none of our `` exponential '' mass loss cases with initial solar masses @xmath149 can be ruled out by helioseismological observations ; in fact , all other things being equal , these helioseismological observations slightly favor the `` exponential '' mass loss case with highest of our initial solar masses ( @xmath141 ) , although this is not statistically significant , as shown by figure [ fig : helio]a , d and the rms values in table [ tab : results ] .
detailed investigation of the effects of the semi - empirical mass loss law @xmath51 recently presented by @xcite and discussed in [ sssec : mdotobs ] above must wait for a future paper .
however , the total amount of mass lost by the sun would be comparable to or smaller than that in our `` exponential '' mass loss case , and the evolution of the mass loss rate as a function of time would be not too dissimilar ( in both cases , most of the mass loss occurs at quite early times ) .
thus a @xmath150 mass loss case should yield an effect on the solar sound speed profile similar to ( or smaller than ) that of our `` exponential '' mass loss cases of figure [ fig : helio]a , d. figure [ fig : helio]b presents our results for our `` step - function '' mass loss models ; recall that this type of mass loss was chosen as the most extreme possibility that remains consistent with all the observational mass loss constraints ( see [ ssec : mdotused ] ) . since the sun remains more massive for a longer period in this case than for the `` exponential '' mass loss cases ( compare fig .
[ fig : mass]b with fig .
[ fig : mass]a ) , there is more of an impact on the sun s internal structure ( compare fig . [
fig : helio]b with fig .
[ fig : helio]a )
. again , the mass - losing models with higher initial masses fit slightly ( but not significantly ) better than the standard non - mass - losing model , as shown by figure [ fig : helio]b and the rms values in table [ tab : results ] .
even our most extreme `` step - function '' case ( @xmath141 ) differs from the standard solar model by an rms@xmath151 , as compared to 0.0018 from other uncertainties . as in the `` exponential '' mass loss case
, none of the `` step - function '' mass loss cases can be ruled out by helioseismic observations .
figure [ fig : helio]c presents our results for the radical `` linear '' mass loss models ; recall that these models violate the solar wind constraints from the lunar rock observations by about an order of magnitude ( see [ ssec : mdotused ] ) . since the sun remains
remains more massive for a very long time compared to the other mass loss cases ( compare fig .
[ fig : mass]c with fig .
[ fig : mass]a , b ) , there is even more of an impact on the sun s internal structure ( compare fig . [
fig : helio]c with fig .
[ fig : helio]a , b ) .
intermediate @xmath56 values have the lowest rms@xmath145 values , though again none of the differences are significant ; but the most extreme `` linear '' @xmath141 differs from the standard solar model by rms@xmath152 , nearly as large as the largest effect of 0.0018 from other uncertainties .
even these radical `` linear '' mass loss cases can not yet be ruled out by comparisons with the helioseismic observations .
helioseismic observations measure the position of the base of the sun s convective envelope , namely @xmath153 @xcite , and the surface helium abundance , namely , a mass fraction @xmath154 ( see discussion in our companion paper our sun iv
the values of @xmath155 and @xmath156 for both our reference standard solar model and for all of our mass - losing models are shown in table [ tab : results ] .
the mass - losing cases all have values of @xmath155 and @xmath156 very close to those of the standard solar model , all of them being consistent with the helioseismic observations .
the lithium depletion in a main sequence star , relative to its initial lithium abundance , can result from three different causes .
( 1 ) there can be significant lithium depletion from pre - main - sequence lithium burning at early times , when the convective envelope reaches deep into the star .
( 2 ) rotationally induced mixing on the main sequence can transport lithium down from the convective envelope to regions hot enough for lithium burning .
( 3 ) mass loss on the main sequence can cause the convective envelope to move inwards and engulf lithium - depleted regions .
the sun s initial lithium abundance is assumed to be equal to the meteoritic abundance , and the depletion factor @xmath157 is obtained by comparing this initial abundance with the present observed solar photospheric lithium abundance .
this observed solar lithium depletion factor is @xmath158 @xcite . for pre - main - sequence lithium depletion ,
our reference standard solar model predicts a lithium depletion factor of @xmath159 , although this is very sensitive to the solar metallicity ( i.e. , @xmath23 value , as well as uncertainties in diffusion ) and to the choice of low - temperature molecular opacities ; values of @xmath157 from 11 to 70 can be obtained ( see our companion paper our sun iv
* ) . for rotation - induced main sequence lithium depletion ,
there is no theoretical prediction ; instead , the observed main sequence lithium depletion is used to constrain the free parameters in the theoretical treatment ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for main sequence mass loss
, the extent of main sequence lithium depletion depends primarily on the initial solar mass , and only weakly on the timescale of mass loss .
@xcite used the observed solar lithium depletion to obtain a limit on solar main sequence mass loss , finding that the maximum mass loss allowed was @xmath160 ( i.e. , a maximum initial solar mass of @xmath161 ) .
however , as discussed in
[ sec : intro ] , such an extreme mass loss case violates the constraint from the requirement that the early earth not lose its water via a moist greenhouse effect , which would occur for @xmath162 .
this constraint is based on a cloud - free climate model ; a very slight increase in @xmath56 might be allowed if clouds were taken into account . as the sun s initial mass
is increased above @xmath163 , there are two competing effects .
higher initial masses have less pre - main - sequence lithium depletion ; on the other hand , the higher the initial mass , the more mass loss has to take place , and thus the more lithium depletion takes place on the main sequence ( as the convective envelope sheds lithium - rich material from the surface and engulfs lithium - depleted material from below ) . as may be seen from table [ tab : results ] , for initial solar masses in the range @xmath164 , the first of these effects dominates , and the total lithium depletion is slightly less than in the standard ( non - mass - losing ) model ; slightly stronger rotation - induced mixing would be required in order to account for the observed lithium depletion . for masses
@xmath165 , the mass loss dominates ; however , even for our most extreme mass loss cases ( @xmath141 ) , the total lithium depletion ( from pre - main - sequence burning plus mass loss effects ) is only @xmath166 .
this is at most factor of 2 more lithium depletion than in the standard ( non - mass - losing ) model with @xmath159 .
the modest amount of mass loss considered here ( @xmath167 ) has only a minor effect on the extent of solar lithium depletion adding even the maximum possible mass loss only increases the lithium depletion by a factor of 2 , relative to the depletion on the pre - main - sequence .
this effect is smaller than the factor - of-2.5 effects on pre - main - sequence lithium depletion caused by uncertainties in other physical parameters , as discussed above and in our companion paper `` our sun iv '' @xcite . in these mass - losing models , rotational mixing
would still be required , to account for the majority of the sun s main - sequence lithium depletion ; therefore the arguments of @xcite ( who showed that mass loss could not be responsible for the majority of the observed hyades lithium depletion ) are not applicable , as discussed in
[ sssec : hyrest ] .
the observed solar lithium depletion thus can not be used to constrain these mass - losing solar models .
the observed solar beryllium abundance is @xmath168 , consistent with no depletion relative to the meteoritic value of @xmath169 .
the uncertainties given for these values imply that solar beryllium can not have been depleted by more than a factor of 2 ( 3-@xmath111 upper limit ) .
a standard solar model has negligible beryllium depletion ( @xmath34% ) ; our mass - losing solar models predict slightly larger depletions , but are still all consistent with the observational limit .
the most extreme of the `` exponential '' mass loss cases ( @xmath2 ) depleted beryllium by a relatively small amount ( a factor of 1.17 ) .
even the most extreme of the `` step - function '' and `` linear '' mass loss cases yielded only @xmath5-@xmath111 beryllium depletion factors ( of 1.53 and 1.63 , respectively ) ; the @xmath170 cases depleted beryllium by negligible amounts , less than 3% . in other words , all the mass losing cases deplete beryllium by significantly less than the observational upper limit of a factor of 2 .
as may be seen from table [ tab : results ] , the modest mass loss considered here has almost no effect on the predicted solar neutrino fluxes .
variations are at most a few percent in the predicted 8b flux and in the predicted capture rate for the chlorine experiment ( as compared to uncertainties of @xmath171% from other causes ) , and less than a percent in the predicted capture rate for the gallium experiment ( as compared to uncertainties of at least several percent from other causes ) the other sources of uncertainty in neutrino fluxes are discussed elsewhere ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
at present , an airless , rapidly - rotating body at earth s orbit would have a temperature of @xmath172c ( 255 k ) , if it had earth s present albedo and emissivity @xcite , but the earth s present mean surface temperature is observed to be @xmath173c ( 288 k ) . in other words ,
at present the greenhouse effect raises earth s surface temperature by @xmath174c .
the main greenhouse gases in the earth s atmosphere are co@xmath7 and h@xmath7o . if the atmospheric co@xmath7 abundance were _ constant _ ( at its present value ) , and the h@xmath7o abundance were determined by its equilibrium vapor pressure , then 2 gyr ago the earth s surface temperature would have been below @xmath76c @xcite . if the early earth s surface temperature were below the freezing point of water , extensive glaciation would be expected ; such glaciation would raise the earth s albedo , delaying the time when the surface temperature reached @xmath76c .
in other words , one would expect earth to be fully glaciated as recently as 1 gyr in the past @xcite .
on the other hand , a number of independent observations indicate that the earth was at least warm enough for liquid water to exist as far back as 4 gyr ago .
sedimentary rocks , which are laid down under water , have been dated to at least 4 gyr ago @xcite .
liquid water is necessary to explain the existence of the widespread microorganisms whose fossils are found in rocks dated as far back as 3.8 gyr ago @xcite . tidal or intertidal stromatolite fossils have been dated to @xmath175 gyr ago , alluvial detrital uraninite grains as far back as 3 gyr , and turbidites and ripple marks have been dated as far back as 3.5 gyr @xcite .
in fact , there is evidence not only that liquid water existed on the early earth , but also that earth was considerably warmer in the past than it is today . to start with
, there is no evidence of glaciation before 2.7 gyr ago @xcite , and it has been suggested that tillites prior to 2 gyr ago are actually due to impacts rather than glaciers @xcite .
archaeobacteria exhibit extreme thermophilic trends @xcite .
high ocean temperatures of @xmath176c in the period 2.6 to 3.5 gyr ago are suggested by sulphur isotope measurements @xcite .
average surface temperatures of tens of degrees celsius in the period 2.5 to 3.5 gyr ago are indicated by deuterium to o ratios @xcite .
temperatures as high as @xmath177c in the period @xmath178 gyr ago are suggested by differences in o isotopic data between coexisting cherts and phosphates @xcite , although the results are subject to interpretation .
the above `` weak sun paradox '' , of a faint young sun and a young earth warm enough for liquid water , has traditionally been explained by invoking a much stronger greenhouse effect for ancient earth .
this would be driven by extremely high ( i.e. , _ non - constant _ ) co@xmath7 concentrations in earth s early atmosphere , with partial co@xmath7 pressures between 0.2 and 10 bars 4.5 gyr ago @xcite .
qualitatively , high co@xmath7 concentrations can be justified on the basis of theoretical feedback mechanisms linking mineral dissolution to liquid water and thus to atmospheric co@xmath7 @xcite .
although such massive amounts of co@xmath7 in the earth s early atmosphere are a possible solution to the `` weak sun paradox '' , they are not mandated ; there is little experimental evidence available on which to base a choice of co@xmath7 concentration @xcite . indeed , very high concentrations may prove to be inconsistent with derived weathering rates @xcite .
the recent work of @xcite places an upper limit of 0.04 bar on the partial pressure of co@xmath7 in the period from 2.75 to 2.2 gyr ago , based on the absence of siderite in paleosols ; earlier work by @xcite estimated a partial pressure of co@xmath7 of 0.004 bar , but with the same upper limit of 0.04 bar . for earth surface temperatures between 5 and 20@xmath179c during that period ,
climate models predict a partial pressure of co@xmath7 between 0.03 and 0.3 bar @xcite , barely consistent with the upper limit of 0.04 bar . in other words , there is little evidence for a strong co@xmath7 greenhouse effect on ancient earth .
actual measurements of co@xmath7 abundances are available only for relatively recent times , i.e. , only for the last @xmath180 gyr ( see , e.g. , * ? ? ?
* ; * ? ? ?
these latter measurements show major variations in the co@xmath7 abundance over the past 0.45 gyr .
the lowest values are comparable to the present - day co@xmath7 abundance of about 350 ppmv , namely , 0.00035 bar ( or the pre - industrial - age value of @xmath181 ppmv note that `` ppmv '' refers to parts per million by volume ) ; the highest values measured over the last 0.45 gyr are @xmath182 ppmv . however , from these measurements , there is no clear evidence of a long - term trend of higher co@xmath7 abundances in the relatively recent past ( i.e. , the last 0.4 gyr ) .
a non - co@xmath7 greenhouse has been suggested for the early earth @xcite .
recently , @xcite calculated that a strong greenhouse contribution from ammonia was possible , if a concentration of @xmath183 \sim 10^{-5}$ ] could be maintained .
normally , the ammonia would be photodissociated by solar uv flux on a timescale of 10 years .
they pointed out that ammonia could be shielded from the uv radiation by high - altitude organic solids produced from photolysis of methane a photochemical smog , similar to that observed in the upper atmosphere of titan @xcite .
however , the ammonia lifetime depends sensitively on two parameters , the fraction @xmath184 of the methane irradiation products that are organic solids , and the sedimentation timescale @xmath74 of the smog ; @xcite take as reasonable values @xmath185 and 0.5 yr@xmath186 yr . for @xmath187 , or @xmath188 yr , the ammonia lifetime is long enough for the required ammonia concentration to be maintained , given a reasonable amount of resupply .
however , for @xmath189 and @xmath190 yr , the ammonia lifetime is less than 200 yr , which would require excessively large amounts of ammonia production to maintain the required ammonia greenhouse effect .
@xcite also note that an atmosphere rich in n@xmath7 , with minor co@xmath7 and ch@xmath191 components , could have adequate self - shielding of nh@xmath192 from photodissociation ( allowing an ammonia greenhouse ) , only as long as the ratio ch@xmath191/co@xmath193 was maintained .
in other words , for the early earth , a smog - shielded ammonia greenhouse is a viable solution to the `` weak sun paradox '' under certain conditions , but fails under others .
as discussed above , it is not clear whether the greenhouse effect could suffice to warm the early earth .
a bright young sun , with stronger illumination of the young earth than from the standard solar model , would require a less extreme early greenhouse effect to prevent the early earth from freezing over . for mars , there are also indications of higher surface temperatures in the past , that are even harder to explain by a greenhouse effect .
there is evidence of large scale flow of liquid water @xmath178 gyr ago , from the drainage channels and valley networks visible on the heavily cratered ancient surface of mars @xcite ; there is some evidence for lakes @xcite , and possibly even oceans 3 to 4 gyr ago @xcite .
even if the channels were formed by subsurface sapping of groundwater , martian surface temperatures significantly higher than today would have been required for liquid water to be present near the surface ; if the suggested evidence of glacial markings were confirmed , this would require temperatures high enough for precipitation to occur @xcite .
@xcite demonstrated that there is an upper limit to the greenhouse warming of mars that is possible from co@xmath7 .
he showed that the maximum possible greenhouse warming occurs at a martian surface co@xmath7 pressure of 5 bars with more co@xmath7 , the added greenhouse warming is outweighed by the cooling effects of increased co@xmath7 condensation in the upper martian atmosphere .
he demonstrated that the requirement of liquid water on mars , i.e. , a surface temperature of at least @xmath194k , demands a solar flux value @xmath195 ( where @xmath196 is the solar flux relative to its present value ) , even with the most favorable greenhouse warming case of a co@xmath7 pressure of 5 bars . at 3.8 gyr ago , when liquid water is thought to have existed on mars , the standard solar model yields a value of @xmath197 ( see fig . [
fig : solarflux ] ) , totally insufficient relative to kasting s minimum value of 0.86 .
@xcite states that the uncertainties in his martian climate model might push the limiting value of @xmath196 from 0.86 down to 0.80 ( albeit for an unreasonably low martian albedo ) , but even this lower @xmath196 requirement is incompatible with the standard solar model .
the standard solar model does not reach this extreme limit of @xmath198 until 2.9 gyr ago , and reaches @xmath199 later still , at 1.9 gyr ago ( see fig .
[ fig : solarflux ] ) in either case , far too late to account for liquid water on mars 3.8 gyr ago . with co@xmath7 pressures either lower or higher than 5 bar
, @xcite shows that even higher solar flux values would be required to yield liquid water .
he presents the solar flux required to obtain liquid water as a function of the co@xmath7 pressure : e.g. , a pressure of 3 bar would require @xmath200 , and a pressure of 10 bar would require @xmath201 .
the standard solar model is totally incapable of yielding such high fluxes 3.8 gyr ago ( see fig . [
fig : solarflux ] ) .
in other words , for a standard solar model , co@xmath7 greenhouse warming can not under any circumstances yield liquid water on early mars .
a preliminary study by @xcite suggested the possibility that relatively small amounts of atmospheric so@xmath7 ( @xmath202bar ) might have served as a powerful source of heating in the upper atmosphere of early mars ( due to its strong absorption in the near uv ) , which might have been sufficient to prevent the condensation of co@xmath7 ( though they noted that so@xmath7 in the presence of liquid water would produce h@xmath7so@xmath191 , which would lead to some countervailing cooling ) .
@xcite pointed out a more important effect of the h@xmath7so@xmath191 : even @xmath203bar of so@xmath7 in the martian atmosphere , 4 orders of magnitude less than proposed by @xcite , would result in a ph acidic enough to attack solid rocks and precipitate gypsum ( caso@xmath191 * 2h@xmath7o ) , removing the so@xmath7 from the atmosphere . some temperature increase on early mars
is expected from geothermal heating @xcite , but by itself it is insufficient @xcite .
with the above greenhouse and geothermal heating apparently incapable of yielding liquid water on early mars , given the illumination from a standard solar model , let us consider the possibility of a non - standard , brighter young sun .
figure [ fig : solarflux ] presents the relative flux values @xmath196 throughout the sun s past history yielded by our mass - losing solar models with initial masses from @xmath0 to @xmath1 ( as well as that from a standard solar model ) .
the requirement that the early earth not lose its water via a moist greenhouse effect leads to an upper limit of @xmath204 , corresponding to the upper edge of the figures .
the lower luminosity constraints from the requirement that liquid water be present 3.8 gyr ago on early mars are shown by the vertical arrows ; the heavy double arrow corresponds to the limit @xmath195 obtained by @xcite , and the lighter single arrow to his extreme ( less probable ) limit @xmath205 . figure [ fig : solarflux]a presents the solar flux @xmath196 at earth and mars ( relative to their present flux ) from the time of solar system formation until the present , for our `` exponential '' mass loss cases ; the flux for the standard solar model ( without mass loss ) is also shown , by the solid line .
note that the mass loss time scale is between 0.755 and 0.551 gyr , with initial mass loss rates from @xmath206 to @xmath94yr@xmath30 , for initial masses from 1.01 to @xmath1 , respectively .
the exponential decline as a function of time of these mass loss rates means that they are generally consistent with the observations of @xmath40 uma , @xmath41 cet , and @xmath42 com presented by @xcite , who obtained upper limits of @xmath207 , @xmath208 , and @xmath209yr@xmath30 , respectively , for the mass loss rates of these three young sun - like stars ( note that the @xmath141 case is only marginally consistent , lying very slightly above the @xmath40 uma limit : see fig .
[ fig : mdot ] ) .
our results demonstrate that , for the `` exponential '' mass loss cases , the @xmath2 case and only this case is marginally consistent with the @xcite mars flux requirement @xmath210 at an age of @xmath178 gyr ago ; if the martian surface co@xmath7 pressure 3.8 gyr ago was either much lower or much higher than 5 bar , even the @xmath2 case would be ruled out .
( if the unlikely extreme kasting flux limit of @xmath211 is used , initial masses @xmath212 would be permissible for a co@xmath7 pressure of 5 bars , and the @xmath2 case would be marginally compatible with pressures between @xmath4 and @xmath213 bar . )
since the semi - empirical mass loss law @xmath51 recently presented by @xcite is strongly peaked at early times ( as discussed in [ sssec : mdotobs ] ) , it tends to yield a relatively small solar mass @xmath178 gyr ago .
a @xmath150 mass loss relation , normalized by the present solar mass loss rate , would yield @xmath214gyr@xmath215 ; by tweaking parameters to their extreme limits , one might obtain @xmath214gyr@xmath216 . comparing the flux relations in figure [ fig : solarflux ] with the solar mass as a function of time in figure [ fig : helio ] , one can estimate that a solar mass of @xmath214gyr@xmath217 is needed in order to satisfy the @xcite mars flux requirement @xmath218gyr@xmath219 .
thus the mass loss formula of @xcite might possibly be capable of satisfying the mars flux requirement .
this possibility will be tested in a future paper .
figure [ fig : solarflux]b presents the extreme `` step - function '' mass loss case .
this case has a constant mass loss rate for the first 1.6 gyr ( @xmath96 to @xmath97yr@xmath30 , for initial solar masses of 1.01 to @xmath88 , respectively ) , and a low mass loss rate thereafter .
these mass loss rates are all consistent with the stellar mass loss observations of @xcite for @xmath40 uma , @xmath41 cet , and @xmath42 com quoted above .
these `` step - function '' cases have a longer mass loss timescale than the `` exponential '' one , and thus yield a higher solar flux for the first 1.6 gyr .
( due to the way the `` step - function '' cases were defined , after the first 1.6 gyr their solar flux is very close to that of the standard solar model . )
our results demonstrate that , for this extreme `` step - function '' mass loss case , initial masses @xmath220 are capable of yielding liquid water on mars until 3.8 gyr ago
. figure [ fig : solarflux]c similarly presents the radical `` linear '' mass loss case .
this case has a high initial mass loss rate ( @xmath221 to @xmath99yr@xmath30 , for initial solar masses of 1.01 to @xmath88 , respectively ) , which remains relatively high throughout much of the sun s lifetime ( since it declines linearly with time to reach the present solar mass loss rate at the present time ) .
these mass loss cases are consistent with the stellar mass loss observations of @xcite for @xmath40 uma , @xmath41 cet , and @xmath42 com quoted above . however , they are _ not _ consistent with the lunar rock observations of the solar wind over the past 3 gyr , violating this latter constraint by an order of magnitude .
our results demonstrate that , for this radical `` linear '' mass loss case , initial masses @xmath212 are capable of yielding liquid water on mars until 3.8 gyr ago .
the lower end of this range is mildly ( but not significantly ) favored by the helioseismology ; note that such cases with @xmath222 have remarkably constant solar flux over the first 3 gyr .
a mass - losing solar model will always be brighter at birth than the standard solar model , since the luminosity @xmath223 at the zero age main sequence ( zams ) is roughly proportional to the mass to the fourth power ( @xmath224 ) . for a mass - losing sun
, the orbital radii of the planets varies inversely with the solar mass ( @xmath225 , due to conservation of angular momentum ) ; the initial flux at the planets is thus proportional to the sixth power of the initial solar mass ( @xmath226 ) .
figure [ fig : solarflux ] illustrates the solar flux at the planets as a function of time , demonstrating how much higher the early solar flux at the planets is in the mass - losing cases than in the standard ( non - mass - losing ) model . for clarity , only the flux for our `` preferred '' initial masses for each type of mass loss
are illustrated in figure [ fig : prefflux ] .
our preferred `` exponential '' case ( with @xmath141 ) predicts a solar flux at the planets about 5% higher at birth than at present , considerably higher than that indicated by the standard solar model ( which predicts a flux 29% lower than at present ) . at 3.8 gyr
ago , the flux for our `` exponential '' case would have been only 16% lower than at present ( cf .
25% for the standard model ) . for our preferred `` step - function '' case ( with @xmath227 ) , the flux at the planets would have been only 10% lower at birth than at present ( cf .
29% for the standard model ) ; at 3.8 gyr ago , the flux would have been only 14% lower than at present ( cf .
25% for the standard model ) .
for these `` exponential '' and `` step - function '' cases , the flux at the planets for the past 3 billion years would be essentially the same as that predicted by the standard solar model . our radical `` linear '' case ( with @xmath227 )
would have had an almost constant solar flux at the planets for the first 3 gyr , namely , only 11% lower than at present ( cf .
29% to 12% lower for the standard model ) ; for this case , the flux would be close to that predicted by the standard solar model only during the last billion years .
figure [ fig : hrd ] presents the evolution in the hr diagram of our preferred exponential and step - function mass - losing cases ( heavy dashed and dot - dashed curves , respectively ) ; these cases are in agreement with the helioseismic observations , with the existence of water on early mars , and with the lunar rock observations of solar mass loss .
( a radical linear case , agreeing with the first two of these constraints but disagreeing with the third one , is shown by the light dotted curve . ) for comparison , the standard solar model is also displayed ( solid line ) .
figure [ fig : hrd ] illustrates that the early evolution of mass - losing solar models is in the opposite direction in the hr diagram to the standard solar model : the mass - losing models initially become less luminous and slightly redder ( instead of more luminous and slightly bluer ) . figure [
fig : hrd ] also illustrates that in the past , the sun s surface temperature changed only by negligible amounts ( 1 or 2% ) , both for the standard and the mass losing cases , in contrast to the relatively large changes in the luminosity .
a slightly higher initial solar mass , producing a brighter young sun , turns out to be a viable explanation for warm temperatures on early earth and mars , that otherwise are difficult to account for ( particularly for mars ) . such a higher initial solar mass leaves a fingerprint on the sun s present internal structure that is large enough to be detectable in principle via helioseismic observations .
our computations demonstrated that all 21 of the mass - losing solar models that we considered were consistent with the helioseismic observations ; in fact , our preferred mass - losing cases were in marginally ( though not significantly ) better agreement with the helioseismology than the standard solar model was . however ,
there are still significant uncertainties in the observed solar composition and in the input physics on which the solar models are based ; these uncertainties have a slightly larger effect on the sun s present internal structure than the fingerprint left from early solar mass loss .
future improvements by a factor of 2 or so in the accuracy of these input parameters could reduce the size of the uncertainties below the level of the fingerprints left by a more massive , brighter young sun , allowing one to determine whether early solar mass loss took place or not . also urgently needed are more measurements of mass loss rates from other young stars similar to the young sun , and more measurements from our solar system that can be used to estimate the solar wind in the past .
we are indebted to prof .
marc h. pinsonneault for helpful discussions on diffusion and for providing us with his diffusion code .
we are grateful to prof .
sarbani basu for discussions of helioseismology , and for providing us with the current helioseismic reference model ; we are also grateful to prof .
dimitri m. mihalas , for providing us with his equation - of - state code .
we wish to thank prof .
charles a. barnes , prof .
yuk l. yung , and dr .
mimi f. gerstell for thoughtful discussions and encouragement .
we wish to acknowledge the support provided by prof .
thomas a. tombrello , chairman of the division of physics , math , and astronomy , and prof .
robert d. mckeown , head of the w. k. kellogg radiation laboratory .
one of us ( i .- j .
s. ) wishes to thank alexandra r. christy , her daughter , and prof .
robert f. christy , her husband , for their supportiveness , and robert f. christy for critical analysis and helpful comments .
one of us ( a. i. b. ) wishes to thank prof .
peter g. martin and prof .
j. richard bond for their support , and m. elaine boothroyd , his wife , for her patience and encouragement .
this work was supported by a grant from the sun - earth connection program of the supporting research and technology and suborbital program in solar physics of the national aeronautics and space administration , and by the national science foundation grant to the kellogg radiation laboratory .
bahcall , j. n. , chen , x. l. , & kamionkowski , m. 1998 , , 57 , 2756 bahcall , j. n. , pinsonneault , m. h. , & basu , s. 2001 , , 555 , 990 bahcall , j. n. , pinsonneault , m. h. , & wasserburg , g. j. 1995 , rev .
phys . , 67 , 781 bahcall , j. n. , & ulrich , r. k. 1988 , rev .
phys . , 60 , 297 basu , s. , pinsonneault , m. h. , & bahcall , j. n. 2000 , , 529 , 1084 boothroyd , a. i. , & sackmann , i .- j . 2002 , , in press ( our sun v ) boothroyd , a. i. , sackmann , , & fowler , w. a. 1991 , , 377 , 318 bowring , s. a. , williams , i. s. , & compston , w. 1989 , geology , 17 , 971 brown , a. , veal , a. , judge , p. , bookbinder , j. a. , & hubeny , i. 1990 , , 361 , 320 charbonnel , c. , vauclair , s. , & zahn , 1992 , , 255 , 191 christensen - dalsgaard , j. , & dppen , w. 1992 , , 4 , 267 cogley , j. g. , & henderson - sellers , a. 1984 , rev . geophys .
space phys .
, 22 , 131 cohen , e. r. , & taylor , b. n. 1986 , codata bulletin no .
63 ( new york : pergamon ) gaidos , e. j. , gdel , m. , & blake , g. a. 2000 , geophys .
lett . , 27 , 501 geiss , j. 1973 , in proc . 13th intl
. cosmic ray conf .
, vol . 5 ( denver : univ . of denver ) , 3375 geiss , j. , & bochsler , p. 1991 , in the sun in time , ed . c. sonnett , m. giampapa , & m. matthews ( tucson : univ .
arizona press ) , 98 goldspiel , j. , & squyres , s. w. 1991 , icarus , 89 , 392 grevesse , n. , & noels , a. 1993 , in origin and evolution of the elements , ed .
n. prantzos , e. vangioni - flam , & m. cass ( cambridge : cambridge univ .
press ) , 15 grevesse , n. , & sauval , a. j. 1998 , space sci .
, 85 , 161 gruzinov , a. v. , & bahcall , j. n. 1997 , , 490 , 437 gruzinov , a. v. , & bahcall , j. n. 1998 , , 504 , 996 guenther , d. b. 1989 , , 339 , 1156 guenther , d. b. , demarque , p. , kim , y .- c . , & pinsonneault , m. h. 1992 , , 387 , 372 holland , h. d. , & zbinden , e. a. 1986 , paleosols and the evolution of the atmosphere , part i. , in physical and chemical weathering in geochemical cycles , ed .
a. lerman & m. meybeck ( nato asi institute ) kerridge , j. f. , signer , p. , wieler , r. , becker , r. h. , & pepin , r. o. 1991 , in the sun in time , ed .
c. sonnett , m. giampapa , & m. matthews ( tucson : univ .
arizona press ) , 389 knauth , l. p. , & epstein , s. 1976 , geochim .
cosmochim .
acta , 40 , 1095 parker , t. j. , gorsline , d. s. , saunders , r. s. , pieri , d. c. , & schneeberger , d. m. 1993 , j. geophys .
res . , 95 , 11061 pinsonneault , m. h. , kawaler , s. d. , sofia , s. , & demarque , p. 1989
, , 338 , 424 pollack , j. b. 1979 , icarus , 37 , 479 swenson , f. j. , & faulkner , j. 1992 , , 395 , 654 swenson , f. j. , faulkner , j. , iglesias , c. a. , rogers , f. j. , & alexander , d. r. 1994 , , 422 , l79 thoul , a. a. , bahcall , j. n. , & loeb , a. 1994 , , 421 , 828 wood , b. e. , & linsky , j. l. 1998 , , 492 , 788 wood , b. e. , linsky , j. l. , mller , , & zank , g. p. 2001
, , 547 , l49 wood , b. e. , mller , , zank , g. p. , & linsky , j. l. 2002 , , 574 , 412 yung , y. l. , nair , h. , & gerstell , m. f. 1997 , icarus , 130 , 222 \1 .
fine - zoned reference & 1.817 & .02030 & .2760 & .2424 & .7135 & .00133 & .00085 & .01698 & & & & 24.24 & 1.008 & 7.87 & 133.7 & 5.31 + 2 .
fine - zoned @xmath228 exp & 1.818 & .02025 & .2754 & .2425 & .7133 & .00116 & .00074 & .01533 & 1 & .00028 & .00179 & 19.22 & 1.009 & 7.87 & 133.7 & 5.32 + 3 .
fine - zoned @xmath229 exp & 1.819 & .02017 & .2748 & .2425 & .7133 & .00126 & .00074 & .01544 & 1 & .00017 & .00158 & 17.43 & 1.009 & 7.87 & 133.7 & 5.32 + 4 .
fine - zoned @xmath230 exp & 1.819 & .02012 & .2743 & .2426 & .7133 & .00118 & .00070 & .01437 & 1 & .00034 & .00276 & 17.81 & 1.009 & 7.88 & 133.7 & 5.32 + 5 .
fine - zoned @xmath231 exp & 1.819 & .02007 & .2738 & .2427 & .7132 & .00118 & .00068 & .01398 & 1 & .00034 & .00311 & 19.91 & 1.012 & 7.89 & 133.8 & 5.33 + 6 .
fine - zoned @xmath232 exp & 1.821 & .02003 & .2734 & .2428 & .7131 & .00112 & .00065 & .01340 & 1 & .00039 & .00366 & 22.96 & 1.026 & 7.91 & 133.9 & 5.35 + 7 .
fine - zoned @xmath233 exp & 1.821 & .01999 & .2730 & .2428 & .7130 & .00105 & .00063 & .01247 & 1 & .00049 & .00464 & 26.99 & 1.069 & 7.92 & 133.9 & 5.35 + 8 .
fine - zoned @xmath1 exp & 1.823 & .01997 & .2726 & .2429 & .7128 & .00096 & .00062 & .01162 & 1 & .00060 & .00550 & 31.82 & 1.173 & 7.94 & 134.1 & 5.37 + 9 . coarse - zoned reference & 1.814 & .02030 & .2760 & .2424 & .7136 & .00140 & .00091 & .01772 & 1 & .00008 & .00074 & 24.36 & 1.010 & 7.89 & 133.8 & 5.33 + 10 .
@xmath228 exp & 1.816 & .02024 & .2754 & .2427 & .7134 & .00128 & .00081 & .01641 & 9 & .00016 & .00134 & 19.22 & 1.009 & 7.91 & 134.0 & 5.35 + 11 .
@xmath229 exp & 1.816 & .02016 & .2748 & .2429 & .7134 & .00124 & .00077 & .01569 & 9 & .00024 & .00208 & 17.26 & 1.009 & 7.89 & 133.8 & 5.33 + 12 .
@xmath230 exp & 1.816 & .02010 & .2742 & .2431 & .7134 & .00120 & .00074 & .01515 & 9 & .00030 & .00263 & 17.82 & 1.009 & 7.89 & 133.8 & 5.33 + 13 .
@xmath231 exp & 1.816 & .02005 & .2738 & .2433 & .7133 & .00115 & .00070 & .01448 & 9 & .00037 & .00332 & 19.72 & 1.012 & 7.90 & 133.8 & 5.34 + 14 .
@xmath232 exp & 1.817 & .02001 & .2734 & .2433 & .7133 & .00111 & .00067 & .01379 & 9 & .00045 & .00402 & 22.82 & 1.026 & 7.91 & 133.9 & 5.35 + 15 .
@xmath233 exp & 1.818 & .01997 & .2730 & .2434 & .7131 & .00105 & .00065 & .01300 & 9 & .00053 & .00484 & 27.47 & 1.068 & 7.93 & 134.0 & 5.36 + 16 .
@xmath1 exp & 1.819 & .01995 & .2726 & .2434 & .7131 & .00101 & .00063 & .01217 & 9 & .00062 & .00569 & 34.25 & 1.170 & 7.95 & 134.1 & 5.38 + 17 .
@xmath228 step & 1.816 & .02019 & .2751 & .2426 & .7134 & .00128 & .00080 & .01626 & 9 & .00018 & .00149 & 19.57 & 1.009 & 7.87 & 133.7 & 5.32 + 18 .
@xmath229 step & 1.815 & .02007 & .2742 & .2432 & .7135 & .00124 & .00077 & .01572 & 9 & .00027 & .00204 & 18.94 & 1.009 & 7.86 & 133.6 & 5.31 + 19 . @xmath230 step & 1.815 & .01999 & .2734 & .2435 & .7135 & .00117 & .00072 & .01463 & 9 & .00035 & .00316 & 20.87 & 1.011 & 7.88 & 133.7 & 5.32 + 20 .
@xmath231 step & 1.816 & .01993 & .2728 & .2437 & .7133 & .00109 & .00067 & .01347 & 9 & .00048 & .00435 & 24.36 & 1.025 & 7.90 & 133.8 & 5.34 + 21 .
@xmath232 step & 1.818 & .01988 & .2722 & .2437 & .7131 & .00101 & .00064 & .01207 & 9 & .00063 & .00580 & 29.75 & 1.074 & 7.94 & 134.0 & 5.37 + 22 . @xmath233 step &
1.821 & .01985 & .2716 & .2437 & .7129 & .00095 & .00064 & .01061 & 9 & .00079 & .00733 & 37.76 & 1.213 & 7.98 & 134.3 & 5.40 + 23 .
@xmath1 step & 1.823 & .01982 & .2711 & .2436 & .7126 & .00089 & .00068 & .00916 & 9 & .00095 & .00883 & 49.63 & 1.534 & 8.01 & 134.4 & 5.43 + 24 .
@xmath228 linear & 1.818 & .02021 & .2749 & .2427 & .7132 & .00116 & .00070 & .01486 & 9 & .00035 & .00293 & 19.70 & 1.009 & 7.94 & 134.1 & 5.37 + 25 .
@xmath229 linear & 1.818 & .01997 & .2731 & .2429 & .7131 & .00122 & .00069 & .01427 & 9 & .00046 & .00356 & 18.49 & 1.009 & 7.92 & 134.0 & 5.36 + 26 .
@xmath230 linear & 1.816 & .01972 & .2713 & .2439 & .7133 & .00115 & .00069 & .01293 & 9 & .00060 & .00496 & 20.11 & 1.011 & 7.87 & 133.6 & 5.32 + 27 .
@xmath231 linear & 1.817 & .01959 & .2700 & .2445 & .7133 & .00105 & .00070 & .01084 & 9 & .00082 & .00713 & 23.34 & 1.027 & 7.90 & 133.8 & 5.35 + 28 .
@xmath232 linear & 1.817 & .01947 & .2687 & .2450 & .7132 & .00099 & .00078 & .00858 & 9 & .00106 & .00958 & 28.35 & 1.081 & 7.95 & 134.1 & 5.38 + 29 .
@xmath233 linear & 1.821 & .01937 & .2675 & .2452 & .7128 & .00099 & .00094 & .00640 & 9 & .00132 & .01200 & 36.71 & 1.250 & 8.00 & 134.4 & 5.43 + 30 .
@xmath1 linear & 1.824 & .01929 & .2663 & .2452 & .7125 & .00108 & .00114 & .00432 & 9 & .00161 & .01476 & 48.90 & 1.631 & 8.07 & 134.7 & 5.48 + 31 .
gs98 reference & 1.775 & .01912 & .2759 & .2418 & .7155 & .00188 & .00120 & .02121 & 9 & .00060 & .00378 & 18.37 & 1.009 & 7.73 & 132.6 & 5.24 + 32 .
gs98 @xmath228 exp & 1.776 & .01906 & .2752 & .2421 & .7154 & .00178 & .00111 & .02014 & 31 & .00035 & .00123 & 14.85 & 1.008 & 7.78 & 132.9 & 5.28 + 33 .
gs98 @xmath229 exp & 1.776 & .01899 & .2746 & .2422 & .7154 & .00175 & .00105 & .01929 & 31 & .00035 & .00201 & 13.70 & 1.008 & 7.72 & 132.5 & 5.23 + 34 .
gs98 @xmath230 exp & 1.776 & .01893 & .2741 & .2425 & .7153 & .00168 & .00101 & .01864 & 31 & .00044 & .00267 & 14.47 & 1.008 & 7.73 & 132.6 & 5.24 + 35 .
gs98 @xmath231 exp & 1.777 & .01889 & .2736 & .2427 & .7153 & .00162 & .00097 & .01799 & 31 & .00049 & .00332 & 16.46 & 1.010 & 7.74 & 132.6 & 5.25 + 36 .
gs98 @xmath232 exp & 1.778 & .01885 & .2732 & .2428 & .7152 & .00157 & .00092 & .01725 & 31 & .00056 & .00407 & 19.58 & 1.025 & 7.75 & 132.7 & 5.26 + 37 .
gs98 @xmath233 exp & 1.778 & .01881 & .2728 & .2428 & .7151 & .00152 & .00088 & .01652 & 31 & .00063 & .00481 & 24.15 & 1.062 & 7.77 & 132.7 & 5.27 + 38 .
gs98 @xmath1 exp & 1.780 & .01879 & .2724 & .2428 & .7150 & .00145 & .00084 & .01573 & 31 & .00070 & .00561 & 30.84 & 1.160 & 7.78 & 132.8 & 5.28 + worst variant model & & & & & & .00366 & .00267 & .03376 & & .00184 & .02614 & & & & & + | the relatively warm temperatures required on early earth and mars have been difficult to account for via warming from greenhouse gases .
we tested whether this problem can be resolved for both earth and mars by a young sun that is brighter than predicted by the standard solar model ( ssm ) .
we computed high - precision solar evolutionary models with slightly increased initial masses of @xmath0 to @xmath1 ; for each mass , we considered three different mass loss scenarios .
we then tested whether these models were consistent with the current high - precision helioseismic observations .
the relatively modest mass loss rates in these models are consistent with observational limits from young stars and estimates of the past solar wind obtained from lunar rocks , and do not significantly affect the solar lithium depletion . for appropriate initial masses ,
all three mass loss scenarios are capable of yielding a solar flux 3.8 gyr ago high enough to be consistent with water on ancient mars .
the higher flux at the planets is due partly to the fact that a more massive young sun would be intrinsically more luminous , and partly to the fact that the planets would be closer to the more massive young sun . at birth on the main sequence , our preferred initial mass
@xmath2 would produce a solar flux at the planets 50% higher than that of the ssm , namely , a flux 5% _ higher _ than the present value ( rather than 30% lower , which the ssm predicts ) . at first
( for @xmath3 gyr ) , the solar flux would _ decrease _ ; subsequently , it would behave more like the flux in the ssm , increasing until the present .
we find that all of our mass - losing solar models are consistent with the helioseismic observations ; in fact , our preferred mass - losing case with @xmath2 is in marginally ( though insignificantly ) better agreement with the helioseismology than is the ssm .
the early solar mass loss of a few percent does indeed leave a small fingerprint on the sun s internal structure
. however , for helioseismology to significantly constrain early solar mass loss would require higher accuracy in the observed solar parameters and input physics , namely , by a factor of @xmath4 for the observed solar surface composition , and a factor of @xmath5 for the solar interior opacities , the @xmath6 nuclear reaction rate , and the diffusion constants for gravitational settling . |
[ cols="<,^,^,^,^ " , ] clumpy torus models ( e.g. , * ? ? ? * ; * ? ? ? * ) and in particular the @xcite models ( also known as clumpy models ) have been shown to fit well the near and mid - ir seds and spectroscopy of seyfert galaxies @xcite .
the clumpy torus models are described by six parameters to account for the torus geometry and the properties of the dusty clouds ( see table [ tab : torusmodels ] ) .
the geometry of the torus is defined with the torus radial thickness @xmath10 , which is the ratio between the outer and inner radii of the torus @xmath11k , and the agn bolometric luminosity @xmath12 .
then the dust sublimation radius in pc is @xmath13 , where the agn bolometric luminosity is in units of @xmath14 . ] , and the width of the angular distribution of the clouds @xmath15 , that is , the angular size of the torus as measured from its equator .
the clouds have an optical depth @xmath16 and are arranged with a radial distribution expressed as a declining power law with index @xmath17 ( @xmath18 ) , with @xmath19 giving the mean number of clouds along a radial equatorial ray .
the last parameter is the viewing angle to the torus @xmath20 .
we refer the reader to the sketch of the clumpy model geometry in figure 1 of @xcite .
we updated the fit of @xcite with two main differences .
first , we used the higher angular resolution mid - ir unresolved flux densities inferred in section [ sec : trecs ] and also included the t - recs mid - ir spectrum for the fit .
second , this galaxy is classified as a seyfert 1.5 ( it shows a broad component in h@xmath21 , * ? ? ? * ) and therefore there is an unobscured view to the agn . to account for it , we added an agn component to the resulting torus model ( see below ) and allowed for a small amount of foreground extinction ( @xmath22mag , see * ? ? ? * ) , to redden the fitted models .
we also included in the nuclear sed the hst / nicmos @xmath23 m unresolved flux @xcite . similarly to our previous work @xcite , we took a bayesian approach to both fit the torus models to the data and derive meaningful confidence levels for the fitted parameters . to this end
, we used the bayesclumpy ( bc ) fitting routine , an interpolated version of the clumpy models @xcite .
the new version of bc allows to fit both photometric points and spectra . for the modelling we used the whole range of parameters probed by the clumpy models .
we list them in table [ tab : torusmodels ] . unlike in some of our previous work , here we allowed for the torus extent to vary in the range @xmath24 , instead of restricting the models to small tori .
this is because larger tori may be required to reproduce the far - ir emission of agn @xcite .
in addition to the six torus model parameters to be fitted plus the foreground extinction , there is an extra parameter to account for the vertical displacement needed to match the fluxes of a given model to the observations .
this vertical shift , which we allow to vary freely , scales with the agn bolometric luminosity ( see * ? ? ?
we assumed uniform distributions in the ranges given in table [ tab : torusmodels ] for the prior distributions of the torus model parameters .
the bc fitting routine translates the probability distributions of the fitted torus model parameters into two model spectra .
the first corresponds to the maximum - a - posteriori ( map ) values that represent the best fit to the data , while the second is the model produced with the median value of the probability distribution of each parameter .
[ fig : torusmodel ] shows the best fit to the nuclear near and mid - ir sed and the mid - ir spectrum of the ngc 1365 agn .
we also plot in this figure the range of acceptable torus models at the @xmath25 confidence level .
as can be seen from this figure , the fit to the photometric and spectroscopic data is very good .
specifically , the clumpy models reproduce well the flat silicate feature at around @xmath26 m ( see * ? ? ?
* their figures 16 and 17 ) .
we can now compare the inferred torus model parameters with observations .
we start with the agn bolometric luminosity .
as explained above , we chose to keep the scaling of the torus models to the observations as a free parameter . because the scaling is proportional to the agn bolometric luminosity , when compared with the observations it provides information about the goodness of the fit ( see * ? ? ? * for more details ) .
the derived agn bolometric luminosity from our fit is @xmath27 .
@xcite infered a @xmath28kev luminosity in the range @xmath29 ( corrected for the assumed distance in this work ) .
our derived agn bolometric luminosity is consistent with the values from x - rays after applying a bolometric correction .
we can also compare the torus radius @xmath30pc and torus angular width @xmath31 from the modelling with observations .
the width of the angular distribution of the clouds in the torus is related to the opening angle of the ionization cones , with some dependence with the geometry of the gas ( see discussion in * ? ? ?
the half opening angle of the ngc 1365 ionization cone modeled by @xcite is @xmath32deg ( measured from the pole ) , which is compatible with the derived angular size of the torus .
the size ( diameter ) of the @xmath33 m emitting source inferred from the modelling of the mid - ir interferometric observations of this galaxy is @xmath34pc @xcite .
we note that the size of the @xmath33 m emitting source , which traces the warm dust within the torus , is expected to be smaller than the _ true _ size of the torus .
this is because the @xmath33 m emission is rather insensitive to cooler material further from the nucleus ( see discussion in * ? ? ?
given the good agreement with the observations , we conclude that the fitted torus model provides a good representation of the agn torus emission . to estimate the agn far - ir flux density in the pacs bands , we extrapolated the fitted torus model beyond @xmath35 m .
as can be seen from fig .
[ fig : torusmodel ] , the torus model flux density peaks at @xmath36 m and falls off rapidly at longer wavelengths .
this is similar to results for other agn ( see e.g. , * ? ? ?
* ) . in table
[ tab : agnfluxes ] we list the extrapolated agn far - ir flux densities together with those in the mid - ir from the t - recs imaging and spectroscopy .
the ranges given in this table for the far - ir fluxes from the bc fit of the ngc 1365 take into account the @xmath37 confidence region of the fitted clumpy torus models .
a quick comparison between the extrapolated agn far - ir fluxes and the aperture photometry of the central regions ( table [ tab : photometry ] ) clearly shows that the agn contribution at these wavelengths is small and decreases for increasing wavelengths . within the central 5.4kpc ( @xmath38 )
the agn is predicted to contribute @xmath39 ( @xmath40 , @xmath41 ) of the total @xmath3 m emission . at @xmath42
m the agn accounts for at most 2% of the emission in the same region and less than 1% at 100 and @xmath43 m .
using the flux densities measured for @xmath38 and the following equation from @xcite , @xmath44 we estimate a @xmath45 m luminosity within the ilr region of @xmath46 .
the ratio between the agn bolometric luminosity and the ir luminosity of this region is then @xmath47 .
this is in good agreement with findings for the majority of local lirgs hosting an agn @xcite .
the central region of ngc 1365 is a bright source of mid- and far - ir emission ( see e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , although its contribution to the integrated emission of the galaxy decreases significantly at longer wavelengths .
the mid - ir morphology , especially in the irac 5.8 and @xmath48 m and in the t - recs images , resembles that of a circumnuclear ring of star formation with a diameter of approximately 1.8kpc ( see figs .
[ fig : ircentralimages ] and [ fig : trecs ] ) .
the agn and the bright mid - ir clusters m4 , m5 , and m6 identified ( see table [ tab : sources ] ) by @xcite out to @xmath49 m are also clearly identified in the gemini / t - recs image at @xmath1 m ( see fig .
[ fig : trecs ] ) .
the region to the southwest of the agn is more diffuse and similar to the h@xmath50 morphology ( see fig .
[ fig : closeups ] , upper panel ) .
we do not , however , identify any bright compact sources there .
this diffuse emission is reminiscent of the optical morphology of the region containing superstar clusters ssc3 and ssc6 along with a number of less luminous optical clusters @xcite . in the spitzer / mips @xmath3 m image the emission from the agn
is still differentiated from that of the mid - ir clusters , although the latter are not resolved individually because of the decreased angular resolution .
the herschel observations of ngc 1365 show that the far - ir agn emission becomes faint when compared to that of the clusters .
this is not only because the clusters in the central region of ngc 1365 are still intrinsically bright in the far - ir , but also because in general the agn emission is observed to fall steeply in the far - ir ( see e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
moreover , in general the far - ir emission of agn appears to be dominated by a starburst component @xcite .
it is also apparent from fig .
[ fig : ircentralimages ] that the group of clusters m5 ... m8 are brighter than the group of m2 and m3 in the far - ir , as is also the case in the mid - ir
. this implies , as will be discussed further in section [ sec : sfr ] , that the obscured star formation rate ( sfr ) is higher in the former group than in the latter .
moreover , the morphology of the sfr tracers ( e.g. , the @xmath51 , @xmath52 , and @xmath42 m emissions ) is similar to that of the molecular gas , especially that of the rare c@xmath53o(j=2 - 1 ) isotope @xcite , where a `` twin - peak '' morphology is observed .
such molecular gas twin peaks are observed in barred galaxies @xcite and are formed where dust lanes intersect nuclear rings of star formation , as is the case of ngc 1365 .
we can also compare the ir morphology , in particular at 8 , 24 , and @xmath42 m , of the ilr region with that of h@xmath50+[nii ] image shown in fig .
[ fig : closeups ] has not been continuum subtracted , and thus we restrict our discussion to the h@xmath50 bright hot spots identified by other works ( e.g. , * ? ? ?
* ; * ? ? ?
since all these emissions probe the sfr in galaxies @xcite , a good morphological correspondence on scales of hundreds of parsecs is expected .
[ fig : closeups ] indeed shows that the h@xmath50 hot spots in the ilr region ( table 4 , and also * ? ? ?
* ; * ? ? ?
* ) are clearly associated with ir emitting regions ( compare with the irac @xmath48 m image as it has the best angular resolution ) .
this seems to be the case out to @xmath54 m . at longer wavelengths the decreased angular resolution of the herschel images
does not allow us to compare the morphologies .
the opposite is not necessary true because the h@xmath50 emission is strongly affected by extinction caused by the prominent dust lane crossing the ilr region .
mid - ir clusters m4 and m5 appear to be detected in the h@xmath50 image , but they are very faint .
this is because they are in one of the regions with the highest extinction , as we shall see in more detail in section [ sec : silicatefeature ] .
the spitzer / irs map of the [ neii]@xmath4 m emission covering the central 2.2kpc of ngc 1365 is shown in fig.[fig : spectralmaps ] .
this emission line is a good tracer of the current sfr because its luminosity is proportional to the number of ionizing photons @xcite .
the [ neii ] emission appears as a partial ring and it is enhanced in the mid - ir clusters .
its morphology is similar to the pacs @xmath42 m morphology and the co `` twin peaks '' discussed by @xcite .
the agn of ngc 1365 , on the other hand is not a bright [ neii]@xmath4 m source , as found for other active galaxies ( see e.g. , * ? ? ?
analogously to other sfr indicators , namely the 24 and @xmath42 m emissions , the region to the northeast of the agn ( clusters m4 , m5 , and m6 ) is brighter in [ neii ] than the region of clusters m2 and m3 .
the regions containing these bright mid - ir clusters account for approximately @xmath55 of the total [ ne ii]@xmath4 m emission in the central 2.2kpc ( see table [ tab : spectroscopy ] ) .
note , however that the emission from the clusters is unresolved at the spitzer / irs sl angular resolution and we did not apply a correction for unresolved emission .
therefore this contribution from the clusters should be taken as a lower limit .
figure [ fig : spectralmaps ] also shows the maps of the 6.2 and the @xmath5 m pah features .
both features probe mostly the emission from b stars rather than that from massive on - going star formation ( o stars , see * ? ? ? * ) .
the @xmath56 m pah morphology is very similar to [ neii ] , whereas the @xmath5 m pah map appears to have a deficit of emission in the region to the northeast of the agn .
the map of the ratio between these two features is in fig .
[ fig : silicatespahratio ] .
although several works found variations in the pah ratios of galaxies not explained by changes in the extinction ( see e.g. , * ? ? ?
* ; * ? ? ?
* ) , in the case of the central region of ngc 1365 it might be entirely due to extinction .
the 6.2 to @xmath5 m pah ratio is strongly increased in the region enclosing clusters m4 , m5 , m6 , which is also the region suffering the highest extinction ( see section [ sec : silicatefeature ] and fig . [ fig : silicatespahratio ] ) . since the relative absorption at @xmath5 m is higher than at @xmath56 m ( see e.g. , table 5 in * ? ? ?
* ) , correcting the @xmath5 m pah map for extinction would produce a morphology similar to that of the @xmath56 m pah map .
the measured ew and pah ratios of the regions not affected by strong extinction are similar to those measured in other starburst galaxies ( see e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we conclude that the star formation activity inside the ilr region of ngc 1365 has been taking place for at least a few tens of million years , as the pah features trace the emission from b stars @xcite , and is taken place currently because of the bright [ neii]@xmath4 m emission requires the presence of young ionizing stars .
the circumnuclear region of ngc 1365 is crossed by a prominent dust lane entering the ring ( see * ? ? ? * and references therein and also figs .
[ fig : pacs100_large ] and [ fig : closeups ] ) .
this dust lane is still apparent in the shortest wavelength irac bands and can be seen passing between clusters m5 and m6 at @xmath57 m ( see fig . [
fig : ircentralimages ] ) .
the brightest mid - ir clusters m4 , m5 , and m6 are located at the edge of the dust features where the bar dust lane enters the ilr region to the northeast of the agn ( * ? ? ?
* and also figs .
[ fig : ircentralimages ] and [ fig : closeups ] ) .
mid - ir clusters m2 and m3 to the southwest of the nucleus are in a region less affected by extinction and are likely to be associated with optical super star clusters and optically - detected hii regions @xcite .
the broad feature at @xmath58 m is believed to be produced by amorphous silicate grains and in _ normal _ star - forming galaxies is observed in mild absorption @xcite . in the simplest dust geometry of a purely absorbing foreground screen
the observed apparent depth of this feature is proportional to the optical extinction .
the general variation of the obscuration inside the ilr region of ngc 1365 revealed by the optical imaging can be traced with the spectral map of the silicate feature . as can be seen from fig .
[ fig : silicatespahratio ] , the apparent strength of the silicate feature is high ( in absorption ) in the region encompassing clusters m4 , m5 , and m6 to the northeast of the agn , almost zero at the location of the agn , and intermediate ( also in absorption ) around clusters m2 and m3 . in table [ tab : spectroscopy ] we list the strength of the silicate feature measured in the sl 1d spectra of selected regions . assuming a foreground screen of dust and adopting the @xcite extinction law ( @xmath59 ) , the observed apparent depths of the star - forming regions in the circumnuclear region of ngc 1365 imply values of the visual extinction of between @xmath60mag and @xmath61mag for the regions of clusters m2+m3,and m4+m5+m6 , respectively .
the high value of the extinction derived for the latter region is entirely consistent with the values derived by @xcite for the individual clusters using hydrogen recombination lines . our value of the extinction in the region containing m2 and m3 is higher than the optical estimate for the l2 and l3 hii regions from @xcite . the apparent strength of the silicate feature of the agn measured from the spitzer / irs sl spectrum is @xmath62 ( table [ tab : spectroscopy ] ) indicating that the feature is present slightly in absorption .
this value , however , is contaminated by extended emission as is apparent from the @xmath48 m image ( see fig .
[ fig : closeups ] ) and the presence of pah features in the irs nuclear spectrum ( fig .
[ fig : slspectra ] ) .
indeed , the gemini / t - recs high angular resolution spectrum of the agn is mostly a featureless continuum with no evidence of pah feature emission ( fig .
[ fig : torusmodel ] ) .
finally , the relatively moderate strength of the silicate feature of the integrated @xmath63kpc central region ( @xmath64 , table [ tab : spectroscopy ] ) of ngc 1365 m ir luminosity of this galaxy is @xmath65 using the iras flux densities from @xcite . ]
is typical of the observed nuclear silicate strengths measured in other local lirgs @xcite and indicates an average visual extinction in this region of @xmath66mag .
the herschel images can be used to trace the spatial variations of the temperature of the dust inside the ilr region of ngc 1365 .
if we assume that the far - ir emission of a galaxy can be approximated with a modified blackbody , then in the case of optically thin emission the flux density can be expressed as @xmath67 where @xmath68 is the blackbody function for a dust temperature of @xmath69 and @xmath21 is the dust emissivity . in the simplest approximation
we can calculate the color temperature @xmath70 of the dust using the ratio of the surface brightness at two wavelengths with the following equation expressed in terms of wavelengths @xmath71 we chose to construct a map of the pacs @xmath72 m to the pacs @xmath42 m ratio as these two wavelengths provide the best angular resolutions with herschel .
we rebinned the pacs @xmath42 m image to the pixel size of the pacs @xmath72 m image .
we matched the psf of the two images by smoothing the @xmath42 m with a gaussian function , although this may introduce some artifacts ( see e.g. , * ? ? ?
* ; * ? ? ?
we then calculated the dust color temperature by solving equation [ eq : colortemperature ] .
we fixed the value of the dust emissivity @xmath73 , as it is found to fit well the integrated seds of local ir galaxies @xcite .
the map of the pacs @xmath72 m to @xmath42 m ratio of the inner @xmath74 is shown in fig [ fig : colortemperature ] .
the observed range of @xmath75 in the central region translates into values of the dust color temperature of between @xmath76k and @xmath77k . within the ilr region of ngc 1365
the highest @xmath78 color temperatures correspond to regions actively forming stars , that is , the regions containing the mid - ir star clusters and the l4 h@xmath50 hot spot .
these bright star forming regions have color temperatures similar to , although slightly higher than , those of hii regions in the spiral arms .
the regions with the coldest @xmath78 in the ilr region might be associated with some of the foreground dust features seen in the optical images ( see fig .
[ fig : pacs100_large ] ) .
finally , the region of the agn does not appear different in terms of the color temperature when compared to the bright star forming regions .
this is probably because the agn is faint in the far - ir ( see next section ) coupled with the relatively coarse angular resolution of this map ( i.e. , that of the pacs @xmath72 m ) . in this section
we study the properties of the dust in the ilr region of ngc 1365 , particularly the dust heated by processes other than the agn .
to do so , we first subtracted the agn emission ( see table [ tab : agnfluxes ] ) from the observed fluxes in the spitzer / mips @xmath3 m band and all the herschel bands .
we then fitted the observed non - agn mid- and far - ir sed using a combination of two modified blackbody following @xcite and @xcite : @xmath79 where @xmath80 and @xmath81 are the relative masses of the warm and cold dust components and @xmath82 and @xmath83 are their temperatures .
as done in section [ sec : colortemperature ] , we fixed the dust emissivity @xmath84 and then used a standard @xmath85 minimization technique to fit the two dust temperatures and the relative masses of the two dust components . for the fit we used the fluxes measured through an @xmath38 aperture , and normalized them to that at @xmath72 m .
for the spitzer / mips @xmath3 m flux density , we added in quadrature the photometric error and the uncertainty associated with subtracting the agn component at this wavelength to a total error budget of @xmath86 .
for the herschel data points because the agn emission in the far - ir is very small compared to the total emission inside the ilr region , the dominant source of error is that associated with the photometric calibration of the data : @xmath87 .
[ fig : sedfit ] shows the best fit to the agn - subtracted sed of the ilr region of ngc 1365 , which was obtained with @xmath88k and @xmath89k , and relative masses of cold to warm dust of @xmath90 .
it is worth noting that the fit using two modified blackbodies is formally better than a fit to the far - ir data ( i.e. , excluding the @xmath3 m data point ) using a single modified blackbody .
this can be explained because there is a non - negligible contribution from the warm dust component to the herschel pacs @xmath42 m flux density , as found for the integrated emission of other nearby galaxies ( see e.g. , * ? ? ?
* ; * ? ? ?
however , the temperature of the warm component is not tightly constrained since varying the warm dust temperature by as much as @xmath91k produces statistically similar good fits ( @xmath92 ) .
this is because the peak of this component at around @xmath93 m is not well sampled with the present data and the mass contribution of the warm component is small .
the temperature of the cold component , on the other hand is well constrained . to get an estimate of the uncertainties on the fitted cold dust temperature , we changed it while fixing the temperature of the warm component . as can be seen from fig .
[ fig : sedfit ] , cold dust temperatures in the range @xmath94k provide acceptable fits to the sed .
previous works showed that the integrated mid - ir and far - ir seds of seyfert galaxies could be fitted with a combination of various dust temperatures probing different heating mechanisms .
these include warm dust associated with the agn torus , cold dust similar to that observed in starburst galaxies , and very cold dust at temperatures characteristic of dust heated by the interstellar medium @xcite . with the superior angular resolution of the herschel observatory
we can now obtain spatially resolved observations of nearby seyfert galaxies and star forming galaxies .
the dust temperature of the ring of star formation in ngc 1365 is similar to that of other nuclear and circumnuclear starbursts with or without an agn with similar ir luminosities and physical sizes , for instance , the ring of ngc 3081 @xcite and the nuclear region of m83 @xcite . on the other hand ,
the dust of mrk 938 , which is part of our survey of seyfert galaxies , has a considerably higher dust temperature of @xmath95k , probably due to the smaller size of the ir emitting region of this galaxy and higher ir luminosity @xcite .
we calculated the dust mass ( @xmath96 ) within the ilr region of ngc 1365 using the following equation from @xcite ( adapted from * ? ? ?
* ) : @xmath97 where @xmath98 is the observed flux density , @xmath99 is the luminosity distance , @xmath100 is the blackbody emission for the best fitting dust temperatures , and @xmath101 is the dust absorption coefficient . as done by @xcite , we evaluated this expression at @xmath102 m using an absorption coefficient of @xmath103 , as interpolated from the dust model of @xcite .
we derived a dust mass in the ilr region of @xmath104 , which accounts for approximately 25% of the total dust mass of this galaxy @xmath105 .
the latter estimate is from @xcite but recalculated for the absorption coefficient used in this work .
the ages and masses of the mid - ir clusters ( @xmath106myr and @xmath107 , * ? ? ?
* ) and the large amount of molecular gas available in the central regions of this galaxy @xcite indicate that there is intense on - going star formation activity there .
we can use the ir observations of ngc 1365 to estimate the obscured sfr in the ilr region , and compare it with the unobscured sfr . as discussed by @xcite , a combination of the observed ( not corrected for extinction ) h@xmath50 luminosity and a mid - ir monochromatic luminosity ( preferably @xmath3 m ) will provide the best estimate of the total sfr in a moderately obscured environment ( see section [ sec : silicatefeature ] ) , such as the ilr region of ngc 1365 .
we note that this empirically calibrated recipe includes contributions from dust heating from all stars and not only the youngest ( see * ? ? ?
* for a full discussion ) . using the h@xmath50 flux of @xcite for a 40-diameter aperture and our @xmath3 m flux ( after subtracting the agn component ) we estimated a total sfr within the ilr region ( inner @xmath108kpc , @xmath38 ) of ngc 1365 of sfr@xmath109 for a salpeter initial mass function ( imf ) .
approximately 85% of this sfr is contributed by the @xmath3 m emission , and thus it originates from dust - obscured star forming regions .
most of the ir emission in the ilr region comes from the star formation ring containing the bright mid - ir clusters identified by @xcite .
indeed , we estimate that within the inner @xmath110 the total sfr is @xmath111 .
this has been calculated using the agn - subtracted @xmath3 m flux density ( @xmath112 ) and the 24-diameter aperture h@xmath50 flux from @xcite .
this corresponds to a sfr surface density in the ring of @xmath113 , for a ring radius of 900pc .
this value of the sfr density is similar to those of other circumnuclear starbursts @xcite and is expected to be larger by factors of @xmath114 compared to the disk of the galaxy @xcite .
in this paper we have studied the ir emission associated with the star formation activity in the ilr region of ngc 1365 , as well as the ir emission of the agn . to this end we have analyzed new far - ir ( @xmath0 m ) herschel / pacs and spire imaging , and high angular resolution ( @xmath115 ) gemini / t - recs mid - ir imaging and spectroscopy of this galaxy .
we have also made use of archival spitzer / irac and mips imaging and irs spectral mapping data .
our main findings for the inner @xmath116kpc region of ngc 1365 are : * the new herschel pacs imaging data at 70 , 100 , and @xmath43 m reveal that the ring of star formation in the ilr region is bright in the far - ir .
the agn is the brightest mid - ir source in the inner 2kpc up to @xmath117 m , but it becomes increasingly fainter in the far - ir when compared with the mid - ir clusters or groups of them in the ring . * the 24 and @xmath42 m emissions as well as the [ ne ii]@xmath4 m line and pah features trace the star - forming ring in the ilr region and have morphologies similar to the co `` twin - peaks '' .
this all indicates that there is intense on - going star formation taking place in the inner few kpc of ngc 1365 . *
the unresolved near and mid - ir nuclear emission and mid - ir spectrum ( i.e. , agn - dominated emission ) of ngc 1365 are well reproduced with a relatively compact torus ( outer radius of @xmath118pc ) with an opening angle of @xmath119deg , and an agn bolometric luminosity @xmath120 using the clumpy torus models .
these parameters are in good agreement with independent estimates in the literature . * using the fitted torus model we quantified the agn emission in the far - ir .
the agn only contributes at most 1% of the @xmath42 m emission within the inner 5.4kpc ( @xmath38 ) , and less than 1% at longer wavelengths . at @xmath3
m the agn accounts for @xmath121 of the emission in the same region .
we estimated that the agn bolometric contribution to the @xmath45 m luminosity in the inner 5.4kpc is approximately 5% .
* the non - agn 24 to @xmath122 m sed of the ilr region ( inner 5.4kpc ) of ngc 1365 is well fitted with a combination of two modified blackbodies with warm and cold temperatures of 54k and 24k , respectively
. however , the cold dust component accounts for most of total dust mass inferred in this region ( @xmath123 ) and has a temperature similar to that of other nuclear and circumnuclear starbursts of similar sizes and ir luminosities . * from the comparison between the sfr from h@xmath50 ( unobscured ) and the sfr from @xmath3 m ( obscured ) we infer that up to @xmath124 of the on - going sfr inside the ilr region of ngc 1365 is taking place in dust - obscured regions in the ring of star formation .
we are grateful to b. elmegreen and e. galliano for providing us with the bvr map of ngc 1365 shown in the right panel of figure 1 .
we also thank andrs asensio ramos for developing the bayesclumpy fitting routine .
we thank an anonymous referee for comments that helped improve the paper .
, m.p .- s . , and p.e .
acknowledge support from the spanish plan nacional de astronoma y astrofsica under grant aya2009 - 05705-e .
h . also acknowledges support from the universidad de cantabria through the augusto gonzlez linares program and aya2010 - 21161-c02 - 01 .
c.r.a . acknowledges the spanish ministry of science and innovation ( micinn ) through project consolider - ingenio 2010 program grant csd2006 - 00070 : first science with the gtc ( http://www.iac.es/consolider-ingenio-gtc/ ) and the spanish plan nacional grant aya2010 - 21887-c04.04 .
m.p . acknowledges the junta de andaluca and the spanish ministry of science and innovation through projects po8-tic-03531 and aya2010 - 15169 , respectively . _
herschel _ is an esa space observatory with science instruments provided by european - led principal investigator consortia and with important participation from nasa .
pacs has been developed by a consortium of institutes led by mpe ( germany ) and including uvie ( austria ) ; ku leuven , csl , imec ( belgium ) ; cea , lam ( france ) ; mpia ( germany ) ; inaf - ifsi / oaa / oap / oat , lens , sissa ( italy ) ; iac ( spain ) .
this development has been supported by the funding agencies bmvit ( austria ) , esa - prodex ( belgium ) , cea / cnes ( france ) , dlr ( germany ) , asi / inaf ( italy ) , and cicyt / mcyt ( spain ) .
spire has been developed by a consortium of institutes led by cardiff univ .
( uk ) and including : univ .
lethbridge ( canada ) ; naoc ( china ) ; cea , lam ( france ) ; ifsi , univ .
padua ( italy ) ; iac ( spain ) ; stockholm observatory ( sweden ) ; imperial college london , ral , ucl - mssl , ukatc , univ .
sussex ( uk ) ; and caltech , jpl , nhsc , univ .
colorado ( usa ) .
this development has been supported by national funding agencies : csa ( canada ) ; naoc ( china ) ; cea , cnes , cnrs ( france ) ; asi ( italy ) ; mcinn ( spain ) ; snsb ( sweden ) ; stfc , uksa ( uk ) ; and nasa ( usa ) . based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
lira p. , videla l. , wu y. , alonso - herrero a. , alexander d. m. , ward m. , 2012 , apj , submitted mason r. e. , levenson n. a. , shi , y .
, packham c. , gorjian v. , cleary k. , rhee j. , werner m. , 2009 , apj , 693 , l136 | we present new far - infrared ( @xmath0 m ) herschel pacs and spire imaging observations as well as new mid - ir gemini / t - recs imaging ( 8.7 and @xmath1 m ) and spectroscopy of the inner lindblad resonance ( ilr ) region ( @xmath2kpc ) of the spiral galaxy ngc 1365 .
we complemented these observations with archival spitzer imaging and spectral mapping observations .
the ilr region of ngc 1365 contains a seyfert 1.5 nucleus and a ring of star formation with an approximate diameter of 2kpc .
the strong star formation activity in the ring is resolved by the herschel / pacs imaging data , as well as by the spitzer @xmath3 m continuum emission , [ ne ii]@xmath4 m line emission , and 6.2 and @xmath5 m pah emission .
the agn is the brightest source in the central regions up to @xmath6 m , but it becomes increasingly fainter in the far - infrared when compared to the emission originating in the infrared clusters ( or groups of them ) located in the ring .
we modelled the agn unresolved infrared emission with the clumpy torus models and estimated that the agn contributes only to a small fraction ( @xmath7 ) of the infrared emission produced in the inner @xmath8kpc .
we fitted the non - agn @xmath9 m spectral energy distribution of the ilr region and found that the dust temperatures and mass are similar to those of other nuclear and circumnuclear starburst regions .
finally we showed that within the ilr region of ngc 1365 most of the on - going star formation activity is taking place in dusty regions as probed by the @xmath3 m emission .
[ firstpage ] galaxies : evolution galaxies : nuclei galaxies : seyfert galaxies : structure infrared : galaxies galaxies : individual : ngc 1365 . |
over the last years there is a high interest in the investigation of exotic e1 modes , pygmy dipole resonance ( pdr ) , compression dipole resonance ( cdr ) , and toroidal dipole resonance ( tdr ) , see e.g. reviews @xcite .
the schematic images of these modes are illustrated in fig . 1 .
pdr is the working title of a collection of marked dipole strength at energies below the dipole giant resonance .
although the sub - peaks in the pdr region can have different microscopic structure , one often visualizes their averaged flow pattern as oscillations of the neutron excess against the nuclear core with n = z @xcite .
so this resonance can exist only in nuclei with large neutron excess . as seen from fig .
1a ) , the schematic pdr flow is irrotational .
this resonance gathers effects from various nuclear matter properties ( symmetry energy , incompressibility , effective masses ) @xcite and so can provide useful complementing information on the nuclear equation of state and thus for astrophysical applications .
the cdr also represents an irrotational flow .
it is , as the name says , a dipole mode , but with compressional character becoming apparent in a higher order radial profile @xcite , see fig .
this resonance can be used as a source of an additional information on the nuclear incompressibility @xcite .
further , the tdr is a remarkable example of _ vortical _ dipole motion , see fig .
for the first time , the toroidal fraction of the nuclear convection current was inspected by v.m .
dubovik and a.a .
cheshkov @xcite .
later s.f .
semenko predicted the tdr in atomic nuclei @xcite .
the tdr is related to cdr @xcite . in the isoscalar ( t=0 ) channel ,
both resonances constitute the low - energy ( tdr ) and high - energy ( cdr ) parts of the isoscalar giant dipole resonance ( isgdr ) observed in @xmath0 reaction @xcite .
after extraction of the center - of - mass corrections ( cmc ) , the tdr and cdr become the dominant dipole modes in the isoscalar channel .
all three exotic modes , pdr , cdr and tdr , have isoscalar ( t=0 ) and isovector ( t=1 ) branches @xcite . in the present review ,
we will concentrate on the isoscalar branches of the modes .
they are easier accessible in scattering experiments and , as shown below , deliver indeed rich physics .
some properties of the pdr / cdr / tdr were already discussed in the literature @xcite . in particular , a modest collectivity of the modes were established and it was worked out that complex configurations play a crucial role in formation of the fine structure of the strengths . in the present review , we will focus on the vortical representative of this family - isoscalar tdr @xcite .
namely , we will outline some fascinating features of this tdr , recently explored by our group : relation to the pdr @xcite , possibility to use tdr as a measure of the dipole nuclear vorticity @xcite , and anomalous deformation splitting of the tdr @xcite .
the analysis is based on the self - consistent calculations within the skyrme quasiparticle random - phase - approximation ( qrpa ) approach @xcite with the factorized @xcite and exact @xcite residual interaction .
[ fig1_image ] the paper is organized as follows . in section 2 , the basic formalism for the tdr and cdr is given . the calculation scheme is outlined . in section 3 , we demonstrate that the pdr can be treated as a manifestation of the tdr at the nuclear surface . in section 4 , the possibility to use the toroidal strength as a measure of the nuclear dipole vorticity is discussed .
it is shown that this prescription is more accurate and robust than the familiar ravenhall - wambach recipe @xcite . in section 5
, we inspect the anomalous deformation splitting of the tdr in axial nuclei ( an opposite order of k - branches as compared to the case of the giant dipole resonance - gdr ) .
this feature can be used as an unambiguous fingerprint of the tdr in experiment .
finally , in section 6 the summary is done .
in the long - wave approximation ( @xmath1 ) , the standard electric multipole operator reads @xmath2 where @xmath3 is the familiar electric operator and @xmath4 \nonumber \\ \nonumber & = & -\frac{1}{2c } \sqrt{\frac{\lambda}{\lambda+1}}\,\frac{1}{2\lambda+3 } \int d^3r \ : r^{\lambda+2}\mathbf{y}_{\lambda\lambda\mu } \\
\label{tm_curlj } & & \cdot \left(\mathbf{\nabla } \times \hat{\mathbf{j}}_{\text{nuc}}\right)\end{aligned}\ ] ] is the toroidal operator @xcite . here
@xmath5 is the operator of the nuclear current consisting of convection and magnetization parts .
further , @xmath6 is the density operator , @xmath7 and @xmath8 are vector and ordinary spherical harmonics . for the sake of brevity
, we will skip below ( up to the cases of a possible confusion ) the coordinate dependence in currents , densities and spherical harmonics .
equation ( [ e+ktor ] ) shows that the toroidal operator comes as a second order ( @xmath9 ) correction to the dominant electric operator ( [ e_oper_main ] ) .
being determined by the curl of the nuclear current @xmath10 , the toroidal flow is obviously vortical and thus decoupled from the continuity equation .
so the toroidal operator can not be presented through the nuclear density alone and needs the knowledge of the current distribution .
the multipole compression operator reads @xcite @xmath11 where @xmath12 is its familiar density - dependent form @xcite .
the compression operator appears as a probe operator for excitation of the isgdr @xcite . unlike the toroidal one
, this operator may be presented in both current- and density - dependent forms .
the operator is determined by the divergence of current @xmath13 and so is irrotational .
note that , though the toroidal and compression operators represent essentially different flows , they are closely related . accordingly there is a coupling between the tdr and cdr @xcite .
now we will focus on the isoscalar dipole e1(t=0 ) excitations . following @xcite ,
the role of the magnetization nuclear current for tdr and cdr in e1(t=0 ) channel is negligible .
so only the convection part @xmath14 of the nuclear current is considered below . in e1(t=0 ) channel , the toroidal and compression operators
are reduced to @xcite @xmath15 , \\
\label{cm_oper } & & \hat m_{\text{com}}(e1\mu ) = -\frac{i}{2\sqrt{3}c } \int d^3r \hat{\bf j}_{c } \\
\nonumber & & \quad \cdot \ ; [ \frac{2\sqrt{2}}{5 } r^2{\mathbf{y}}_{12\mu } - ( r^2 - \langle r^2\rangle_0 ) { \mathbf{y}}_{10\mu } ] , \\ \label{cmp_oper } & & \hat m'_{\text{com}}(e1\mu ) = \frac{1}{10 } \int d^3r \hat{\rho } \ ; [ r^3-\frac{5}{3}\langle r^2\rangle_0 r ] y_{1\mu } \ ; .\end{aligned}\ ] ] here @xmath16 is the ground - state squared radius , @xmath17 is the ground state density , @xmath18 is the mass number .
the operators ( [ tm_oper])-([cmp_oper ] ) have the center of mass correction ( c.m.c . )
proportional to @xmath19 .
the toroidal and compression matrix elements for @xmath20 transitions between the ground state @xmath21 and rpa excited state @xmath22 are determined through the current transition density ( we skip below the index @xmath23 ) @xmath24\ ] ] involving the upper @xmath25 and lower @xmath26 radial current components . then @xmath27 \
\\ \label{cm_me } \langle \nu |\hat m_{\text{com}}(e1)|0\rangle & = & -\frac{1}{6c } \int dr r^2 \\ \nonumber & \cdot & [ \frac{2\sqrt{2}}{5}r^2 j_{12}^{\nu } - ( r^2 -\langle r^2\rangle_0 ) j_{10}^{\nu } ] \ ; .\end{aligned}\ ] ] ravenhall - wambach proposed that just the current component @xmath28 determines the vorticity of the dipole motion @xcite .
then the flow can be treated as fully vortical ( @xmath290 , @xmath30=0 ) , fully irrotational ( @xmath310 , @xmath320 ) , and mixed ( @xmath290 , @xmath320 ) . following this prescription ,
both tdr and cdr are of a mixed ( irrotational / vortical ) character , which contradicts their predominantly curl- and gradient - like velocities fields exhibited in fig . 1 ,
see also detailed discussion in @xcite .
the calculations were performed within the fully self - consistent skyrme qrpa method @xcite .
we use the exact residual interaction for spherical nuclei @xcite and the separable residual interaction for axial deformed nuclei @xcite .
below these schemes are referred as rpa and srpa , respectively .
both versions are self - consistent because : i ) the mean field and residual interaction are obtained from the same skyrme functional , ii ) the residual interaction includes all the terms of the initial skyrme functional as well as the coulomb direct and exchange terms .
both time - even and time - odd densities are take into account .
the @xmath33-force volume pairing is treated at the bcs level @xcite .
the skyrme force sly6 @xcite providing a satisfactory description of the giant dipole resonance ( gdr ) @xcite is used . for a deformed nucleus , the equilibrium quadrupole axial deformation @xmath34
is determined by minimization of the total energy of the system .
the srpa code employs a mesh in cylindrical coordinates .
the calculations use a large configuration space with particle - hole ( two - quasiparticle ) energies up to 70 mev . the energy - weighted sum rule for e1(t=1 ) transitions
is fully exhausted .
the spurious mode lies around 1 mev , i.e. safely beyond the tdr and cdr regions .
the toroidal and compression e1(t=0 ) strength functions read @xmath35 where @xmath36 is the toroidal ( @xmath37 ) or compression ( @xmath38 ) transition operator given by expressions ( [ tm_oper ] ) and ( [ cm_oper ] ) , respectively . the neutron and
proton effective charges are @xmath39 . the sum in ( [ sf_tdr ] ) runs through all the qrpa @xmath40-states .
further , @xmath41 is a lorenz weight with the averaging parameter @xmath42 .
the folding by the lorentz function simulates the smoothing effects beyond qrpa ( coupling to complex configurations and escape widths ) . in srpa calculation for axial deformed nuclei ,
we use the constant averaging parameter @xmath42=1 mev @xcite .
this suffices to demonstrate the deformation splitting of the tdr , discussed in sec .
however our rpa code for spherical nuclei allows to use a more reasonable lorentz double folding exhibiting a linear dependence of @xmath42 on the excitation energy @xmath43 above the first emission threshold @xcite .
this folding is implemented in strength distributions discussed in sec .
3 . for the description of the gdr and pdr
, we calculate within the rpa the photoabsorption @xmath44 @xcite . the neutron and proton effective charges @xmath45 and @xmath46 are used .
the lorentz energy dependent double folding @xcite is applied .
more details on the calculation scheme can be found elsewhere @xcite .
in the studies @xcite , it was reported that the pdr and tdr occupy the same energy region and thus are related .
this point was thoroughly investigated for a doubly magic nucleus @xmath47pb .
the strength functions , transition densities and current fields were inspected .
the calculations confirmed the experimental finding @xcite that the pdr is separated into isoscalar lower - energy and mixed ( isoscalar / isovector ) higher - energy branches .
the main attention was paid to the low - energy isoscalar part of the pdr @xcite .
it was shown that though the transition densities in this energy region are typical for the pdr ( dominance of the neutron flow at the nuclear surface ) , the corresponding current fields ( current transition densities ) are obviously of toroidal character .
thus it was suggested that the pdr is a peripheral manifestation of the toroidal flow . at first glance
, this suggestion looks incorrect because pdr and tdr represent essentially different irrotational and vortical kinds of nuclear motion .
this is clearly seen in the panels a ) and c ) of fig .
1 . however the same panels indicate that both motions are quite similar at the nuclear surface , see left and right boundary regions marked by the dash curves .
so the pdr could be a local peripheral manifestation of the tdr .
following fig .
1 , pdr and cdr flows are similar at the top and bottom boundary regions .
so perhaps the pdr has also some dipole compression fraction .
sn : the photoabsorption embracing the gdr and pdr ( a ) ; the e1(t=0 ) strength functions ( [ sf_tdr ] ) for the tdr ( b)and cdr ( c ) .
the panel ( b ) also shows the 1ph strength depicted by the dash line .
, title="fig:"][fig2_sf ] in this section , we briefly repeat the above analysis but now for a doubly - magic nucleus @xmath48sn characterized by a significant neutron excess . let s first consider the relevant strength functions exhibited in fig .
2 . in the panel
a ) the photoabsorption @xmath49 embracing the gdr and pdr is shown .
the toroidal and compression strengths ( [ sf_tdr ] ) are given in the panels b)-c ) .
it is easy to see that pdr and low - energy tdr / cdr peaked parts share the same energy region and so can be coupled .
the panel b ) also shows the particle - hole ( 1ph ) unperturbed toroidal strength .
one sees that , in the rpa case , the low - energy tdr peak is significantly downshifted relative to its 1ph counterpart .
this signals a noticeable collectivity of the tdr .
more detailed information of the dipole strength in this energy region can be obtained from the inspection of the transition densities ( td ) and current fields ( cf ) .
note that td and cf do not depend on the toroidal ( [ tm_oper ] ) and compression ( [ cm_oper ] ) operators and are fully determined by the rpa wave functions .
it is not worth to consider these patterns for individual @xmath40-states because they can vary from state to state and so hide common features of the flow .
instead we consider td and cf properly averaged ( summed ) over a relevant energy intervals .
the technique of computing of the summed td / cf is described in detail in @xcite .
note that this technique circumvents a possible ambiguity in the signs of the rpa wave functions . in the present study for @xmath48sn ,
the td / cf are calculated at the energy interval 6 - 10 mev covering the pdr and low - energy tdr and cdr peaks .
-weighted proton ( solid curve ) and neutron ( dash curve ) transition densities in @xmath48sn , calculated with the force sly6 .
the td are summed in the energy interval 6 - 10 mev.,title="fig:"][fig3_td ] figure 3 shows the rpa and 1ph proton and neutron transition densities in @xmath48sn . in the rpa case , we see a typical picture when the proton and neutron td are strong and in phase inside the nucleus ( @xmath504 - 6 fm ) but the neutron td strictly dominates at the nuclear boundary ( @xmath506 - 10 fm ) .
just this td behavior is often used as a justification of the pdr view in terms of the oscillation of the neutron excess against the n = z core , see fig .
a noticeable difference between the rpa and 1ph td indicates that the impact of the residual interaction is important .
the cf are summed in the energy interval 6 - 10 mev . ]
the td are still too rough characteristics since they illustrate only a radial dependence of the flow but not its angular distribution . at the same time , the latter is crucial to recognize the actual flow pattern .
thus we have to inspect the cf which reveal more details .
the cf for the energy interval 6 - 10 mev in @xmath48sn are exhibited in fig .
the proton , neutron , isoscalar , and isovector flows at the plane @xmath51 are shown .
we see that the nuclear motion at 6 - 10 mev has a clear isoscalar toroidal character , cf . with fig .
so it is quite possible that the pdr is rooted in the tdr and is actually its manifestation at the nuclear surface ( where both pdr and tdr flows look as similar irrotational patterns ) if the surface is dominated by neutron density .
note that by definition the pdr exists only in nuclei with a neutron excess .
instead the tdr is a general feature of all the nuclei including those with n = z .
the total pdr , with its isoscalar and mixed parts , is probably a complicated mixture of different kinds of the dipole motion : toroidal , compression , gdr tail , etc @xcite .
however our analysis of cf obviously shows that the toroidal contribution is of a prior importance .
all the dipole motions should be coupled and so generated simultaneously in various reactions .
a ratio between their contributions in the dipole response should depend on the applied reaction ( photoabsorption , @xmath0 , @xmath52 , etc ) .
note that some signs of the toroidal flow in the pdr region can be noticed in the velocity fields of previous rpa calculations @xcite .
a distinctive toroidal motion in the pdr region was obtained in the calculations @xcite within the quasiparticle phonon model ( qpm ) @xcite . besides
, the pdr / tdr interplay was recently discussed in the semiclassical exploration @xcite .
however the studies @xcite did not consider a possible relation between pdr and tdr while the study @xcite was done at a phenomenological level . besides , unlike our scheme , the previous rpa and qpm calculations inspected td / cf for individual states or used summed td / cf without a special care for an ambiguity in the sign of the dipole states .
there are two basic kinds of the nuclear flow : irrotational and vortical @xcite .
the irrotational motion is pertinent to collective low - energy excitations and regular electric gr @xcite .
the vortical flow takes place ( besides a plane nuclear rotation ) in single - particle excitations @xcite and exotic gr like e.g. electric tdr @xcite and twist magnetic quadrupole resonance @xcite ) .
despite some previous studies ( see e.g. @xcite ) , our knowledge about nuclear vorticity is still rather poor .
even the measure of the vorticity is disputable . in hydrodynamics ,
the vorticity is defined as curl of the velocity field @xcite , @xmath53 however nuclear physics deals not with velocities but currents . in this connection , ravenhall and wambach
have proposed the @xmath54-component of the nuclear current as a measure and indicator of the nuclear vortcity ( to be called below as rw vorticity ) @xcite .
they have shown that @xmath54 is unrestricted by the continuity equation ( ce ) @xmath55 in the integral sense and so is suitable for a divergence - free ( vortical ) pattern @xcite . for a long time , the rw prescription was used for estimation of the vortical contribution in various excitations , see e.g. @xcite .
however our last studies have shown that this prescription has serious shortcomings @xcite .
in particular , the rw scheme obviously fails for the tdr and cdr . indeed , following eqs .
( [ tm_me])-([cm_me ] ) , these modes involve both @xmath54 and @xmath56 components of the nuclear current and so have to be of a mixed ( irrotational / vortical ) nature .
at the same time , the tdr is basically vortical while the cdr is irrotational .
their velocities are the gradient and curl functions , respectively @xcite . and
their current fields closely correspond to vortical and irrotational images given in fig . 1 .
moreover , the thorough analysis of the current components @xmath57 and @xmath58 has shown that both them have strong contributions in the low - energy toroidal and high - energy compression regions @xcite . besides , the divergences and curls of these components turned out to be of the same order of magnitude .
this means that @xmath54 has no any strong advantage over @xmath56 to represent the nuclear vorticity .
neither @xmath54 nor @xmath56 alone can represent a vortical or irrotational flow .
instead eqs .
( [ tm_me ] ) and ( [ cm_me ] ) suggest that only the proper combinations of @xmath54 are @xmath56 are suitable for these aims . in this connection
, we propose the toroidal strength as a natural and robust measure of the nuclear vorticity .
following eqs .
( [ tm_me ] ) , the toroidal matrix element is determined by the combination of @xmath59 and @xmath60 , resulting in a clear vortical flow demonstrated in figs .
besides , as shown in @xcite , the toroidal and hydrodynamical criteria of the vorticity are closely related .
one may also provide some formal arguments to justify the toroidal pattern as the standard for vorticity .
we know that the electric current transition density is decomposed into longitudinal and transversal components , @xmath61 where @xmath62 and @xmath63 are some scalar functions @xcite . as compared to the prescription @xcite , this definition looks more plausible for the search of ce - unrestricted divergence - free current because we immediately obtain @xmath64 as a natural candidate .
the current components can be expanded in the basis of eigenfunctions @xmath65 ( @xmath66 = - , 0 , + ) of the vector helmholtz equation ( in analogy to the expansion of the vector - potential , in @xcite )
. then the transversal component reads @xmath67 where @xmath68 is the electric transversal form - factor and integration by @xmath69 is assumed . in the long - wave approximation ( @xmath1 ) , the transversal component is reduced to the longitudinal one .
after subtraction of this long - wave part from @xmath70 , we get at @xmath71 the toroidal current density @xcite .
the transversal character of such toroidal current is confirmed by expression ( [ tor_rel_1 ] ) .
being independent from @xmath72 and thus decoupled from ce , the toroidal current is a relevant vortical part of the complete nuclear current , suitable as a robust measure of the nuclear vorticity .
the gdr in axial nuclei exhibits the deformation splitting into k=0 and k=1 branches @xcite . in prolate nuclei ,
the branch k=0 has a lower energy than the k=1 one , @xmath73 .
the opposite sequence , @xmath74 , takes place in oblate nuclei .
this splitting can be easily explained in terms of dipole oscillations along @xmath75- and @xmath76-axes .
the deformation splitting should also take place in the tdr .
however , the calculations for the tdr ( prolate @xmath77yb @xcite ) have surprisingly revealed an anomalous sequence @xmath78 , i.e. opposite to the order typical for gdr in oblate axial nuclei .
it was shown that this sequence takes place in both srpa and 2qp toroidal strength functions .
so the effect is not caused by the residual interaction .
u. in both panels , the total strength ( solid curve ) and its k=0 ( dash curve ) and k=1 ( dotted curve ) branches are depicted . ] here we discuss the anomalous deformation splitting of tdr in prolate @xmath79u .
the corresponding srpa and two - quasiparticle ( 2qp ) strength functions are depicted in fig .
the strengths are smoothed by a lorentz weight with the constant averaging parameter @xmath42= 1 mev .
the deformation parameter @xmath34=0.286 is obtained by minimization of the total energy of the system .
note that the calculated @xmath34 excellently reproduces the experimental value @xmath80=0.2863(24 ) @xcite .
figure 5 shows that , in accordance with previous results for rare - earth nuclei @xcite , the tdr in @xmath79u also exhibits the anomalous splitting .
indeed , in rpa case , the main k=1 peak lies significantly lower ( 5 - 10 mev ) than the main k=0 peak ( 10 - 17 mev ) . in the low - energy region 5 - 12
mev , the k=1 strength strongly dominates over k=0 one .
though both branches are distributed in a wide energy interval and have more or less the same strength at e@xmath81 12 mev , it is easy to see that the k=1 centroid is certainly lower than the k=0 one . as seen from fig .
5a ) , a similar picture emerges for the unperturbed ( without residual interaction ) two - quasiparticle ( 2qp ) strength . the anomalous splitting of the tdr might be related with the fact that tdr , unlike the regular @xmath82 giant resonances determined by @xmath83-fields , is not the tassie mode . following ( [ tor_rel_1 ] ) ,
the toroidal multipole field rather includes @xmath84 .
this means that simple arguments explaining the deformation splitting in the gdr do not work here .
however , following expression ( [ cm_divj1 ] ) , the cdr is also not the tassie mode . at the same time
, the calculations do not show in the cdr the distinctive `` opposite order '' splitting in deformed nuclei @xcite . perhaps here
the vortical character of the tdr is a decisive factor .
this problem deserves further exploration .
anyway , the anomalous splitting of the tdr can be used for its experimental discrimination from other dipole modes .
some remarkable properties of the isoscalar toroidal dipole resonance ( tdr ) , recently studied by our group , were briefly discussed .
the main attention was paid to a ) the relation of the tdr and low - energy dipole strength ( also denoted as pdr ) @xcite , b ) the possibility to use the toroidal flow as a measure of the nuclear dipole vorticity @xcite , and c ) anomalous deformation splitting of the tdr @xcite .
the analysis was based on the calculations within the self - consistent quasiparticle random - phase - approximation ( qrpa ) method @xcite with the skyrme force sly6 @xcite .
two qrpa versions , exact rpa for spherical nuclei @xcite and separable rpa ( srpa ) for deformed nuclei @xcite were used . as relevant examples ,
the spherical @xmath48sn and deformed @xmath79u were considered .
the vortical tdr was discussed together with its irrotational counterpart , the compression dipole resonance ( cdr ) .
our study has confirmed previous findings @xcite .
namely , for the case of @xmath48sn , we showed that the tdr and pdr share the same energy region .
the field of the nuclear convection current in this region is clearly toroidal . both tdr and pdr flows have much in common at the nuclear surface .
so it is quite possible that the pdr is actually a local manifestation of the tdr at the nuclear boundary . besides we presented the arguments that the familiar measure of the nuclear vorticity proposed by ravenhall and wambach @xcite ( @xmath28 component of the nuclear current ) is not relevant and obviously fails in the tdr / cdr case . instead , the toroidal current and strength
are much better suited for this aim . finally , using the srpa results for @xmath79u , we considered the anomalous deformation splitting of the tdr . in accordance to previous study @xcite
, we have found that in prolate @xmath79u the k=1 branch of the tdr has lower energy than the k=0 branch , i.e. we received the opposite order of the branches as compared to the regular giant dipole resonance .
the nature of this feature is not yet clear and needs a further study . anyway
this feature may be used for an experimental discrimination of the tdr .
note that for dipole excitations in the pdr region the coupling with complex configurations ( ccc ) can be important @xcite .
however , we think that the toroidal features we have found are too strong to be spoiled by this effect .
this is confirmed e.g. by qpm calculations which include the ccc @xcite but nevertheless report a strong toroidal flow at the pdr region .
the isoscalar tdr and cdr dominate the e1(t=0 ) channel and can be observed in @xmath85 reaction , see e.g. @xcite .
so , despite of their second - order character @xcite , the pdr and cdr are accessible and important examples of the dipole excitations .
the tdr is the only vortical mode in e1(t=0 ) channel .
the tdr and pdr share the same energy region and are certainly related .
now the pdr is intensively investigated as it provides an important information for the nuclear equation of state and astrophysical applications @xcite .
the analysis of the tdr / pdr interplay can be essential for a better understanding of the pdr features and related topics .
the work was partly supported by dfg ( re 322/14 - 1 ) , heisenberg - landau ( germany - bltp jinr ) , and votruba - blokhintsev ( czech republic - bltp jinr ) grants .
j.k . and a.r .
are grateful for the support of the czech science foundation ( p203 - 13 - 07117s ) .
99 n. paar , d. vretenar , e. khan , and g. colo , rep .
. phys . * 70 * , 691 ( 2007 ) .
d. savran , t. aumann , and a. zilges , prog .
. phys . * 70 * , 210 ( 2013 ) .
reinhard and w. nazarewicz , phys .
c*87 * , 014324 ( 2013 ) .
harakeh et al , phys .
lett . * 38 * , 676 ( 1977 ) .
s. stringari , phys .
b*108 * , 232 ( 1982 ) .
v.m . dubovik and a.a .
cheshkov , sov .
nucl . * 5 * , 318 ( 1975 ) .
semenko , sov .
j. nucl . phys . *
34 * , 356 ( 1981 ) .
d. vretenar , n. paar , p. ring , and g.a .
lalazissis , phys . rev .
c*63 * , 047301 ( 2001 ) .
j. kvasil , v.o .
nesterenko , w. kleinig , p .-
reinhard , and p. vesely , phys .
c*84 * , 034303 ( 2011 ) .
m. uchida et al , phys .
c*69 * , 051301(r ) ( 2004 ) .
j. kvasil , v.o .
nesterenko , a. repko , w. kleinig , p .-
reinhard , and n. lo iudice , phys .
t*154 * , 014019 ( 2013 ) .
a. repko , p .-
reinhard , v.o .
nesterenko , and j. kvasil , phys .
c*87 * , 024305 ( 2013 ) .
nesterenko , a. repko , p .-
reinhard , and j. kvasil , epj web of conferences * 93 * , 01020 ( 2015 ) .
reinhard , v.o .
nesterenko , a. repko , and j. kvasil , phys .
c*89 * , 024321 ( 2014 ) .
j. kvasil , v.o .
nesterenko , w. kleinig , and p .-
reinhard , phys . scr . * 89 * , 054023 ( 2014 ) .
skyrme , phil . mag . * 1 * , 1043 ( 1956 ) .
d. vauterin and d.m .
brink , phys .
c*5 * , 626 ( 1972 ) .
m. bender , p .- h .
heenen , and p .-
reinhard , rev .
phys . * 75 * , 121 ( 2003 ) .
nesterenko , j. kvasil , and p .-
reinhard , phys .
c*66 * , 044307 ( 2002 ) .
nesterenko , w. kleinig , j. kvasil , p. vesely , p .-
reinhard , and d.s .
dolci , phys .
c*74 * , 064306 ( 2006 ) .
a. repko , j. kvasil , v.o .
nesterenko , and p .-
reinhard , arxiv:1510.01248[nucl - th ] .
ravenhall and j. wambach , nucl .
a*475 * , 468 ( 1987 ) .
p. ring and p. schuck , _ nuclear many body problem _ , ( springer - verlag , new york , 1980 ) .
m. bender , k. rutz , p .-
reinhard , and j.a .
maruhn , eur .
j. a*8 * , 59 ( 2000 ) .
e. chabanat , p. bonche , p. haensel , j. meyer , and r. schaeffer , nucl .
a*635 * , 231 ( 1998 ) .
w. kleinig , v.o .
nesterenko , j. kvasil , p .-
reinhard , and p. vesely , phys .
c*78 * , 044313 ( 2008 ) .
j. kvasil , v.o .
nesterenko , w. kleinig , d. bok , and p .-
reinhard , int .
e. * 20 * , 281 ( 2011 ) .
a. bohr and b.r .
mottelson , _ nuclear structure _ vol .
2 ( benjamin , new york , 1974 ) .
j. endres et al , phys .
lett . * 105 * , 212503 ( 2010 ) .
v.derya et al , phys .
b*730 * , 288 ( 2014 ) .
n. ryezayeva et al , phys .
lett . * 89 * , 272502 ( 2002 ) .
soloviev , _ theory of atomic nuclei _
, ( oxford , pergamon press , 1976 ) .
m. urban , phys .
c*85 * , 034322 ( 2012 ) .
harakeh and a. van der woude , _ giant resonances _ ( clarendon press , oxford , 2001 ) .
g. holzwarth and g. ekart , z. phys .
a*283 * , 219 ( 1977 ) .
caparelli and e.j.v .
de passos , j. phys .
g : nucl . part
* 25 * , 537 ( 1999 ) .
landau and e.m .
lifshitz , _ course of theoretical physics : hydrodynamics _ vol . 6 , ( butterworth - heinemann , oxford , 1987 ) .
j.m . eisenberg and w. greiner , _ excitation mechanisms of the nucleus : electromagnetic and weak interaction _ , ( north - holland , 1988 ) .
m. danos , nucl .
phys . * 5 * , 23 ( 1958 ) .
s. raman , c.w .
nestor , jr . , and p. tikkanen , at .
data nucl .
data tables * 78 * , 1 ( 2001 ) .
tassie , austr . j. phys .
* 9 * , 407 ( 1956 ) .
j. kvasil , v.o .
nesterenko , w. kleinig , d. bozik , p .-
reinhard , and n. lo iudice , eur .
j. a*49 * , 119 ( 2013 ) .
e. litvinova , p. ring , and v. tselyaev , phys .
c*78 * , 014312 ( 2008 ) .
arsenyev , a.p .
severyukhin , v.v .
voronov , and nguen van giai , epj web of conferences * 38 * , 17002 ( 2012 ) . | we review a recent progress in investigation of the isoscalar toroidal dipole resonance ( tdr ) . a possible relation of the tdr and low - energy dipole strength ( also called a pygmy resonance )
is analyzed .
it is shown that the dipole strength in the pygmy region can by understood as a local manifestation of the collective vortical toroidal motion at the nuclear surface .
application of the tdr as a measure of the nuclear dipole vorticity is discussed .
finally , an anomalous splitting of the tdr in deformed nuclei is scrutinized . |
it is known that most low - mass main - sequence stars are members of binary or multiple systems ( duquennoy & mayor 1991 ; fisher & marcy 1992 ) , and in particular in the solar neighborhood , the fraction goes up to @xmath0 ( abt 1983 ) .
this suggests that binary formation is the primary branch of the star formation process ( mathieu 1994 ) .
significant advances in high - angular - resolution infrared imaging technology have enabled large surveys of young binary stars on a variety of star - forming regions ( mathieu 1992 , 1994 ) .
in addition , right after the discovery of the first extrasolar planetary system around a pulsar ( wolszczan & frail 1992 ) , and particularly after the first extrasolar planet discovered around a main sequence star ( mayor & queloz 1995 ; marcy & butler 1998 ) , observational activity was greatly stimulated .
more recently , advances in observational techniques and instrumentation , such as the hst ( wfpc2 & nicmos ) imaging ( padgett 1997 , 1999 ; reid 2001 ; borucki 2010 ) , submillimeter imaging ( smith 2000 ) , optical and infrared long - baseline interferometry ( quirrenback 2001a , b ) , millimeter and submillimeter interferometry ( launhardt 2000 , launhardt 2001 , guilloteau 2001 ) , adaptive optics ( simon 1999 ; close 2001 ) , spatial astrometry ( sderhjelm 1999 ; quist & lindegren 2000 , 2001 ) , and microlensing ( alcock 2001 ; dong - wook 2008 ; rattenbury 2009 ) , are also available in binary studies .
thanks to all this new technology and observational work , we have now the possibility to study and understand better the physics of binary systems and the surrounding discs built during the formation stage . in the recent past , several planets in binary or multiple star systems have been discovered ( correia 2008 ; deeg 2008 ; desidera & barbieri 2007 ; fischer 2008 ; raghavan 2006 ; konacki 2005 ; mugrauer 2005 ; eggenberger 2004 ; sigurdsson 2003 ; udry 2002 ; sigurdsson & phinney 1993 ; lyne 1988 , etc . ) .
both suspected of formed in _ situ _ , or acquired by dynamical processes ( pfahl & muterspaugh 2006 ) . for a review of observational techniques
see muterspaugh ( 2010 ) .
in addition , several recently discovered circumstellar discs where planets are assumed to be formed , lie around close binaries ( wright 2011 ; prato & weinberger 2010 ; desidera & barbieri 2007 ; doyle 2011 ; queloz 2000 ; hatzes 2003 ) . in these cases
the presence of a companion star should have a very strong influence on both discs and planets
. from the several hundreds of extrasolar planets confirmed so far ( see http://exoplanet.eu , http://planetquest.jpl.nasa.gov , www.exoplanets.org ) to be around main sequence stars , about 10% are known to reside in binary systems with a wide range of orbital separations . in almost all cases ,
the planet orbits in s - type configurations ( dvorak 1986 ) , while the second star acts as a perturber to the planetary system . a circumbinary planet ( p - type orbit )
has recently also been detected in kepler 16 ( doyle et al . 2011 ) .
this has motivated the search for stable periodic orbits around binary systems where planets ( and discs in general ) can settle down in a stable configuration .
most theoretical studies have focused on binaries in near - circular orbits ( hnon 1970 ; lubow & shu 1975 ; paczyski 1977 ; papaloizou & pringle 1977 ; rudak & paczyski 1981 ; bonell & bastien 1992 ; bate 1997 ; bate & bonnell 1997 ) .
even very precise analytical methods to approximate periodic orbits in circular binaries are available ( nagel & pichardo 2008 ) .
due to the lack of conservation of the jacobi integral , the case of eccentric binaries is qualitatively more complicated .
artymowicz & lubow ( 1994 ) and pichardo ( 2005 , 2008 , hereafter psa1 and psa2 ) calculate the extent of zones in phase space available for stable , non self - intersecting orbits around each star and around the whole system . in this study
we use the results of psa1 and psa2 , to calculate stable regions for planets or discs in binary systems in the solar neighborhood with known orbital parameters such as , mass ratio , eccentricity and semimajor axis .
although the existence of stable zones , as the ones we are calculating here , is a necessary , but not a sufficient condition to the existence of planets ( or discs in general ) , if any stable material ( planets , gas , etc . ) , exists in a given system , irrespective of their formation mechanism , they would be necessarily located within the limits of the stable zones . in this direction , a fruitful line of investigation
is the intersection between phase space available zones for the long term evolution of planetary systems and habitable regions allowed by the binary system ( haghighipour 2007 ; haghighipour 2010 ) .
we present a table with the compilation of all the binaries in the solar neighborhood with known orbital parameters from different sources . in the same table
are presented the results for our calculated circumstellar and circumbinary stable zones around each binary system .
this paper is organized as follows . in section [ method ] , we explain briefly the method employed to calculate regions of stable non self - intersecting orbits around binary stars . the binary star sample is presented in section [ sample ] .
results , including stable regions of orbital stability for circumstellar and circumbinary planetary discs ( and discs in general ) , and an application to observations of five binaries with observed planets , are given in section [ results ] . in section [ conclusions ] we present our conclusions .
in the simpler case of circular binary orbits , the potential is time - independent in the co - rotating frame and thus the jacobi energy is conserved .
this allows the existence of closed orbits that , when stable , spawn the orbital structure of the system ( carpintero & aguilar 1998 ) . in the general case of binaries in eccentric orbits ,
the problem is more complex , as now the potential is time - dependent .
however , we can exploit the fact that the time - dependency is strictly periodic .
a time - periodic potential in 2-d can be casted as a 3-d system with an autonomous hamiltonian , with the addition of time and the original hamiltonian as two extra dimensions in phase space .
regular orbits will lie on 3-d manifolds and be multiple periodic with three frequencies , one of which is given by the binary orbital frequency .
if we take snapshots at a fixed binary phase , the projections of a regular orbit will lie on a 2-d manifold .
if the orbit has an additional isolating integral of motion , this projection will now lie on a 1-d manifold : an `` invariant loop '' ( maciejewski & sparke 1997 , 2000 ) .
stable invariant loops represent the generalization to periodically time - varying potentials of stable periodic orbits in steady potentials .
psa1 and psa2 , implemented a numerical method to find them .
the equations of motion are solved in an inertial reference frame with cartesian coordinates with the origin at the binary barycenter .
an ensemble of test particles is launched when the binary star is at periastron , and from the line that joins both stars at that moment , to search for invariant loops .
a more detailed explanation of the method and a study of the phase space in binary systems is in those references . in this work
we employ the formulae from psa1 ( eq . 6 ) and psa2 ( eq .
these relations provide the maximum radius of circumstellar stable zones and the inner radius of the circumbinary stable zone , in terms of the mass ratio ( @xmath1 , where @xmath2 and @xmath3 are the masses of the primary and secondary stars , respectively ) , and the eccentricity ( @xmath4 , where @xmath5 and @xmath6 are the semimajor and semiminor axes of the binary orbit ) .
the radius of the outer limit of the circumstellar stable zones from psa1 , is given by , @xmath7 \ , , \label{eq2.1}\end{aligned}\ ] ] and a similar study in psa2 but for the circumbinary region , gives a relation for the inner radius , @xmath8^{0.043}. \label{eqcb}\end{aligned}\ ] ] in these relations @xmath9 is the approximation of eggleton to the maximum radius of a circle circumscribed within the roche lobe ( eggleton 1983 ) : @xmath10 and @xmath11 in figure [ fig.binscheme ] we present a schematic figure of the geometry of the system .
we show in this figure some of the variables involved in the problem . ) .
the orange and green curves are the primary and secondary stars .
the gray circular areas around each star represent the circumstellar stable zones ( eq . [ eqce ] ) .
the red dot at the center represents the system barycenter .
in the upper part we show a schematic rectangle that we will use in table [ tabla_muestra ] , to show graphically the relative extent and situation of the stable regions , in relation to the stars , for each entry in our sample .
dotted lines join the corresponding parts in the upper and lower representations.,width=279 ] we must emphasize that the regions of stable orbits we report here , are the regions were these invariant loops exist , and furthermore , we restrict ourselves only to non self - intersecting orbits , where gas could settle and planets may form .
it is in this sense that the term `` stable region '' should be understood in this work . using these formulae and restriction ,
we have calculated circumbinary and circumstellar radii for our sample with a total of 161 binary systems in the solar neighborhood with known orbital parameters .
the previous equations require , besides the stellar mass ratio , the semimajor axis and orbital eccentricity for each binary system , two parameters that are difficult to determine observationally .
there is a diversity of methods that have been used to estimate them .
our sample is a compilation from different sources ( jancart 2005 ; martin 1998 ; strigachev & lampens 2004 ; bonavita & desidera 2007 ; holman & wiegert 1999 ; mason 1999 ; latham 2002 ; balera 2006 ; cakirli 2009 ; milone 2005 ; daz 2007 ; konacki 2010 ; desidera & barbieri 2007 ; doyle 2011 ) .
we include all binaries with an estimation of these parameters within 100 pc from the sun ( currently 161 systems ) . in table
[ tabla_muestra ] we present our sample . columns 1 and 2 , are the name of the systems in the hipparchos catalogue and an alternate name , if it exists .
columns 3 to 6 are our input data : the semimajor axis , orbital eccentricity and stellar masses as reported in the references listed in column 10 .
columns 7 to 9 list our results : the circumprimary , circumstellar , and circumbinary boundaries of the stable regions . in the circumstellar cases ,
the value is the radius of the outer boundary . in the circumbinary ,
is the radius of the inner boundary .
all lengths are given in au and masses in solar masses . finally , the last column is a shcematic figure that gives the relative sizes and positions of the stable zones ( see figure [ fig.binscheme ] ) .
for instance , for bd -8@xmath12 4352 ( 9th entry in the table ) the circumstellar regions are symmetric and cover a good fraction of the inner hole of the circumbinary region , this is is due to the low eccentricity of the system and its high @xmath13(=0.5 ) .
in contrast , the binary called gamma vir ( 6th entry in the table ) , present circumstellar regions slightly asymmetric , due to the small mass difference between the stars , and quite narrow , due to its high eccentricity .
the presence of a stellar companion , particularly in an eccentric orbit , severely curtails the size and shape of the stable zones .
while a single star has circular stable orbits at all radii ( neglecting finite stellar size and tidal distortion effects ) , the presence of a stellar companion severely curtails the region where stable , non - self intersecting orbits may exist , both in extent and , to a lesser extent , shape . it is unclear at present the way in which these effects impose restrictions in the process of star formation .
what is clear is that the effect is in the sense of inhibiting , rather than promote it . from the observational side ,
the statistics of the observed systems suggests that binarity does indeed has en effect on planetary masses and orbits ( eggenberger 2004 ) , even restricting terrestrial planet formation for binary periastron smaller than 5 au affecting discs to within @xmath141 au of the primary star ( quintana 2007 , quintana & lissauer 2010 ) .
the goal of this study is to determine the extent of circumstellar and circumbinary regions of stable , non - self intersecting orbits , as plausible regions where planets could have formed and may exist .
it could also indicate possible regions of planetary formation .
figure [ fig.samplesn ] shows the binary semimajor axes _ vs. _ the orbital eccentricity of our entire sample , split in mass ratio intervals .
as it is already known , the region of small semimajor axes and high eccentricites is depleted , i.e. very close binaries , tidally locked in general , have eccentricities close to zero ( duquennoy & mayor 1991 ) . a large percentage ( about @xmath15 ) of binaries in the sample
have low eccentricities ( @xmath16 ) , small semimajor axes ( @xmath17 au ) and large mass ratios ( @xmath18 ) , as shown in the histograms in figure [ fig.histsnobs ] .
consequently ( as seen in figure [ fig.histsnrad ] ) , circumstellar stable regions in this sample have small radii ( @xmath19 au ) , with both stable regions ( circumprimary and circumsecondary ) in most systems having similar size .
on the other hand , the majority of circumbinary gaps have radii smaller that @xmath1450 au .
llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + & hd 10361 & 52.2 & 0.53 & 0.77 & 0.75 & 5.61 & 5.54 & 173.15 & 4 & + & hd 145958a & 124 & 0.39 & 0.9 & 0.89 & 18.17 & 18.08 & 393.95 & 4 & + & hd 145958b & 124 & 0.39 & 0.89 & 0.9 & 18.17 & 18.08 & 393.95 & 4 & + & hd 146362 & 130 & 0.76 & 2.19 & 1.12 & 6.98 & 5.14 & 452.9 & 4 & + & @xmath29 cet & 1.57 & 0.27 & 1.3 & 1.3 & 0.28 & 0.28 & 4.75 & 5 & + + & @xmath30 vir & 37.84 & 0.88 & 0.94 & 0.9 & 0.78 & 0.77 & 135.53 & 5 & + & @xmath31 com & 12.49 & 0.5 & 1.43 & 1.37 & 1.45 & 1.42 & 41.08 & 5 & + & @xmath29 crb & 13.98 & 0.28 & 0.79 & 0.78 & 2.51 & 2.49 & 42.43 & 5 & + & bd -8@xmath12 4352 & 1.35 & 0.05 & 0.42 & 0.42 & 0.33 & 0.33 & 3.41 & 5 & + & bd 45@xmath12 2505 & 4.58 & 0.73 & 0.29 & 0.29 & 0.25 & 0.25 & 15.93 & 5 & + + & @xmath32 equ & 4.73 & 0.42 & 1.66 & 1.59 & 0.66 & 0.64 & 15.18 & 5 & + & kpr 37 & 9.67 & 0.15 & 1.2 & 0.89 & 2.26 & 1.97 & 27.23 & 5 & + & 99 her & 16.39 & 0.74 & 0.89 & 0.52 & 0.98 & 0.77 & 56.97 & 5 & + & 9 pup & 10.00 & 0.69 & 0.98 & 0.87 & 0.67 & 0.63 & 34.49 & 5 & + & @xmath31 cma & 19.89 & 0.59 & 2.11 & 1.04 & 2.12 & 1.54 & 66.70 & 5 & + + & @xmath31 cen & 23.57 & 0.52 & 1.12 & 0.95 & 2.71 & 2.52 & 77.86 & 5 & + & @xmath33 boo & 33.14 & 0.51 & 0.86 & 0.73 & 3.85 & 3.58 & 109.35 & 5 & + & g9 - 42 & 0.44 & 0.81 & 0.77 & 0.04 & 0.02 & 0.01 & 1.45 & 7 & + & g62 - 30 & 0.79 & 0.59 & 0.68 & 0.04 & 0.1 & 0.03 & 2.49 & 7 & + & g165 - 22 & 0.1 & 0.08 & 0.82 & 0.07 & 0.03 & 0.01 & 0.25 & 7 & + + & g65 - 52 & 0.48 & 0.26 & 0.62 & 0.04 & 0.12 & 0.04 & 1.36 & 7 & + & g178 - 27 & 0.33 & 0.43 & 0.68 & 0.05 & 0.06 & 0.02 & 1 & 7 & + & g15 - 6 & 0.73 & 0.45 & 0.67 & 0.06 & 0.13 & 0.04 & 2.25 & 7 & + & g66 - 65 & 0.61 & 0.9 & 0.7 & 0.04 & 0.01 & 0 & 2.05 & 7 & + & g141 - 8 & 0.8 & 0.58 & 0.77 & 0.02 & 0.11 & 0.02 & 2.43 & 7 & + llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + & g18 - 35 & 4.48 & 0.37 & 0.75 & 0.07 & 0.93 & 0.32 & 13.44 & 7 & + & v821 cas & 0.044 & 0.13 & 2.046 & 1.626 & 0.01 & 0.01 & 0.12 & 10 & + & sv cam & 0.016 & 0 & 0.862 & 0.646 & 0.0045 & 0.004 & 0.03 & 11 & + & bs dra & 0.060 & 0 & 1.294 & 1.276 & 0.02 & 0.016 & 0.11 & 11 & + 473 & hd 38 & 73.01 & 0.22 & 0.76 & 0.74 & 14.41 & 14.24 & 215.35 & 3 & + + 1349 & hd 1273 & 1.25 & 0.57 & 0.98 & 0.55 & 0.13 & 0.1 & 4.18 & 1 & + 1955 & hd 2070 & 0.54 & 0.33 & 1.13 & 0.48 & 0.1 & 0.07 & 1.66 & 1 & + 2237 & hd 2475 & 7.08 & 0 & 1.56 & 1.24 & 1.96 & 1.76 & 12.87 & 2 & + 2552 & & 41.32 & 0 & 0.4 & 0.29 & 11.6 & 10.02 & 75.05 & 3 & + 2762 & hd 3196 & 5.13 & 0.77 & 1.71 & 1.14 & 0.25 & 0.21 & 17.96 & 2 & + + 2941 & ads520 & 9.57 & 0.22 & 0.7 & 0.7 & 1.87 & 1.87 & 28.23 & 5 & + 3821 & hd 4614 & 72 & 0.49 & 0.99 & 0.51 & 9.54 & 7.05 & 235.07 & 4 & + 3850 & hd 4747 & 6.7 & 0.64 & 0.82 & 0.04 & 0.73 & 0.19 & 21.21 & 4 & + 5531 & & 5.00 & 0.72 & 1.17 & 1.16 & 0.29 & 0.29 & 17.36 & 8 & + 5842 & hd 7693 & 23.4 & 0.04 & 0.89 & 0.84 & 5.96 & 5.81 & 57.89 & 4 & + + 7078 & hd 9021 & 0.64 & 0.31 & 1.21 & 0.7 & 0.12 & 0.09 & 1.97 & 1 & + 7751 & hd 10360 & 52.2 & 0.53 & 0.77 & 0.75 & 5.61 & 5.54 & 173.15 & 4 & + 7918 & hd 10307 & 7.1 & 0.42 & 0.8 & 0.14 & 1.26 & 0.57 & 22.13 & 2 & + 8903 & hd 11636 & 0.63 & 0.88 & 1.86 & 1.05 & 0.01 & 0.01 & 2.25 & 1 & + 8903 & hd 11636 & 0.66 & 0.9 & 2.07 & 1.28 & 0.01 & 0.01 & 2.37 & 2 & + + 48904 & & 0.027 & 0.01 & 0.365 & 0.348 & 0.007 & 0.006 & 0.06 & 9 & + 9480 & wds
02019 + 7054 & 22.5 & 0.39 & 1.92 & 1.19 & 3.59 & 2.88 & 71.31 & 6 & + * 10138 & * hd 13445 & * 18.4 & * 0.4 & * 0.77 & * 0.49 & * 2.86 & * 2.33 & * 58.53 & * 13 & + 10321 & hd 13507 & 4.3 & 0.14 & 1 & 0.05 & 1.34 & 0.36 & 11.18 & 4 & + 11231 & hd 15064 & 0.64 & 0.29 & 1.01 & 0.68 & 0.12 & 0.1 & 1.95 & 1 & + * * * * * * * * * * llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + 12062 & hd 15862 & 2.04 & 0.26 & 0.95 & 0.44 & 0.43 & 0.3 & 6.11 & 1 & + 12114 & hd 16160 & 15 & 0.75 & 0.76 & 0.09 & 1.01 & 0.39 & 50.26 & 4 & + 12153 & hd 16234 & 4.22 & 0.88 & 11 & 9.41 & 0.09 & 0.08 & 15.11 & 2 & + 12623 & hd 16739 & 1.27 & 0.66 & 1.13 & 1.39 & 0.1 & 0.09 & 4.35 & 2 & + 12777 & hd 16895 & 249.5 & 0.13 & 1.24 & 0.43 & 66.5 & 41.09 & 684.25 & 4 & + + 13769 & hd 18445 & 1.06 & 0.56 & 0.78 & 0.18 & 0.13 & 0.07 & 3.47 & 4 & + * 14954 & * hd 19994 & * 120 & * 0.26 & * 1.35 & * 0.35 & * 27.34 & * 14.83 & * 354.86 & * 13 & + 18512 & hd 24916 & 174.55 & 0 & 0.35 & 0.17 & 52.36 & 37.68 & 315.65 & 3 & + 19206 & & 9.290 & 0.69 & 0.960 & 0.790 & 0.64 & 0.58 & 32.1 & 8 & + 20087 & hd 27176 & 7.05 & 0.17 & 1.76 & 0.95 & 1.66 & 1.26 & 20.08 & 2 & + + 20935 & hd 28394 & 0.99 & 0.24 & 1.13 & 1.11 & 0.19 & 0.19 & 2.95 & 1 & + 22429 & hd 30339 & 0.13 & 0.25 & 1.39 & 0.07 & 0.03 & 0.01 & 0.36 & 4 & + 23395 & wds 05017 + 2640 & 10.28 & 0.33 & 1 & 0.72 & 1.79 & 1.54 & 31.9 & 6 & + 23835 & hd 32923 & 2.86 & 0.9 & 1.11 & 1.03 & 0.05 & 0.05 & 10.28 & 4 & + 24419 & hd 34101 & 1.75 & 0.08 & 0.9 & 0.21 & 0.52 & 0.27 & 4.52 & 1 & + + 25662 & hd 35956 & 2.6 & 0.62 & 0.98 & 0.18 & 0.28 & 0.13 & 8.58 & 4 & + 27913 & hd 39587 & 5.9 & 0.45 & 1.05 & 0.14 & 1.01 & 0.41 & 18.41 & 4 & + 29860 & hd 43587 & 11.6 & 0.8 & 1.06 & 0.34 & 0.54 & 0.32 & 40.39 & 4 & + 33451 & hd 51825 & 9.3 & 0.43 & 1.61 & 1.26 & 1.31 & 1.17 & 29.93 & 2 & + 34164 & hd 53424 & 1.7 & 0.27 & 1.09 & 0.66 & 0.34 & 0.27 & 5.13 & 1 & + + 38657 & hd 64468 & 0.56 & 0.26 & 0.81 & 0.14 & 0.13 & 0.06 & 1.64 & 4 & + 39064 & hd 65430 & 4 & 0.32 & 0.83 & 0.06 & 0.92 & 0.29 & 11.66 & 4 & + 39893 & & 1.81 & 0.21 & 0.95 & 0.52 & 0.4 & 0.31 & 5.29 & 1 & + 44248 & hd 76943 & 10.51 & 0.1 & 1.53 & 0.92 & 2.69 & 2.13 & 28.27 & 2 & + 44892 & hd 78418 & 0.184 & 0.2 & 1.173 & 1.011 & 0.04 & 0.04 & 0.53 & 12 & + *
* * * * * * * * * llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + 45571 & hd 80671 & 4.22 & 0.51 & 3.66 & 3.66 & 0.47 & 0.47 & 13.92 & 2 & + 52940 & hi52940 & 2.6 & 0.37 & 1.12 & 0.12 & 0.53 & 0.2 & 7.84 & 4 & + 54204 & wds 11053 - 2718 & 6.04 & 0.35 & 1.93 & 1.93 & 0.95 & 0.95 & 18.91 & 6 & + 55016 & hd 97907 & 6.87 & 0.42 & 2.62 & 2.32 & 0.97 & 0.92 & 22.05 & 2 & + 56809 & hd 101177 & 240.39 & 0.05 & 1.95 & 1.36 & 63.9 & 54.21 & 605.53 & 3 & + + 60129 & hd 107259 & 10.48 & 0.08 & 6.01 & 0.67 & 3.37 & 1.26 & 26.45 & 2 & + 63406 & hd 112914 & 1.59 & 0.33 & 0.82 & 0.23 & 0.32 & 0.18 & 4.86 & 1 & + 65026 & wds 13198 + 4747 & 14.36 & 0.23 & 0.66 & 0.58 & 2.85 & 2.68 & 42.58 & 6 & + 65343 & hd 116495 & 39.65 & 0.84 & 0.61 & 0.58 & 1.17 & 1.15 & 140.96 & 3 & + 66077 & & 51.51 & 0 & 0.39 & 0.35 & 13.91 & 13.24 & 93.65 & 3 & + + 66492 & & 46.59 & 0.61 & 0.54 & 0.39 & 4.23 & 3.65 & 157.58 & 3 & + 66640 & wds
13396 + 1044 & 10.7 & 0.55 & 1.24 & 1.19 & 1.09 & 1.07 & 35.68 & 6 & + * 67275 & * hd 120136 & * 245 & * 0.91 & * 1.35 & * 0.4 & * 4.38 & * 2.52 & * 868.91 & * 13 & + 67422 & hd 120476 & 33.2 & 0.45 & 0.76 & 0.75 & 4.3 & 4.27 & 107.59 & 3 & + 67422 & hd 120476 & 33.15 & 0.44 & 0.83 & 0.76 & 4.45 & 4.27 & 107.08 & 4 & + + 68682 & hd 122742 & 5.46 & 0.55 & 1.11 & 0.55 & 0.63 & 0.45 & 18.11 & 2 & + 68682 & hd 122742 & 5.3 & 0.48 & 0.92 & 0.54 & 0.7 & 0.55 & 17.28 & 4 & + 69226 & hd 123999 & 0.124 & 0.19 & 1.411 & 1.368 & 0.026 & 0.025 & 0.359 & 12 & + 71094 & & 13.98 & 0.16 & 1.89 & 1.16 & 3.28 & 2.62 & 39.6 & 2 & + 71681 & hd 128621 & 22.76 & 0.51 & 1.12 & 0.89 & 2.67 & 2.4 & 75.04 & 4 & + + 71683 & hd 128620 & 22.76 & 0.51 & 1.12 & 0.89 & 2.67 & 2.4 & 75.04 & 4 & + 71729 & hd 129132 & 8.28 & 0.4 & 3.34 & 3.29 & 1.19 & 1.18 & 26.4 & 2 & + 72659 & hd 131156 & 32.8 & 0.51 & 0.92 & 0.79 & 3.8 & 3.54 & 108.17 & 4 & + 72848 & hd 131511 & 0.53 & 0.51 & 0.79 & 0.45 & 0.07 & 0.05 & 1.74 & 1 & + 72848 & hd 131511 & 0.52 & 0.51 & 0.93 & 0.45 & 0.07 & 0.05 & 1.71 & 4 & + * * * * * * * * * * llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + 73440 & hd 133621 & 1.25 & 0.22 & 1.03 & 0.15 & 0.32 & 0.14 & 3.56 & 1 & + 74392 & wds 15123 + 1947 & 14.93 & 0.25 & 3.54 & 2.51 & 2.98 & 2.55 & 44.69 & 6 & + 75312 & wds 15232 + 3018 & 16.15 & 0.26 & 1.26 & 1.18 & 3.02 & 2.93 & 48.64 & 6 & + 75379 & hd 137502 & 0.91 & 0.68 & 1.26 & 0.68 & 0.07 & 0.05 & 3.12 & 1 & + 76382 & ads9716 & 19.15 & 0.59 & 1.14 & 1.14 & 1.73 & 1.73 & 64.54 & 5 & + + 76852 & hd 140159 & 12.4 & 0.15 & 2 & 1.98 & 2.7 & 2.69 & 34.96 & 2 & + 77152 & hd 140913 & 0.55 & 0.54 & 1.17 & 0.04 & 0.08 & 0.02 & 1.67 & 4 & + 78727 & wds 16044 + 1122 & 15.64 & 0.71 & 0.92 & 0.92 & 0.94 & 0.94 & 54.18 & 6 & + 79101 & hd 145389 & 2.24 & 0.47 & 3.47 & 1.31 & 0.33 & 0.21 & 7.23 & 1 & + 80346 & & 2.07 & 0.67 & 0.5 & 0.13 & 0.18 & 0.1 & 6.98 & 1 & + + 80686 & hd 147584 & 0.12 & 0.06 & 1.05 & 0.37 & 0.04 & 0.02 & 0.3 & 1 & + 81126 & hd 149630 & 6.33 & 0.53 & 3.04 & 1.5 & 0.77 & 0.55 & 20.89 & 2 & + 82817 & hd 152771 & 1.38 & 0.05 & 0.33 & 0.56 & 0.38 & 0.3 & 3.47 & 2 & + 82860 & hd 153597 & 0.33 & 0.21 & 1.18 & 0.52 & 0.08 & 0.05 & 0.96 & 1 & + 83895 &
hd 155763 & 7.09 & 0 & 5.94 & 3.65 & 2.05 & 1.64 & 12.86 & 2 & + + 84140 & hd 155876 & 5.01 & 0.75 & 0.38 & 0.37 & 0.25 & 0.25 & 17.5 & 2 & + 84720 & hd 156274 & 91.65 & 0.78 & 0.79 & 0.47 & 4.33 & 3.41 & 321.18 & 4 & + 84949 & hd 157482 & 4.87 & 0.67 & 1.15 & 2.62 & 0.29 & 0.42 & 16.6 & 2 & + 85141 & hd 157498 & 9.29 & 0.58 & 1.79 & 1.75 & 0.87 & 0.86 & 31.22 & 2 & + 86201 & hd 160922 & 0.082 & 0 & 1.460 & 1.180 & 0.02 & 0.011 & 0.143 & 12 & + + 86221 & wds 17370 + 2753 & 9 & 0.21 & 0.64 & 0.63 & 1.8 & 1.79 & 26.4 & 6 & + 86400 & hd 1360346 & 0.39 & 0.23 & 0.72 & 0.39 & 0.08 & 0.06 & 1.15 & 1 & + 86722 & hd 161198 & 3.97 & 0.94 & 0.94 & 0.34 & 0.04 & 0.03 & 14.21 & 2 & + 86974 & hd 161797 & 22 & 0.32 & 1.15 & 0.13 & 4.92 & 1.85 & 65.17 & 4 & + 87895 & hd 163840 & 2.14 & 0.41 & 0.99 & 0.68 & 0.32 & 0.27 & 6.84 & 1 & + llllllllllc object & alter .
name & @xmath5 & @xmath20 & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 & ref & scheme + hip & & ( au ) & & ( m@xmath26 ) & ( m@xmath26 ) & ( au ) & ( au ) & ( au ) & & @xmath27 @xmath28 + 89937 & hd 170153 & 1.05 & 0.41 & 1.18 & 0.77 & 0.16 & 0.13 & 3.35 & 1 & + 90355 & hd 169822 & 0.84 & 0.48 & 0.91 & 0.3 & 0.12 & 0.07 & 2.71 & 4 & + 91768 & hd 173739 & 49.51 & 0.53 & 0.39 & 0.34 & 5.43 & 5.1 & 164.2 & 3 & + 92418 & hd 174457 & 1.9 & 0.23 & 1.07 & 0.06 & 0.52 & 0.14 & 5.26 & 4 & + 92835 & hp dra & 0.123 & 0.06 & 1.102 & 1.099 & 0.03 & 0.03 & 0.31 & 11 & + + * 93017 & * ads 11871 & * 22.96 & * 0.25 & * 1.65 & * 1.58 & * 4.34 & * 4.26 & * 68.77 & * 5 & + 93506 & wds 19026 + 2953 & 13.36 & 0.2 & 2.97 & 2.4 & 2.81 & 2.55 & 38.93 & 6 & + 93574 & hd 175986 & 9.01 & 0.39 & 1.89 & 1.65 & 1.35 & 1.27 & 28.62 & 2 & + 95028 & hd 181602 & 0.85 & 0.37 & 1.4 & 0.5 & 0.15 & 0.1 & 2.65 & 1 & + 95575 & hd 183255 & 0.62 & 0.15 & 0.78 & 0.38 & 0.15 & 0.11 & 1.74 & 1 & + + 95995 & hd 184467 & 1.45 & 0.37 & 1.22 & 0.46 & 0.26 & 0.17 & 4.53 & 2 & + 96302 & hd 184759 & 4.68 & 0.82 & 3.34 & 1.59 & 0.18 & 0.13 & 16.48 & 2 & + 96471 & hd 184860 & 1.4 & 0.67 & 0.77 & 0.03 & 0.14 & 0.03 & 4.42 & 4 & + * 98001 & * hd 188753 & * 11.65 & * 0.47 & * 1.3 & * 1.11 & * 1.48 & * 1.38 & * 37.98 & * 2 & + 99965 & hd 193216 & 1.24 & 0.08 & 0.88 & 0.56 & 0.32 & 0.26 & 3.26 & 1 & + + 103641 & hd 200077 & 0.587 & 0.66 & 1.186 & 0.941 & 0.04 & 0.04 & 2.01 & 12 & + 104019 & wds 21044 + 1951 & 12.86 & 0.39 & 1.67 & 1 & 2.06 & 1.63 & 40.74 & 6 & + 105969 & hd 204613 & 2.06 & 0.13 & 1.01 & 0.49 & 0.52 & 0.38 & 5.68 & 1 & + 107354 & hd 206901 & 8.24 & 0.31 & 1.56 & 2.6 & 1.53 & 1.21 & 25.32 & 2 & + 108473 & hd 208776 & 4.2 & 0.27 & 1.14 & 0.51 & 0.87 & 0.61 & 12.62 & 4 & + + 109176 & hd 210027 & 0.12 & 0 & 1.25 & 0.8 & 0.03 & 0.03 & 0.22 & 1 & + 110893 & hd 239960 & 9.51 & 0.42 & 0.28 & 0.15 & 1.46 & 1.1 & 30.4 & 3 & + 110893 & ads15972 & 9.53 & 0.41 & 0.27 & 0.17 & 1.51 & 1.20 & 30.41 & 5 & + 111170 & hd 213429 & 1.74 & 0.38 & 1.08 & 0.7 & 0.28 & 0.23 & 5.5 & 1 & + 113718 & hd 217580 & 1.16 & 0.54 & 0.76 & 0.18 & 0.15 & 0.08 & 3.78 & 1 & + * * * * * * * * * * * * * * * * * * * * [ cols="<,<,<,<,<,<,<,<,<,<,^ " , ] \(1 ) jancart 2005 , ( 2 ) martin .
1998 , ( 3 ) strigachev and lampens 2004 , ( 4 ) bonavita and desidera 2007 , ( 5 ) holman and wiegert 1998 , ( 6 ) mason 1999 , ( 7 ) latham 2002 , ( 8) balega 2006 , ( 9 ) diaz 2007 , ( 10 ) cakirli 2009 , ( 11 ) milone 2005 , ( 12 ) konacki 2010 , ( 13 ) desidera and barbieri 2007 , ( 14 ) doyle 2011 , ( 15 ) faigler 2011 , ( 16 ) muterspaugh 2010 the perturbing effect of stellar companions lead to a widespread belief that the presence of planets in binary systems was very unlikely .
now we know that binaries can have planets , and thus should have stable regions around them where planet formation took place .
queloz 2000 , and hatzes 2003 , discovered two giant planets in the binary systems gj 86 and @xmath30cephei .
since then , binaries have become important targets in the search for extrasolar planets , particularily given their abundance . until now about 70 binary systems with planets have been discovered ( wright 2011 ) , but only for eight of them , orbital parameters ( semimajor axis , eccentricity and mass ratio ) are available ( desidera & barbieri 2007 , doyle 2011 , muterspaugh 2010 ) .
table [ tabla_muestra ] shows our prediction for the extent of the stable orbits regions for these systems ( in bold type letter ) . in all cases , where the semimajor axis of the planet is known ,
the observed planet is located within the predicted stable zone .
we present figures with the calculated stable regions constructed with invariant loops , for the five cases where planets are confirmed , and the values for the semimajor axis of observed planets are known . in particular , notice the case of hd 120136 , this is an open binary with a very high eccentricity , usually treated as single star for this reason .
the large eccentricity results in a very narrow , circumstellar stable region .
even so , the discovered planet lies in a p - type orbit inside our predicted circumstellar stable region .
this binary system is located at a distance of 10.9 @xmath34 0.08 pc .
the companion of this object , discovered by els 2001 , has a semimajor axis of 18.4 au , with an eccentricity of 0.4 .
the spectral type of the most massive component is k1 with a mass of 0.77 m@xmath35 and it is a white dwarf ( mugrauer & neuhauser 2005 ) .
the companion has a mass of 0.49 m@xmath35 . at the moment
, one planet was found in this system with a mass of @xmath36 = 4.0 m@xmath37 .
it has a semimajor axis of 0.113 au and an eccentricity of 0.042 ( bonavita & desidera 2007 ) . for this binary the calculated stable zone located around each star is @xmath38 = 3.06 au around the principal component , @xmath39 = 2.49 au around the companion , in both cases
this radii is the outermost radii possible to have stable orbits , and @xmath25 = 58.53 au as the innermost radii available for circumbinary orbits . in figure
[ fig.hd13445 ] , we present the circumprimary , circumsecondary and circumbinary regions of orbital stability , for planets in this case , to settle down , calculated at periastron with our method .
the stellar orbits are marked in green .
the known planet is orbiting the primary star in a very small orbital radius , indistinguishable in this figure .
= 0.77 m@xmath35 , m@xmath40 = 0.49 m@xmath35 , @xmath20 = 0.4 , @xmath5 = 18.4 au , title="fig:",width=279 ] = 0.77 m@xmath35 , m@xmath40 = 0.49 m@xmath35 , @xmath20 = 0.4 , @xmath5 = 18.4 au , title="fig:",width=279 ] this binary system is located at a distance of 22.6 pc .
its semimajor axis is 120 au and it has eccentricity of 0.26 .
the mass of the primary star is 1.35 m@xmath35 with a spectral type f8 v , the mass of the secondary star is 0.35 m@xmath35 .
the planet orbiting this object has a semimajor axis of 1.428 au with an eccentricity of 0.30 and @xmath41= 1.69 m@xmath37 ( desidera & barbieri 2007 ) . in figure [ fig.hd19994 ] , circumstellar , circumbinary stable regions , and planetary orbit ( in red )
are shown .
stellar orbits ( green ) are also indicated .
= 1.35 m@xmath35 , @xmath42 = 0.35 m@xmath35 , @xmath20 = 0.26 , @xmath5 = 120 au , title="fig:",width=279 ] = 1.35 m@xmath35 , @xmath42 = 0.35 m@xmath35 , @xmath20 = 0.26 , @xmath5 = 120 au , title="fig:",width=279 ] this system is located at 15.62 pc , and it has the largest eccentricity for the cases of binary systems with known orbital parameters , which here are semimajor axis 245 au , eccentricity of 0.91 and masses 1.35 m@xmath35 for the primary and 0.4 m@xmath35 for the secondary component .
the planet observed in this system has a semimajor axis of 0.048 au while our approach predicts a maximum radii of 4.86 au , the eccentricity planet is 0.023 and @xmath44 is 4.13 ( desidera & barbieri 2007 ) . in figure [ fig.hd120136 ] , we present the circumprimary , circumsecondary and circumbinary stable regions , for planets in this case , calculated at periastron .
the stellar orbits are marked in green .
the known planet is orbiting the primary star in a very small orbital radius , indistinguishable in this figure .
the upper panel shows the circumstellar regions , primary star to the left secondary to the right .
the bottom panel shows the circumbinary region , notice the change in scale .
the green curves show the stellar orbits , the system is presented at periastron .
orbital parameters : @xmath45 = 1.35 m@xmath35 , @xmath42 = 0.4 m@xmath35 , @xmath20 = 0.91 , @xmath5 = 245 au , title="fig:",width=279 ] boo ) .
the upper panel shows the circumstellar regions , primary star to the left secondary to the right .
the bottom panel shows the circumbinary region , notice the change in scale .
the green curves show the stellar orbits , the system is presented at periastron .
orbital parameters : @xmath45 = 1.35 m@xmath35 , @xmath42 = 0.4 m@xmath35 , @xmath20 = 0.91 , @xmath5 = 245 au , title="fig:",width=279 ] this system is located at a distance of 14.1 pc . the spectral type of the main component is k , and its mass is 1.59m@xmath35 , while the mass of the companion is 0.4 m@xmath35 , the semimajor axis is 18.5 au with an eccentricity of 0.36 .
the planet in this system has a semimajor axis of 2.14 au , with an eccentricity 0.12 and @xmath44 1.77 m@xmath37 ( desidera & barbieri 2007 ) . in figure
[ fig.hd222404 ] , circumstellar stable regions , circumbinary stable region ( gray ) , and planetary orbit ( in red ) are shown .
stellar orbits ( green ) are also indicated .
= 1.59 m@xmath35 , @xmath42 = 0.4 m@xmath35 , @xmath20 = 0.36 , @xmath5 = 18.5 au , title="fig:",width=279 ] = 1.59 m@xmath35 , @xmath42 = 0.4 m@xmath35 , @xmath20 = 0.36 , @xmath5 = 18.5 au , title="fig:",width=279 ] .
[ fig.hd222404 ] recently it was found in this system a planet in circumbinary orbit , this the first observed of this kind ( on the circumbinary disc ) .
the primary star has a mass of 0.69 m@xmath35 , and the companion has a mass of 0.20 m@xmath35 , the semimajor axis is 0.22 au with an eccentricity of 0.16 ( doyle 2011 ) .
the discovered planet has a semimajor axis of 0.71 au , with an eccentricity of 0.0069 and a mass of 0.33 m@xmath37 . in figure
[ fig.kepler16 ] , circumstellar stable regions , circumbinary stable region , and planetary orbit ( in red ) are shown .
stellar orbits ( green ) are also indicated .
= 0.69 m@xmath35 , @xmath42 = 0.20 m@xmath35 , @xmath20 = 0.16 , @xmath5 = 0.22 au , title="fig:",width=275,height=283 ] = 0.69 m@xmath35 , @xmath42 = 0.20 m@xmath35 , @xmath20 = 0.16 , @xmath5 = 0.22 au , title="fig:",width=275,height=283 ]
we have compiled a sample of binary stars with known orbital parameters ( semimajor axes , eccentricities and stellar masses ) of the solar neighborhood and present some basic statistics .
we calculate on this binary stars sample the extent of regions of stable non - self intersecting orbits where planets may exist . for this purpose
, we have applied the concept of `` invariant loops '' and used the formulae of psa1 and psa2 .
our approximation is ballistic , thus , the application is straightforward to debris discs ( planets , cometary nucleii , asteroid belts , etc . ) . in the case of gas discs
, further physics may constraint discs sizes , however , what we are providing here are the regions where the most important orbits of the binary , i.e. , the ones that represent the backbone of the dynamical system ( the ones that are followed for the most of the orbits ) , lay . we have computed the spatial limits of these circumstellar and circumbinary zones for a sample of 161 binaries in the solar neighborhood where orbital data is known and presented it in the form of a table where all the relevant parameters are provided .
we compare our results with observations in the 5 cases where planets have been discovered in binary systems , and where semimajor axis for the planets are provided .
we find that all the planets lay down within our computed regions of stability .
in particular , for hd 120136 , our predicted region of circumstellar stability is very small , and yet the discovered planet lays within it .
although confrontation with a larger database is desirable , the current statistics is fully consistent with our results , proving reliable our approach .
the tool of `` invariant loops '' may be very helpful in the search for planets in binary systems .
we thank our referee , rudolf dvorak for the review of this work that greatly improved it .
we also thank linda sparke and antonio peimbert for enlightening discussions and gustavo arciniega for technical support .
we thank project unam through grant dgapa - papiit in110711 - 2 .
alcock , c. ; allsman , r. a. ; alves , d. r. ; axelrod , t. s. ; becker , a. c. ; bennett , d. p. ; cook , k. h. ; drake , a. j. ; freeman , k. c. ; geha , m. ; griest , k. ; lehner , m. j. ; marshall , s. l. ; minniti , d. ; nelson , c. a. ; peterson , b. a. ; popowski , p. ; pratt , m. r. ; quinn , p. j. ; stubbs , c. w. ; sutherland , w. ; tomaney , a. b. ; vandehei , t. ; welch , d. 2001 , apj , 552 , 259 cieza , lucas a. ; padgett , deborah l. ; allen , lori e. ; mccabe , caer e. ; brooke , timothy y. ; carey , sean j. ; chapman , nicholas l. ; fukagawa , misato ; huard , tracy l. ; noriega - crespo , alberto ; peterson , dawn e. ; rebull , luisa m. 2009 , apj , 696l , 84 doyle , laurance r. ; carter , joshua a. ; fabrycky , daniel c. ; slawson , robert w. ; howell , steve b. ; winn , joshua n. ; orosz , jerome a. ; prsa , andrej ; welsh , william f. ; quinn , samuel n. ; latham , david ; torres , guillermo ; buchhave , lars a. ; marcy , geoffrey w. ; fortney , jonathan j. ; shporer , avi ; ford , eric b. ; lissauer , jack j. ; ragozzine , darin ; rucker , michael ; batalha , natalie ; jenkins , jon m. ; borucki , william j. ; koch , david ; middour , christopher k. ; hall , jennifer r. ; mccauliff , sean ; fanelli , michael n. ; quintana , elisa v. ; holman , matthew j. ; caldwell , douglas a. ; still , martin ; stefanik , robert p. ; brown , warren r. ; esquerdo , gilbert a. ; tang , sumin ; furesz , gabor ; geary , john c. ; berlind , perry ; calkins , michael l. ; short , donald r. ; steffen , jason h. ; sasselov , dimitar ; dunham , edward w. ; cochran , william d. ; boss , alan ; haas , michael r. ; buzasi , derek ; fischer , debra 2011 , sci .
1602 fischer , debra a. ; vogt , steven s. ; marcy , geoffrey w. ; butler , r. paul ; sato , bunei ; henry , gregory w. ; robinson , sarah ; laughlin , gregory ; ida , shigeru ; toyota , eri ; ( and 5 coauthors ) , 2007 , apj , 669 , 1336 launhardt , r. , sargent , a.i . , henning , t. , zylka , r. , zinnecker , h. 2000 . in birth and evolution of binary stars , iau symp .
b. reipurth , h. zinnecker , pp .
san francisco : astron .
soc . pac . | using the results of pichardo ( 2005,2008 ) , we determine regions of dynamical stability where planets ( or discs in general ) could survive in stable orbits around binary stellar systems .
we produce this study for 161 binary stars in the solar neighborhood with known orbital parameters .
additionally , we constructed numerically the discs ( invariant loops ) around five binary systems with known orbital parameters and with confirmed planets : hip 10138 , hip 4954 , hip 67275 , hip 116727 and kepler 16 , as a test to the approximation of pichardo et al .
( 2005,2008 ) .
in each single case , the reported position of the planets lay within our calculated stability regions . this study intends to provide a guide in the search for planets around binary systems with well know orbital parameters , since our method defines precise limits for the stable regions , where discs may have established and planets formed .
[ firstpage ] circumstellar matter , discs binary : stars , solar neighborhood , exoplanets . |
since the close of the past century , an increasing number of observational results have transformed radically our view and understanding of the universe . from these observations a consistent picture has emerged indicating that we live in a _ dark pressure _ ( or _ dark energy _ ) dominated universe , whose the current rate of expansion is an increasing function of time .
if confirmed , this extra component , or rather , its gravitational effects , will be the very first observational piece of evidence for new physics beyond the domain of the standard model of particle physics .
such a expectation has , in turn , given rise to many speculations on the fundamental nature of the dark component which dominates the current dynamics of the universe ( see , e.g , @xcite for a recent review ) . among several alternatives to this dark energy problem , a very interesting one suggests modifications in gravity instead of any adjustment to the energy content of the universe .
examples of modified gravity models include , among others , scenarios with higher order curvature - invariant modifications of the einstein - hilbert action @xcite and braneworld ( bw ) cosmologies @xcite .
while in the first case the _ natural _ matter dilution in the expanding universe is avoided by adding high order terms to the gravitational sector of the theory , in the latter example our 4-dimensional universe is thought of as a surface or a brane embedded into a higher dimensional bulk space - time on which only gravity could propagate .
if such extra dimensions were unambiguosly detected from cosmological observations , bw models could not only give a solution to the so - called dark energy problem but also provide a natural explanation for the huge difference between the two fundamental energy scales in nature , namely , the electroweak and planck scales [ @xmath2 ( see also @xcite ) .
recently , many attempts to observationally detect or distinguish brane effects from the usual dark energy physics have been discussed in the literature . in ref .
@xcite , for instance , sahni and shtanov investigated a class of bw models which admit a wider range of possibilities for the dark pressure than do the usual dark energy scenarios .
as shown by these authors ( also in ref .
@xcite ) , a new and interesting feature of this class of models is that the acceleration of the universe may be a transient phenomena , which can not be achieved in the context of our current standard scenario , i.e. , the @xmath1cdm model but could reconcile the supernova evidence for an accelerating universe with the requirements of string / m - theory @xcite ( see also @xcite for other theoretical and observational aspects of bw models ) .
another particularly interesting bw model is the one proposed by dvali _
@xcite , which is widely refered to as dgp model .
this scenario describes a self - accelerating 5-dimensional bw model with a noncompact , infinite - volume extra dimension in which the dynamics of gravitational interaction is governed by a competition between a 4-dimensional ricci scalar term , induced on the brane , and an ordinary 5-dimensional einstein - hilbert action . for scales below a crossover radius @xmath3 ( where the induced 4-dimensional ricci scalar dominates ) ,
the gravitational force experienced by two punctual sources is the usual 4-dimensional @xmath4 force whereas for distance scales larger than @xmath3 the gravitational force follows the 5-dimensional @xmath5 behavior where @xmath6 is the number of spatial dimensions . ] .
the theoretical consistency of the model , and in particular of its self - accelerating solution , is still a matter of debate in the current literature ( see , e.g. , @xcite ) .
from the observational viewpoint , however , dgp models have been successfully tested in many of their predictions , ranging from local gravity to cosmological observations @xcite ( see also @xcite for a recent review on the dgp phenomenology ) .
in this paper we are particularly interested in testing the viability of dgp scenarios from cosmological time measurements , i.e. , observations of lookback time to galaxy clusters at intermediary and high redshifts and recent estimates of the total age of the universe . in order to build up the lookback time sample we use the age estimates of @xmath7 160 galaxy clusters at six redshifts lying in the interval @xmath8 , as compiled by capozziello _
@xcite , and assume the total expanding age of the universe to be @xmath9 gyr , as obtained by mactavish _
@xcite from an analysis involving the most recent large - scale structure ( lss ) and cosmic microwave background ( cmb ) data .
we also show that the constraints from lookback time observations on the parameters of the model are orthogonal to those obtained from recent measurements of the baryon acoustic oscillations at @xmath0 @xcite , providing very restrictive bounds on dgp cosmologies . in agreement with other independent analyses
, it is shown that spatially closed dgp models are largely favoured by the current data . 0.1 in
in dgp models , the modified friedmann s equation due to the presence of an infinite - volume extra dimension reads @xcite @xmath10^{2 } = h^{2 } + \frac{k}{r(t)^{2}},\ ] ] where @xmath11 and @xmath12 are , respectively , the hubble parameter and the energy density of the cosmic fluid ( which we will assume to be composed only of nonrelativistic particles ) , @xmath13 is the spatial curvature parameter , @xmath14 is the planck mass and @xmath15 ( @xmath16 is the 5-dimensional reduced planck mass ) is the crossover scale defining the gravitational interaction among particles located on the brane . note that for values of @xmath17 , dgp and the standard cold dark matter ( scdm ) models are analogous so that the cosmological evolution for the early stages of the universe is exactly the same in both scenarios . equation ( 1 ) can be rewritten as @xmath18^{2 } , \end{aligned}\ ] ] where @xmath19 ( @xmath20 is the current value of the hubble parameter ) , @xmath21 and @xmath22 stand for the matter and curvature density parameters , respectively , and @xmath23 is the density parameter associated to the crossover radius @xmath3 .
note still that the general normalization condition in dgp scenarios is given by @xmath24^{2 } = 1,\ ] ] while for a flat universe ( @xmath25 ) , it reduces to @xmath26 . as noticed in ref .
@xcite , the above described cosmology can be exactly reproduced by the standard one plus an additional dark energy component with a time - dependent equation of state parameter @xmath27 where @xmath28 and @xmath29 . as shown by several authors @xcite , for values of the length scale @xmath30
( see also table i ) , the presence of an infinite - volume extra dimension as described above leads to a late - time acceleration of the universe , in agreement with most of the current distance - based cosmological observations . in the next section ,
differently from the above analyses , we test the viability of these scenarios from time measurements , i.e. , measurements of lookback time to galaxy clusters and recent estimates of the age of the universe ( see also @xcite for an analysis involving age estimates of high-@xmath31 objects in the context of bw models ) .
in his seminal _ observational tests for world models _ , sandage @xcite defines the lookback time as the difference between the present age of the universe ( @xmath32 ) and its age ( @xmath33 ) when a particular light ray at redshift @xmath31 was emitted . for many years , this interesting relation ( which involves the main cosmological parameters )
was used as a cosmological probe only in a qualitative way due exclusively to the lack of reliable age data . only more recently , with the advent of large telescopes and new technics , it was possible to estimate with reasonable precision ages of high-@xmath31 objects , including galaxies , quasars and galaxy clusters ( see , for instance , @xcite ) . here
, in order to discuss age constraints on the bw scenarios presented above , we use the age estimates of @xmath7 160 galaxy clusters at six redshifts distributed in the interval @xmath8 , as compiled by capozziello _
the clusters at intermediary redshifts ( i.e. , @xmath34 , @xmath35 and @xmath36 ) have their age estimated from the color of their reddest galaxies , whereas for the two at lower redshifts and the high-@xmath31 galaxy clusters at @xmath37 @xcite , it was used the color scattering from a large sample of low redshifts sdss clusters imaged @xcite ( see sec .
iv of @xcite for more details on the age sample ) . the total age of the universe is assumed in our analysis to be @xmath9 gyr , as obtained by mactavish _
@xcite from a joint analysis involving recent lss data ( the matter power spectra from the 2dfgrs and sdss redshift surveys ) and results of the most recent cmb experiments ( wmap , dasi , vsa , acbar , maxima , cbi and boomerang ) . the lookback time - redshift relation , as defined above , is given by @xmath38 where @xmath39 gyr and @xmath40 stands for the density parameters @xmath21 and @xmath41 ( throughout this paper we assume @xmath42 , as provided by the hst key project @xcite ) .
now , following capozziello
_ et al . _
@xcite , let us consider an object ( e.g. , a galaxy , a quasar or a galaxy cluster ) at redshift @xmath43 whose the age @xmath44 is defined as the difference between the age of the universe at @xmath43 and the one when the object was born ( at its formation redshift @xmath45 ) , i.e. , @xmath46\end{aligned}\ ] ] or , equivalently , @xmath47 ( see the schema presented in fig .
1 ) . from the above expressions ( and fig .
1 ) , it is now straightforward to define the observed lookback time to an object at @xmath43 as @xmath48 - [ t^{obs}_o - t_l(z_f ) ] \nonumber \\ & & = t^{obs}_o - t(z_i ) - df , % \nonumber\end{aligned}\ ] ] where @xmath49 stands for the _ incubation time _ or _ delay factor _ , which accounts for our ignorance about the amount of time since the beginning of the structure formation in the universe until the formation time ( @xmath50 ) of the object .
0.1truecm 0.1truecm 0.1 in in order to estimate the best fit to the set of parameters @xmath51 we define the likelihood function @xmath52,\ ] ] with the @xmath53 function given by @xmath54^{2 } } { \sigma_i^{2 } + \sigma_{t^{obs}_o}^{2 } } } + \nonumber \\ & & \quad \quad \quad \quad + \frac{\left[t_o(\mathbf{p } ) - t^{obs}_o\right]^{2}}{\sigma_{t^{obs}_o}^{2}}.\end{aligned}\ ] ] in the above expression , @xmath55 gyr is the uncertainty in the individual lookback time to the i@xmath56 galaxy cluster of our sample while @xmath57 gyr stands for the uncertainty on the total expanding age of the universe ( @xmath58 ) . as discussed in ref .
@xcite , two important aspects concerning the above equations should be emphasized at this point .
first , that we have included a prior on the total age of the universe the second term of eq .
( [ chi2 ] ) .
the reason for this additional term is the well - known fact that the evolution of the age of the universe with redshift ( @xmath59 ) may differ considerably from scenario to scenario , which means that it is possible that cosmological models that are able to explain age estimates of high-@xmath31 objects may not be compatible with the total expanding age at @xmath60 ( and vice - versa )
see @xcite for an example .
the second aspect concerns the delay factor @xmath49 .
note that while the observed lookback time @xmath61 [ eq .
( [ lookobs ] ) ] depends explicitely on @xmath49 its theoretical value @xmath62 [ eq .
( [ looktheo ] ) ] does not .
moreover , it is also worth mentioning that in principle it must be different for each object in the sample . here , however , the delay factor @xmath49 is assumed as a nuisance " parameter , so that we marginalize it over the interval [ 0 , 20 ] gyr . in fig .
2a we show the binned data of the lookback time plotted as a function of redshift for a fixed value of @xmath63 gyr and some selected values of the density parameters @xmath21 and @xmath41 . for the sake of comparison , the einstein - de sitter model and standard @xmath1cdm prediction ( with @xmath64 and @xmath65 )
are also shown .
figure 2b shows the first results of our statistical analysis .
confidence regions ( @xmath66 and @xmath67 ) in the @xmath68 space are displayed by considering the lookback time measurements discussed above .
note that the area corresponding to the confidence intervals is not very restrictive , a fact that is expected and understood in terms of the conservative uncertainty assumed ( @xmath55 gyr ) for the individual lookback time ( see fig .
2a ) gyr is a conservative assumption , in the sense that if the uncertainty on the galaxy cluster age were so large , then there also should be observed a large scatter in the color - magnitude relation for the reddest cluster galaxies , which is not .
however , to take into account other systematic uncertainties as , for instance , those relative to the evolutionary model , we consider that the value assumed , although conservative , is well applicable . ] .
the best - fit parameters for this analysis are @xmath69 and @xmath70 , which corresponds to a spatially closed scenario , in agreement with other independent analyses .
( see , e.g. , @xcite and references therein ) .
if , however , the gaussian prior on the matter density parameter @xmath71 ( as given by wmap results @xcite ) is assumed , the new best - fit model is found at @xmath64 and @xmath72 . by restricting the analysis to the flat case ,
we note that the data favour a lower value of the matter density parameter , i.e. , @xmath73 ( @xmath74 ) at 95.4% ( c.l . ) . also , if we compare the flat cases of dgp and @xmath1cdm scenarios , we find that the former is disfavored , with @xmath75 .
as well known , the acoustic peaks in the cosmic microwave background ( cmb ) anisotropy power spectrum is an efficient way for determining cosmological parameters ( e.g. , @xcite ) . because the acoustic oscillations in the relativistic plasma of the early universe will also be imprinted on to the late - time power spectrum of the non - relativistic matter @xcite , the acoustic signatures in the large - scale clustering of galaxies yield additional tests for cosmology . in particular
, the characteristic and reasonably sharp length scale measured at a wide range of redshifts provides an estimate of the distance - redshift relation , which is a geometric complement to the usual luminosity - distance from type ia supernove @xcite . using a large spectroscopic sample of 46,748 luminous , red galaxies ( lrgs ) covering 3816 square degrees out to a redshift of z=0.47 from the sloan digital sky suvey , eisenstein et al .
@xcite have successfully found the peaks , described by the @xmath76-parameter , which is independent of cosmological models , i.e. , @xmath77^{2/3 } \\
= 0.469 \pm 0.017 , \nonumber\end{aligned}\ ] ] where @xmath78 is the redshift at which the acoustic scale has been measured , @xmath79 is the dimensionless comoving distance to @xmath80 , and the function @xmath81 is defined by one of the following forms : @xmath82 , @xmath83 , and @xmath84 , respectively , for open , flat and closed geometries ( we refer the reader to ref .
@xcite for more details on bao physics ) . the dotted lines in fig .
2b represent the constraints from this measurement on the parametric space @xmath68 ( see also @xcite ) . the important point to be noted in fig .
2b is that lookback time data and the bao prior provide orthogonal statistics in the plane @xmath85 .
this , therefore , suggests that possible degeneracies between these parameters may be broken by combining these two kinds of observations in a joint statistical analysis .
the results of such an analysis are shown in fig . 2c .
note that the parameter space now is considerably reduced relative to the former analysis , with the best - fit model occurring at @xmath86 and @xmath87 .
such a model corresponds to an accelerating universe with @xmath88 , a total expanding age @xmath89 gyr , and a transition redshift of the order of @xmath90 . at 95.4% c.l .
, we also obtain @xmath91 and @xmath92 . from eq .
( [ rc ] ) , we also note that the joint best - fit value for @xmath41 ( @xmath93 ) leads to an estimate of the crossover scale @xmath3 in terms of the hubble radius @xmath94 , i.e. , @xmath95 if now we restrict our analysis to the flat case ( @xmath25 ) , we find @xmath96 or , equivalently , @xmath64 , i.e. , in good agreement with current @xmath21 estimates from cmb data @xcite .
this particular value of @xmath41 corresponds to a crossover distance between 4-dimensional and 5-dimensional gravities of the order of @xmath97 . in table
i we summarize the main estimates for @xmath3 obtained in this paper .
lcrl test & @xmath21 & @xmath41 & @xmath98 + + lookback time & 0.98 & 0.54 & 0.68 + lookback time + @xmath21 & 0.27 & 0.26 & 0.98 + lookback time + bao & 0.25 & 0.25 & 1.0 + + & flat case & & + + lookback time & 0.14 & 0.18 & 1.17 + lookback time + bao & 0.27 & 0.13 & 1.38 + +
alternative cosmologies from brane world model provide a possible mechanism for the present acceleration of the universe congruously suggested by various cosmological observations . in this paper
we have focused our attention on one of these scenarios , the so - called dgp model , in which the acceleration is attributed to gravitational _ leakage _ into extra dimensions .
we have explored the constraints on the parameter space of the dgp scenario from measurements of lookback time to galaxy clusters and the age of the universe .
it is shown that the lookback time observations provide complementary and interesting constraints on the parameters of the model , though there is a degeneracy between @xmath21 and @xmath99 .
when we further combined the recent measurements of the baryon acoustic oscillations at @xmath0 found in the sloan digital sky survey data , a very stringent constraint on both @xmath21 and @xmath99 is obtained .
the resulting parameters range as @xmath91 and @xmath92 at 95.4% confidence level , indicating a closed universe , which is consistent with other independent analyses .
if , however , a flat universe is considered a priori ( as usually done in the context of dark energy models ) , we find @xmath96 or , equivalently , @xmath64 , which is in good agreement with current estimates of the matter density parameter from cmb data @xcite . finally , it is also important to emphasize that although the current lookback time data do not provide very restrictive bounds on the density parameters @xmath100 and @xmath101 of the dgp model , the method discussed in sec .
iii ( and also in refs .
@xcite ) , along with new and more precise age measurements of high-@xmath31 objects will certainly provide a new and complemetary tool to test the reality of the current cosmic acceleration as well as to distingush among the many alternative world models .
the present work , therefore , highlights the cosmological interest in the observational search for old collapsed objects at low , intermediary and high redshifts .
the authors are very grateful to v.f .
cardone , m. fairbairn , a. goobar and g.s .
frana for valuable discussions .
np is supported by pronex / cnpq / fapern .
z - hz is supported by the national natural science foundation of china , under grant no .
10533010 , and by srf , rocs , sem of china .
jsa is supported by cnpq ( brazilian agency ) under grants no .
307860/2004 - 3 and 475835/2004 - 2 and by fundao de amparo pesquisa do estado do rio de janeiro ( faperj ) no .
e-26/171.251/2004 .
v. sahni and a. a. starobinsky , int . j. mod
d9 , 373 ( 2000 ) ; p. j. e. peebles and b. ratra rev .
phys . * 75 * , 559 ( 2003 ) ; t. padmanabhan , phys . rept . * 380 * , 235 ( 2003 ) .
hep - th/0212290 ; j. a. s. lima , braz .
* 34 * , 194 ( 2004 ) .
astro - ph/0402109 ; e. j. copeland , m. sami and s. tsujikawa , hep - th/0603057 . n. arkani - hamed , s. dimopoulos , g. r. dvali , phys
b429 * , 263 ( 1998 ) ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. dvali , phys . lett . *
b436 * 257 ( 1998 ) ; l. randall , science * 296 * , 1422 ( 2002 ) .
t. shiromizu , k. maeda , and m. sasaki , phys .
d 62 , 024012 ( 2000 ) ; r. dick , class . quantum grav .
18 , r1 ( 2001 ) ; c. j. hogan , class .
. grav . * 18 * , 4039 ( 2001 ) ; c. j. hogan , astro - ph/0104105 ; k. ichiki , m. yahiro , t. kajino , m. orito and g. j. mathews , phys . rev . *
d66 * , 043521 ( 2002 ) ; k. freese and m. lewis , phys .
b540 * , 1 ( 2002 ) ; z. -h .
zhu and m. -k .
fujimoto , , * 581 * , 1 ( 2002 ) ; d. langlois , astrophys .
space sci . * 283 * , 469 ( 2003 ) ; a.a .
sen and s. sen , , 1 ( 2003 ) ; y. wang , k. freese , p. gondolo , and m. lewis , , 25 ( 2003 ) ; z. -h .
zhu and m. -k .
fujimoto , , 52 ( 2003 ) ; m. d. maia , gr - qc/0401032 ; y. gong and c .- k .
duan , mnras * 352 * , 847 ( 2004 ) ; m. d. maia , e. m. monte and j.m.f .
maia , phys .
* b585 * , 11 ( 2004 ) ; z. -h . zhu and m. -k .
fujimoto , , 12 ( 2004 ) ; m. d. maia , e. m. monte , j. m. f. maia and j. s. alcaniz , class .
* 22 * , 1623 ( 2005 ) .
astro - ph/0403072 ; j. s. alcaniz , a. dev and d. jain , , 26 ( 2005 ) ; koivisto , h. kurki - suonio and f. ravndal , , 064027 ( 2005 ) . c. deffayet , g. dvali and g. gabadadze , phys
* d65 * , 044023 ( 2002 ) ; p. p. avelino and c. j. a. p. martins , apj * 565 * , 661 ( 2002 ) ; j. s. alcaniz , phys .
d * 65 * , 123514 ( 2002 ) .
astro - ph/0202492 ; a. lue , phys . rev . * d67 * , 064004 ( 2003 ) ; a. lue and g. d. starkman , phys . rev . * d67 * , 064002 ( 2003 ) ; a. lue , r. scoccimarro and g. d. starkman , phys . rev . * d69 * , 124015 ; e. v. linder , phys . rev . lett . * 90 * , 091301 ( 2003 ) j. s. dunlop et al .
, nature * 381 * , 581 ( 1996 ) ; h. spinrad et al .
, 581 ( 1997 ) ; j. a. peacock et al . , mnras * 296 * , 1089 ( 1998 ) ; r. jimenez et al . , mnras * 305 * , l16 ( 1999 ) ; j.s . alcaniz and j.a.s .
lima , , l87 ( 1999 ) .
astro - ph/9902298 ; r. jimenez and y. sukyoung , , 670 ( 2000 ) ; j.a.s .
lima and j. s. alcaniz , mon . not .
. soc . * 317 * , 893 ( 2000 ) .
astro - ph/0005441 ; l. a. nolan , j. s. dunlop , r. jimenez , mon . not .
roy . astron . soc . *
323 * , 385 ( 2001 ) .
j. s. alcaniz and j.a.s .
lima , , l133 ( 2001 ) .
astro - ph/0101544 ; g. hasinger , n. schartel and s. komossa , , l77 ( 2002 ) ; r. jimenez and a. loeb , , * 573 * , 37 ( 2002 ) ; j. s. alcaniz , j. a. s. lima and j. v. cunha , mon . not .
. soc . * 340 * , l39 ( 2003 ) .
astro - ph/0301226 ; a. cimatti et al , nature , * 430 * , 184 ( 2004 ) . | the observed late - time acceleration of the universe may be the result of unknown physical processes involving either modifications of gravitation theory or the existence of new fields in high energy physics . in the former case ,
such modifications are usually related to the possible existence of extra dimensions ( which is also required by unification theories ) , giving rise to the so - called brane cosmology . in this paper
we investigate the viability of this idea by considering a particular class of brane scenarios in which a large scale modification of gravity arises due to a gravitational _ leakage _ into extra dimensions . to this end , differently from other recent analyses , we combine _ orthogonal _ age and distance measurements at intermediary and high redshifts .
we use observations of the lookback time to galaxy clusters , indirect estimates of the age of the universe from the most recent large - scale structure ( lss ) and cosmic microwave background ( cmb ) data , along with the recent detection of the baryon acoustic oscillations at @xmath0 . in agreement with other recent analyses
we show that , although compatible with these age and distance measurements , a spatially closed scenario is largely favoured by the current observational data . by restricting our analysis to a spatially flat universe
, we also find that the standard @xmath1cdm model is favoured over the particular braneworld scenario here investigated . |
we learn much of our information about astrophysical objects by observing the light they emit .
observations of the polarization properties of light can tell us the geometry of the emitter , strength of the magnetic field , density of plasma , and temperature .
the proper and correct theory of optical activity is essential for making accurate predictions . while the low - temperature propagation characteristics of plasma are well - established @xcite , the theory of relativistic effects has not been fully studied . in this paper
i discuss the propagation effects through a homogeneous magnetized relativistic plasma .
a non - magnetized case emerges as a limit of the magnetized case .
the discussion is divided into three separate topics .
two linear plasma propagation effects are faraday rotation and faraday conversion @xcite . traditionally , these effects are considered in their lowest orders in the ratio @xmath4 of the cyclotron frequency @xmath1 to the circular frequency of light @xmath5 i d est in a high - frequency approximation . the distribution of particles is taken to be thermal @xmath6 with the dimensionless temperature @xmath7 in the units of particle rest mass temperature @xmath8 the faraday rotation measure @xmath9 and conversion measure are known in a non - relativistic @xmath10 and an ultra - relativistic @xmath11 limits @xcite .
i derive a surprisingly simple analytic expression for arbitrary temperature @xmath12 the smallness of @xmath13 @xmath14 in the real systems led some authors @xcite to conclude that the high - frequency approximation will always work .
however , there is a clear indication that it breaks down at high temperatures @xmath15 it was claimed that the eigenmodes of plasma are linearly polarized for high temperatures @xmath11 @xcite , because the second order term @xmath16 becomes larger than the first order term @xmath17 due to the @xmath7 dependence .
the arbitrarily large @xmath7-factor may stand in front of higher order expansion terms in @xmath4 of the relevant expressions .
i find the generalized rotation measure as a function of @xmath4 and @xmath7 without expanding in @xmath4 and compare the results with the known high - frequency expressions .
the high-@xmath7 behavior of the plasma response is indeed significantly different .
plasma physics involves complicated calculations .
this led to a number of errors in the literature @xcite , some of which have still not been fixed . in the article
i check all the limiting cases numerically and analytically and expound all the steps of derivations .
thus i correct the relevant errors and misinterpretations made by previous authors , hopefully not making new mistakes .
the analytical and numerical results are obtained in mathematica 6 system .
it has an enormous potential in these problems @xcite .
the paper is organized as follows .
the formalism of plasma response and calculations are described in
[ method ] .
several applications to observations can be found in [ applications ] .
i conclude in [ discussion ] with a short summary and future prospects .
i assume the traditional geometry depicted on figure [ fig_rot ] : * euclidean basis ( @xmath18 @xmath19 @xmath20 ) , * magnetic field along the third axis @xmath21 * a wave vector of the wave @xmath22 with an angle @xmath23 between @xmath24 and @xmath25 the basis is rotated from ( @xmath18 @xmath19 @xmath20 ) to ( @xmath26 @xmath27 @xmath28 ) , so that the wave propagates along @xmath29 in the new basis .
the transformation has the form @xmath30 which can be conveniently written as @xmath31 vectors and tensors then rotate according to @xmath32 the propagation of weak electromagnetic ( em ) waves in a homogeneous magnetized plasma can be fully described by the response tensor @xmath33 it expresses the linear proportionality between the induced current density and the vector potential @xmath34 the spatial projection of such defined 4-d tensor @xmath35 is equal to the 3-d tensor @xmath36 defined by @xmath37 i consider trubnikov s form of the response tensor @xcite .
i work in a low - density regime , where the plasma response is calculated for a vacuum wave with @xmath38 i take the tensor @xmath39 from the first - hand derivations @xcite , make the transformation ( [ transform ] ) , and take the 1-st and 2-nd components in both indices .
thus the projection onto the ( @xmath40 ) plane in cgs units is @xmath41,\ ] ] @xmath42 @xmath43 @xmath44 and @xmath45^{1/2},\ ] ] where @xmath46 is the sign of the charge , @xmath47 is the n - th bessel function of the second kind in the component @xmath48 and the opposite sign of @xmath49 term by an error . the author has corrected his formulas in @xcite . ] .
the quantity @xmath50 is the dimensionless inverse temperature , @xmath51 where @xmath52 the actual temperature of particles .
the response of plasma is usually characterized by the dielectric tensor .
its projection onto the ( @xmath53 ) plane is @xmath54 the wave equation for transverse waves in terms of @xmath55 is @xmath56 where @xmath57 and @xmath58 are the components of the electric field along @xmath59 and @xmath60 and @xmath61 @xcite .
let me first calculate the limiting expression for @xmath35 in the high - frequency limit @xmath62 i denote @xmath63 substitute the definitions ( [ podst ] ) into the expression ( [ main ] ) , and expand the response tensor @xmath35 in @xmath4 .
i retain only up to the 2-nd order of the expansion , which gives the conventional generalized faraday rotation @xcite
. the first terms of the series of @xmath64 @xmath65 and @xmath66 read @xmath67 @xmath68 @xmath69 @xcite used the approximation @xmath70 instead of the expansion ( [ rad ] ) and obtained the approximate high - t expressions as his final answers .
however , one can take the emergent integrals , if one considers the exact expansions ( [ rad],[exp_t],[exp_r ] ) .
three terms appear in the expanded expression for @xmath71 [ calc ] @xmath72,\ ] ] @xmath73,\ ] ] @xmath74.\ ] ] the 2-nd term ( [ calcb ] ) originates from the expansion of @xmath75 in @xmath76 to the first order @xmath77 integrals ( [ calca],[calcb],[calcc ] ) can be evaluated knowing that [ ident ] @xmath78=n!i^{n+1}\frac{k_{n-1}(\rho)}{\rho^2},\ ] ] @xmath79=n!i^{n+1}\frac{k_{n-2}(\rho)}{\rho^3}.\ ] ] i substitute the high - frequency expansions ( [ rad],[exp_t],[exp_r ] ) into the expression ( [ dielectric ] ) for the projection of the dielectric tensor @xmath80 with the projection of the response tensor @xmath35 ( [ main ] ) and take the integrals ( [ calca],[calcb],[calcc ] ) analytically .
the components of the dielectric tensor ( [ dielectric ] ) in the lowest orders in @xmath3 are then [ diag ] @xmath81 @xmath82 @xmath83 where the plasma frequency @xmath84 in cgs units is @xmath85 the results reproduce the non - relativistic limits for @xmath86 [ nonrel ] @xmath87 @xmath88 @xmath89 where all bessel functions of @xmath50 approach unity @xcite . the corresponding relativistic limits @xmath90 of the same components
are [ rel ] @xmath91 @xmath92 @xmath93 consistent with @xcite . the ultra - relativistic non - magnetized dispersion relation then reads @xmath94 according to the relation ( [ dispersion ] ) . the expression ( [ disp_nonmagn ] ) is consistent with @xcite , chapter 32 .
the plasma propagation effects can usually be described in terms of only the difference of the diagonal components and the non - diagonal component of @xmath95 i define @xmath96 to be a vector of @xmath97 @xmath98 @xmath99 i introduce the multipliers @xmath100 and @xmath101 to correct the expressions , when the high - frequency limit breaks .
i write the difference between the diagonal components with a multiplier @xmath100 as @xmath102 and the non - diagonal component with a multiplier @xmath101 as @xmath103 both multipliers equal unity in the high - frequency limit @xmath104 now we can turn to a more general case .
the ultra - relativistic expressions ( [ rela],[relb],[relc ] ) allow me to trace the t - factors in front of the first 3 expansion coefficients of the dielectric tensor in @xmath105 the coefficient at @xmath106 is @xmath107 times larger than at @xmath105 thus at temperature @xmath108 the 2-nd order becomes larger than the 1-st order for the ratio @xmath109 this indicates that the expansion in @xmath4 may become invalid at these plasma parameters .
the multipliers @xmath100 and @xmath101 are likely to be far from @xmath110 i consider only the real parts of these multipliers , since the imaginary parts correspond to absorption .
the contour plots of the numerically calculated @xmath100 and @xmath111 for somewhat arbitrary @xmath112 are shown on figure [ fig_diag_contour ] and figure [ fig_nondiag_contour ] , respectively .
let me define @xmath113 to be the following combination of the parameters @xmath114 for the fiducial @xmath115 @xmath112 the parameter @xmath113 is just temperature @xmath116 i first identify the boundaries , where the high - frequency limit is valid .
then i find a fit for the multipliers at higher @xmath117 the expression ( [ diagdifffit ] ) for the difference @xmath118 is accurate within @xmath119 for @xmath120 if we set @xmath121 the expression ( [ nondiagfit ] ) for @xmath122 is accurate within @xmath119 for @xmath123 if we set @xmath124 the accuracy depends on the parameter @xmath113 rather than on the individual parameters @xmath97
@xmath125 @xmath126 the expression @xmath127 extends the applicability domain of the formula ( [ diagdifffit ] ) up to @xmath128 figure [ fig_diag_fit ] shows the fit for @xmath129 in comparison with the numerical results .
the expression @xmath130 extends up to @xmath131 the domain of the formula ( [ nondiagfit ] ) .
figure [ fig_nondiag_fit ] shows the fit for @xmath132 in comparison with the numerical results .
the expression for the response tensor ( [ main ] ) is written for a vacuum wave with @xmath133 in the real plasma , the wave is modified by the plasma response .
a more general self - consistent response tensor should be used @xcite .
one needs to solve a dispersion relation similar to the relation ( [ dispersion ] ) to obtain the eigenmodes .
thus the eigenmodes and the response tensor should be computed self - consistently .
one should not forget about the antihermitian and longitudinal components of the dielectric tensor @xmath80 that modify the dispersion relation .
the above calculation is applicable also to a non - magnetized plasma .
dispersion relation of em waves in a non - magnetized plasma reads @xmath134 in a high - frequency approximation @xmath135 .
the opposite limit of @xmath136 was considered by @xcite .
now we turn to the magnetized case .
@xcite only considered the first terms of in the expansion of @xmath35 in @xmath4 to get the eigenmodes .
i do the next step : consider the full expression in @xmath4 in the low - density regime @xmath137 but consider only the hermitian part of @xmath35 in computations .
the ellipticity @xmath138 determines the type of eigenmodes .
if @xmath139 then the eigenmodes are linearly polarized unless @xmath23 is close to @xmath140 if @xmath141 then the eigenmodes are circularly polarized for @xmath23 far from @xmath142 let me consider the fiducial model with @xmath143 and @xmath144 figure [ fig_ratio_fit ] shows the ratio @xmath145 calculated in a high - frequency approximation ( see [ high_freq ] ) ( dashed line ) and in a general low - density approximation ( see [ low_dens ] ) ( solid line ) .
the high - frequency approximation produces the linear eigenmodes already at @xmath146 consistently with @xcite .
however , the general low - density limit produces the eigenmodes with @xmath147 up to very high temperatures @xmath148 unexpectedly , the sign of the diagonal difference @xmath149 changes at about @xmath150
the calculated transrelativistic propagation effects have far - reaching consequences in many topics of astronomy .
let me concentrate on four applications : propagation delay , faraday rotation measure of light from the galactic center ( gc ) , circularly polarized light from the gc , diagnostics of jets .
propagation delay is an important effect in pulsar dispersion @xcite .
the relativistic part of this delay can be obtained from the dispersion relation ( [ nonmagn ] ) .
i retain only the first - order correction in @xmath97 since @xmath10 in the interstellar medium @xcite .
since @xmath151 at low @xmath97 the non - relativistic dispersion measure ( dm ) should be modified as @xmath152 this shows that the gas density is slightly underestimated , if the non - relativistic formulas are used .
however , the relativistic correction to the dm is small and can be neglected in most practical cases when @xmath153 the effects in magnetized plasma are also relevant for pulsars .
relativistic plasmas exhibit a generalized faraday rotation for a general orientation of the magnetic field @xcite .
one can decompose it into two effects : faraday rotation and faraday conversion .
the former operates alone at @xmath154 the latter operates alone at @xmath155 and both should be considered together for the intermediate angles .
the transfer equations ( mueller calculus ) for the stokes parameters @xmath156 @xmath157 @xmath158 @xmath159 were devised to treat together the propagation effects , emission , and absorption @xcite .
good approximations for emission and absorption have been long known @xcite .
now one can combine them with the proper approximations of the propagation effects given by @xmath160 @xmath161 and do the radiative transfer calculations . here
@xmath80 stands for the hermitean part given by the relations ( [ gaunt_diag],[gaunt_nondiag ] ) with the real multipliers @xmath129 and @xmath162 one of the most interesting objects for such calculations is our galactic center sgr a*. the transfer equations were recently solved for a simple time - independent dynamical model of the gc accretion @xcite .
the authors treat the ordinary and extraordinary modes as linearly polarized .
they assume these eigenmodes constitute a basis , where either @xmath163 or @xmath164 components of emissivity and propagation coefficients vanish .
actually , @xmath163 components vanish ( @xmath165 ) already in the basis @xmath166 since the projection of the magnetic field onto @xmath167 is parallel to @xmath59 ( see @xcite p.184 ) . as i have shown in the
[ modes ] , plasma modes are far from being linearly polarized at temperatures @xmath168 estimated for the gc @xcite .
thus , the propagation coefficients should be taken from equations ( [ diagdifffit ] ) and ( [ nondiagfit ] ) .
the faraday conversion coefficient @xmath169 can not be defined via emissivities and faraday rotation coefficient @xmath170 as in @xcite .
the faraday rotation measure was calculated from a simulated accretion profile in @xcite .
however , the paper considered only the faraday rotation and did not carry out the self - consistent treatment of propagation .
it is impossible to disentangle the effects of faraday rotation and faraday conversion in a relativistic plasma .
the crucial part of any radiative transfer is the proper transfer coefficients .
it allows one to estimate the electron density near the accreting object @xcite .
several formulas were suggested for the temperature dependence of the component @xmath122 responsible for faraday rotation .
these formulas were yet given for the high - frequency approximation ( see [ high_freq ] ) .
let me compare them with the exact temperature dependence ( [ nondiag ] ) @xmath171 and its limits .
the limits are @xmath172 as @xmath173 and @xmath174 as @xmath175 the results of this comparison are shown on figure [ fig_mildly_bridge ] .
@xcite @xcite with the 3-d response tensor @xmath176 that has the opposite sign .
] divided the thermal distribution into ultra - relativistic and non - relativistic parts as marked by the electron energy @xmath177 they sum the contributions of both species with calculated densities . to make a plot , i take their effective temperature @xmath178 of plasma above @xmath179 to be just temperature @xmath180 and not the average kinetic energy as @xcite suggest .
this brings @xmath178 to lower values and decreases the rotation measure .
even with this decrease the rotation measure is severely overestimated at @xmath181 the convergence to the relativistic limit is not achieved even at @xmath182 the paper @xcite found the simpler fitting formula that reproduces the limits .
their expression is quite accurate . in @xcite
should be redefined as @xmath183 , otherwise the lower limit is not reproduced . ]
the increase in the circular polarization of sgr a * at frequency 1thz is predicted by @xcite . the phase of faraday conversion approaches unity and the destructive interference does not occur at this frequency .
the result seems to be qualitatively correct regardless of the expression for the conversion measure , but the proper expressions ( [ diagdifffit ] ) and ( [ nondiagfit ] ) should be used for quantitative predictions . the better treatment of propagation effects may also play a role in observations of jets . as we saw in
[ low_dens ] , the propagation effects in thermal plasma can not be described in the lowest orders in @xmath125 if the temperature @xmath7 is sufficiently high .
power - law distribution of electrons can have a quite high effective temperature .
thus the high - frequency limit @xcite may not approximate well the hermitian part of the response tensor .
careful analysis of jet observations @xcite may be needed .
it should be based at least on the expressions for @xmath80 in a general low - density regime .
this paper presents several new calculations and amends the previous calculations of propagation effects in uniform magnetized plasma with thermal particle distribution equation ( [ distrib ] ) . the expression ( [ main ] ) for the correct response tensor
is given in a high - frequency approximation .
the exact temperature dependence ( [ diag ] ) and ( [ nondiag ] ) is found in first orders in @xmath3 in addition to the known highly - relativistic and non - relativistic results .
the higher order terms may be important for relativistic plasmas in jets and hot accretion flows .
the fitting expressions ( [ gaunt_diag ] ) and ( [ gaunt_nondiag ] ) are found for the dielectric tensor components ( [ diagdifffit ] ) and ( [ nondiagfit ] ) at relatively high temperatures . the results of numerical computations are given only when the corresponding analytical formulas are found .
one can always compute the needed coefficients numerically for every particular frequency @xmath2 , plasma frequency @xmath184 cyclotron frequency @xmath185 and distribution of electrons .
however , the analytic formulas offer a simpler and faster way of dealing with the radiative transfer for a non - specialist .
the eigenmodes were not considered in much detail , since radiative transfer problems do not require a knowledge of eigenmodes . however
the knowledge of eigenmodes is needed to compute the self - consistent response tensor ( see [ exact ] ) .
the response tensor in the form ( [ main ] ) can be expanded in @xmath3 and @xmath186 this expansion is of mathematical interest and will be presented in a subsequent paper as well as the expressions for a power - law electron distribution .
propagation through non - magnetized plasmas will also be considered separately .
100 azzam , r. m. a. , & bashara , n. m. 1987 , `` ellipsometry and polarized light '' ( amsterdam : north holland ) ballantyne , d. r. , ozel , f. , psaltis , d. 2007 , apj , 663l , 17 beckert , t. , falcke , h. 2002 , a&a , 388 , 1106 bellan , p. m. 2006 , `` fundamentals of plasma physics '' , ( cambridge : cambridge university press ) bergman j. , & eliasson b. 2001 , phpl , 8 , 1482 cox , d. p. , & reynolds , r. j. 1987 , ann .
a&a , 25 , 303 huang , l. , liu , s. , shen , z .- q .
, cai , m. j. , li , h. , & fryer , c. l. 2008 , apj , 676l , 119 jones , t. w. , odell , s. l. 1977 , apj , 214 , 522 landau , l. d. , lifshits , e. m. 1980 , `` physical kinetics '' ( oxford : pergamon press ) marichev , o. , 2008 , wolfram blog from may 6 http://blog.wolfram.com/2008/05/06/two-hundred-thousand-new-formulas-on-the-web/ melrose , d. b. , mcphedran , r. c. 1991 , `` electromagnetic processes in dispersive media '' , ( cambridge university press : cambridge ) melrose , d. b. 1997 , j. plasma physics , 57 , 479 melrose , d. b. 1997 , phys . rev .
e , 56 , 3527 melrose , d. b. 1997 , j. plasma physics , 58 , 735 melrose , d. b. 2008 , `` quantum plasmdynamics : magnetized plasmas '' , in preparation phillips , j. a. , & wolszczan , a. 1992 , apj , 385 , 273 rybicki , g. b. , & lightman , a. p. 1979 , `` radiative processes in astrophysics '' , ( wiley - interscience : new york ) quataert , e. , & gruzinov , a. 2000 , , 545 , 842 sazonov , v. n. 1969 , soviet astron .
, 13 , 396 sharma , p. , quataert , e. , hammett , g. w. , & stone , j. m. 2007 , , 667 , 714 sharma , p. , quataert , e. , & stone , j. m. 2007 shcherbakov , r. v. 2008 , apj , accepted swanson , d.g . 2003 , `` plasma waves '' ( bristol : institute of physics pub . ) trubnikov , b. a. 1958 , `` magnetic emission of high temperature plasma '' .
thesis , moscow institute of engineering and physics [ english translation in aec - tr-4073 , us atomic energy commission , oak ridge , tennessee , 1960 ] .
trubnikov , b. a. 1996 , `` theory of plasma '' , ( moscow : energy atom press ) ( in russian ) wardle , j. f. c. , homan , d. c. , ojha , r. , & roberts , d. h. 1998 , nature , 395 , 457 wolfe , b. , & melia , f. , 2006 , apj , 637 , 313 | the transfer of polarized radiation in magnetized and non - magnetized relativistic plasmas is an area of research with numerous flaws and gaps .
the present paper is aimed at filling some gaps and eliminating the flaws .
starting from a trubnikov s linear response tensor for a vacuum wave with @xmath0 in thermal plasma , the analytic expression for the dielectric tensor is found in the limit of high frequencies .
the faraday rotation and faraday conversion measures are computed in their first orders in the ratio of the cyclotron frequency @xmath1 to the observed frequency @xmath2 .
the computed temperature dependencies of propagation effects bridge the known non - relativistic and ultra - relativistic limiting formulas .
the fitting expressions are found for high temperatures , where the higher orders in @xmath3 can not be neglected .
the plasma eigenmodes are found to become linearly polarized at much larger temperatures than thought before .
the results are applied to the diagnostics of the hot ism , hot accretion flows , and jets . |
the incomparably intensive studies of graphene @xcite paved the way towards applications of this unique material in spintronics @xcite .
progress in this field involves design of novel graphene - based spintronic devices @xcite .
however , applicational potential of graphene in spin electronics is closely connected with progress of knowledge about its magnetic properties @xcite . in this context ,
graphene nanostructures ( quantum dots , nanoflakes ) attract special attention @xcite .
this is due to the fact that geometric confinement and particular form of edge can cause emergence of novel phenomena and alter substantially the magnetic characteristics predicted for infinite graphene layers ( e.g. @xcite ) .
one of the crucial factors serving the development of graphene - based spintronics is the ability to control its magnetic properties .
various approaches to manipulation of magnetism were considered in the literature , including optical methods @xcite and influence of the external magnetic field @xcite or hydrogenation @xcite . however , a fundamentally important issue in spintronic applications is controlling the magnetic properties by means of external electric field .
let us mention that the influence of the electric field on the magnetic properties has been noticed experimentally in some systems useful for spintronics like , for example , metallic @xcite and semiconducting thin films and other nanostructures @xcite .
it has also been a subject of theoretical studies , among which we put particular emphasis on those devoted to modifying an indirect ruderman - kittel - kasuya - yosida interaction in the presence of electric field @xcite .
owing to highly unusual electronic structure , graphene and its nanostructures offer new possibilities of using the electric field to influence the energy levels available for the charge carriers and , in particular , to control the resulting magnetic properties ( see the review ref . as well as refs .
) . motivated by this possibilities
, we perform a theoretical study of a graphene nanostructure with two magnetic planes attached at its both ends , in search of the modifications of magnetic coupling caused by the influence of external electric field originating from the gates .
our aim is to find conditions in which the sign of the coupling can be switched between ferro- and antiferromagnetic one using the electric field . for our study
we select a monolayer graphene nanostructure being a short fragment of armchair graphene nanoribbon .
we mention that the magnetic properties of such structures attracted already some attention in the literature @xcite and that nanoribbon - shaped structures with well - defined edges can be obtained experimentally @xcite .
( a ) schematic view of a short fragment of armchair graphene nanoribbon , characterized by zigzag terminations .
the structure consists of @xmath0 dimer lines , each composed of @xmath1 atoms , where @xmath2 dimer line is symmetry axis .
( b ) view of the nanostructure placed between gates producing uniform electric field in zigzag direction and with magnetic planes attached to outermost atoms of zigzag terminations . ]
the subject of our interest in the present work is a graphene nanostructure in a form of a short fragment of an armchair - edged graphene nanoribbon ( in fact constituting a graphene quantum dot ) .
the schematic view of the system is presented in fig .
[ fig : fig1](a ) .
the structure has zigzag - like terminations and is composed of @xmath0 atomic rows ( dimer lines ) extending in armchair direction .
we consider only odd @xmath0 values , so that the structures possess symmetry axis along the dimer line in the center , the position of which we denote by @xmath2 .
the remaining dimer lines are numbered by @xmath3 .
the number of atoms along each dimes line is equal to @xmath1 , being an even number . the numbers @xmath0 and @xmath1 characterize uniquely the nanostructure , which is composed of @xmath4 carbon atoms , belonging to two inequivalent , interpenetrating sublattices ( as marked with filled and empty circles in fig .
[ fig : fig1 ] ) . at the zigzag terminations ,
the magnetic planes are attached to the graphene nanostructure in such a way that the exchange interaction between magnetic moments in planes and spins of charge carriers in the nanostructure occurs only at the outermost atoms of each termination , i.e. at the ends of dimes lines with even values of @xmath5 .
what is more , the nanostructure is placed between the gates providing an uniform electric field along zigzag direction ( perpendicular to the dimes lines ) .
both the magnetic planes and the gates are marked in the schematic drawing fig .
[ fig : fig1](b ) . in order to characterize the indirect magnetic interactions mediated by a graphene nanostructure , it is crucial to describe its electronic structure , in particular in the sector involving the @xmath6 orbitals . in our work
we employ a tight - binding model supplemented with an on - site hubbard term in mean field approximation , which is commonly used in studies of graphene nanostructures @xcite ( including description of indirect magnetic coupling @xcite ) .
the hamiltonian of the model is the following : @xmath7 the operator @xmath8 ( @xmath9 ) creates ( annihilates ) an electron with spin @xmath10 at site @xmath11 , while @xmath12 represents the number of electrons . in tight - binding term
, @xmath13 is the hopping integral between nearest - neighbour carbon atoms belonging to a pair @xmath14 . for atomic pairs at the armchair edges of the nanostructure ( i.e. for @xmath15 )
, its value is taken as @xmath16 .
this selection depicts the edge deformation of the armchair graphene nanoribbons consisting in shortening of the interatomic distances along the armchair edges @xcite ; we use the value of @xmath17 ( what corresponds to bond deformation of approximately 3.5 % @xcite ) . for the rest of the nearest - neighbour pairs ,
we assume @xmath18 . the parameter @xmath19 is taken in further calculations as a convenient energy scale , to which all the other quantities can be normalized .
the value of @xmath19 amounts to approximately 2.8 ev @xcite in graphene .
the on - site hubbard repulsion energy @xmath20 is an effective parameter , whose value takes into consideration the long - range nature of coulombic interactions and is estimated as @xmath21 @xcite , which value we accept in our calculations .
the external uniform electric field of the gates is denoted by @xmath22 , while @xmath23 is the distance between the subsequent parallel atomic rows in armchair direction ( i.e. distance between the dimer lines ) and @xmath24 is the number of dimer line to which @xmath11-th atom belongs .
the distance between nearest - neighbour carbon atoms we denote by @xmath25 ( which is about 1.4 @xcite ) .
it is instructive to mention that the normalized field of @xmath26 corresponds to the electric field of approximately 2.3 v / .
the interaction between the spins of charge carriers and magnetic moments of both magnetic planes attached to the zigzag terminations of the nanostructure is parametrized by exchange energy @xmath27 .
as mentioned , the coupling involves only the spins of the charge carriers present at zigzag terminations for even values of @xmath5 ( and the number of such sites at each termination is @xmath28 ) .
therefore , @xmath29 and @xmath30 denote only the summation over the outermost lattice sites of zigzag terminations .
the magnetic moments associated with the planes interacting with charge carrier spins possess magnitude of @xmath31 and are denoted by @xmath32 and @xmath33 , respectively .
let us state that the problem of selection of the value of exchange energy between magnetic impurity spin and charge carrier spin in graphene is present also in other theoretical works , for example those related to spin relaxation in graphene @xcite , where the value of @xmath34 ev was selected as representative .
on the other hand , a value of @xmath35 ev was used in interpretation of experimental results on spin current scattering @xcite .
the hamiltonian ( eq . [ eq : eq1 ] ) can be presented in a form of @xmath36 , with two interdependent hamiltonians @xmath37 , @xmath38 for spin up and spin - down charge carriers ( see e.g. the detailed description in ref .
the presence of @xmath39 electrons on @xmath6 orbitals is assumed , what corresponds to charge neutrality of the structure ( one electron per carbon atom ) .
both hamiltonians can be diagonalized numerically in single - particle approximation , what allows for self - consistent determination of the energy eigenvalues @xmath40 for both orientations of spin . as a consequence
, the total energy of the charge carriers in the ground state can be calculated as @xmath41 , where average values of charge concentrations @xmath42 correspond to self - consistent solution and the single - particle eigenenergies are assumed to be sorted in ascending order . in order to reach a ground state of the system , the self - consistent calculations are performed separately for all the possible numbers of electrons with spin up ( @xmath43 ) and spin down ( @xmath44 ) satisfying @xmath45 and the state with the lowest total energy is accepted . moreover ,
the calculations are repeated for random initial conditions ( charge distribution ) for the purpose of finding a true ground state . if diagonalization of is performed both for antiparallel and parallel orientation of spins of magnetic planes @xmath32 and @xmath33 ,
then an indirect coupling energy between the planes can be determined .
it is expressed by the formula @xmath46 , where @xmath47 ( @xmath48 ) is the total energy of charge carriers for parallel ( antiparallel ) orientation of spins @xmath49 and @xmath50@xcite .
let us put emphasis on the fact that such a calculation bases on a non - perturbative approach . in our work
the numerical diagonalization of the hamiltonian ( eq . [ eq : eq1 ] ) was performed with the help of the lapack package @xcite .
the presented model allows to study the influence of the external in - plane electric field on the indirect , graphene nanostructure - mediated coupling between magnetic planes for various sizes @xmath0 and @xmath1 .
this enables identification of the most interesting cases for which the variation of coupling under the influence of the electric field is most significant . in our analysis
we focus on rather small structures , composed of a few tens of carbon atoms .
what is more , due to the the non - perturbative nature of the calculations , a non - trivial analysis of the influence of exchange energy @xmath27 on the resulting @xmath51 is possible , thus various contributions to coupling can be identified .
at the beginning it is instructive to study some properties of the graphene nanostructures in question without including the interaction with magnetic planes .
this involves in particular the influence of the electric field on the charge and spin distribution . since the magnetic planes are attached to the outermost parts of the zigzag terminations of the structures ( to the lattice sites with even values of @xmath5 ) , it is most interesting to analyse the concentration of charge carriers particularly at these lattice sites .
the distribution of total charge density @xmath52 ( for both spin - up and spin - down electrons ) along the zigzag termination of the nanostructure is illustrated in fig .
[ fig : fig2 ] for various values of external electric field in a nanostructure characterized by @xmath53 and @xmath54 .
for @xmath55 the distribution of charge is homogeneous .
when in - plane electric field parallel to the zigzag terminations of the nanostructure is switched on , the charge redistribution starts .
the charge density at the symmetry axis ( @xmath2 ) remains insensitive to the field , while the most pronounced changes occur at the edges of the nanostructure ( @xmath56 ) , where they are first observable for weak field .
close to the field of @xmath57 , we deal with almost linear dependence of charge density on distance , as the edge near the negative potential gate is almost completely free from electrons , while at the opposite side , near the positive potential gate the electronic density is almost doubled .
it is visible that the changes of electronic density are non - linear in external field @xmath22 , unless the lattice sites lie close to the symmetry axis .
charge density at the lattice sites located at zigzag terminations of nanostructure with @xmath53 and @xmath54 for various strengths of external electric field .
] this feature can be followed in the dependence of the total electronic density @xmath52 at selected carbon sites on the in - plane electric field , which is shown in fig .
[ fig : fig3](a ) ( for the nanostructure with @xmath53 and @xmath54 ) .
the data concern the sites at the nanostructure terminations , where the charge carrier spins interact with magnetic planes . such sites are marked in the inset to fig . [ fig : fig3](a ) .
it is visible that at the site located at the symmetry axis of the nanoribbon , for @xmath2 , the electronic density is completely untouched by the external electric field . for the sites closer to the gate with negative potential the electronic density deceases , while the opposite changes occur at the sites located closer to the gate possessing positive potential . for the sites located at @xmath58 , the changes are symmetric with respect to the value of @xmath59 .
when the sites located closest to the edge ( @xmath56 ) are considered , the dependence remains linear up to @xmath60 , and then abrupt jumps of electronic density occur .
for the field approaching @xmath57 , electronic densities at those edge sites reach the saturation values ( equal to 0 for the site close to the negative potential gate and equal to 2 for the site closest to the positive potential gate ) . for the sites located at @xmath61 , the initial slope of the dependence
is greatly reduced , but the abrupt changes in @xmath52 occur at weaker field , first of them approximately at @xmath62 .
high - field jumps take place at similar values as for @xmath56 , but the saturation values are already not reached at @xmath57 . for the sites closest to the symmetry axis ( @xmath63 ) , the initial dependence of @xmath52 on @xmath22 is linear with small slope , and this slope increases and reaches a constant value at about @xmath64 . ( a ) dependence of ( a ) charge density and ( b ) spin density at lattice sites on the zigzag termination of the nanostructure [ see inset in ( a ) ] on the normalized electric field . both plots are prepared for the structure with @xmath53 and @xmath54 . ]
spin density distribution for the nanostructure with with @xmath53 and @xmath54 , in absence of interaction with external magnetic planes , for various electric fields for which nonzero spin at zigzag terminations emerges .
two different colours correspond to opposite orientations of electron spins . ]
energies of a few occupied electronic states close to the highest occupied one as a function of normalized electric field , for the nanostructure with with @xmath53 and @xmath54 , in absence of interaction with external magnetic planes .
the case ( a ) is for charge neutrality , while case ( b ) is for two electrons added to the system .
] this picture can be supplemented with the dependence of spin density @xmath65 on the electric field , presented in fig .
[ fig : fig3](b ) for the same sites as in fig .
[ fig : fig3](a ) .
the spin densities at sites @xmath58 are identical , thus full symmetry with respect to the nanostructure axis @xmath2 is present . up to @xmath62
there is no spin polarization at the considered sites .
however , at this field value , a range of nonzero spin polarization builds up and preserves up to @xmath66 .
this happens simultaneously for all the sites , however , the values of spin polarization are much higher for @xmath2 and @xmath61 than for the sites numbered by odd @xmath5 values .
this observation is particularly important in context of the fact that the magnetic planes are attached to the sites numbered by even values of @xmath5 .
another range of nonzero @xmath67 is present between @xmath68 and @xmath69 ( that time the polarization is significantly stronger close to the edges than close to the symmetry axis of the nanostructure ) .
the next range of nonvanishing polarization commences at the field of @xmath70 .
let us comment that these results do not imply necessarily appearance of a nonzero total spin of the nanostructure .
such a polarization does not take place in absence of electric field for the systems in question , what is due to the fact that the nanoflakes contain equal number of carbon atoms from both sublattices and no net spin polarization is predicted for charge neutrality conditions in such nanostructures @xcite .
however , in some range of the electric fields ( for example @xmath71 and @xmath72 ) , the total spin of the nanostructure equal to 1 is predicted . in order to illustrate that , we present the spin density distribution for a whole nanostructure in fig . [ fig : fig3b ] , for three values of electric field .
all three values fall within the range of appearance of nonzero spin polarization of the zigzag termination .
it is visible that for the largest and the smallest value ( i.e. for @xmath73 and @xmath74 ) , the total spin of the charge carriers in the nanostructure is zero , with opposite spin polarization of both edges . on the contrary , for @xmath75 , a nonzero total
spin emerges .
all the predicted spin polarizations manifest themselves mostly at the outermost sites of the nanostructure , mainly its zigzag terminations . in order to gain more insight into physical mechanism leading to the polarization
, we studied the energies of the discrete electronic states as a function of the electric field .
in particular , we focused our attention on the lowest field range in which magnetic polarization appears in the nanostructure with @xmath53 and @xmath54 .
the results are shown in fig .
[ fig : fignew ] , where the energies of a few occupied states close to the highest occupied one are plotted against normalized electric field . fig .
[ fig : fignew](a ) corresponds to the case of charge - undoped nanoribbon , while fig .
[ fig : fignew](b ) is prepared for a nanostructure doped with two additional electrons . for the system with two additional electrons
, no spin polarization emerges due to external electric field and all the states are spin - degenerate ( this case is shown as a reference one ) .
however , in a charge - neutral system , the situation is different . for the state with the lowest energy shown in fig .
[ fig : fignew](a ) , the zeeman splitting occurs in some @xmath22 range , so that spin - degeneracy is lifted .
however , the picture is slightly more complex in the case of the state with higher energy ( the highest occupied state ) .
namely , it lies in the vicinity of another state ( the one plotted with dashed line in fig . [
fig : fignew](b ) .
when both states experience zeeman splitting , it becomes more energetically favourable for one of the electrons to skip to the next available state ( plotted with dashed line both in fig .
[ fig : fignew](a ) and fig .
[ fig : fignew](b ) ) .
such a situation takes place because the energy of that state for spin up becomes lower that the energy of the previous state for spin down . therefore , net polarization with spin equal to 1 occurs
it can be concluded from the analysis of the fig .
[ fig : fignew ] that the appearance of the spin polarization is associated with a single electronic state and requires a pair of states with small separation in energy .
let us now turn the attention to the question of size of the graphene nanostructure mediating the interaction between the magnetic planes .
extending the mediating graphene nanostructure would lead eventually to achieving a limit of an infinite graphene monolayer , which is actually a zero - gap semiconductor and conducts current .
therefore , under such conditions , the model hamiltonian which we employ to describe the influence of the electric field on the system would no longer be physically justified . on the contrary ,
the applicability of the model is limited to systems with well - separated discrete energy states ( especially with pronounced highest occupied molecular orbital - lowest unoccupied molecular orbital gap ) . in particular , the spatial extension of the mediating nanostructure in the direction parallel to the electric field
should be strongly limited due to screening effects .
let us mention here the results of calculations in ref .
, concerning the screening of the parallel electric field in graphene nanostructures of various chirality .
it has been found that the screening appears the least pronounced and the potential profile across the structure is most similar to linear one in the case of an amchair nanoribbon ( see fig .
3(a ) in ref . ) . however ,
increasing the width of the structure would certainly limit the validity of the model which we employ to describe the effect .
another factor enforcing the usage of small graphene nanostructures is the fact that the switching of the indirect coupling is vitally dependent on the behaviour of a single energy state under the influence of the electric field , as explained before .
such a state develops spin polarization and spin density associated with individual lattice sites decreases with the increase of the number of carbon atoms building the structure .
therefore , the spin density at each outermost site of zigzag terminations , being in contact with magnetic plane , would be significantly limited in excessively large structures .
this limitation applies to the possibility of increasing the length of the nanostructure ( i.e. the distance between magnetic planes ) . on the other hand , limitations for increasing the width of the nanostructure
have been discussed in the above paragraph .
these considerations are aimed at justifying the interest in the smallest graphene nanostructures in the present paper .
the creation of nonuniform spin polarization at the zigzag termination of the structure under the influence of the electric field is particularly useful from the point of view of achieving electric - field control over the indirect coupling by attaching the magnetic planes to the edges .
since the outermost lattice sites at both zigzag terminations belong to different sublattices ( see fig . [
fig : fig1 ] ) , without electric field the indirect coupling is expected to indicate antiferromagnetic character ( e.g. @xcite ) .
this kind of coupling results from typical indirect rkky mechanism , which consists in creation of spin polarization by one of the magnetic planes and the interaction of the other plane with such polarization .
therefore , this process is describable ( for low @xmath27 exchange energies ) by means of second order perturbation calculus .
however , the presence of the electric - field induced spin polarization at the lattice sites where charge carrier spins interact with magnetic moments in the attached planes opens the room for appearance of another coupling mechanism .
namely , the first - order perturbative coupling mediated by spin - polarized states can become active .
such a coupling mechanism has already been predicted mainly in graphene nanoflakes @xcite , but also in armchair graphene nanoribbons of finite length @xcite .
therefore , the described effect of electric field on the electronic properties of the studied nanostructures potentially allow for continuous changing the indirect interaction strength and sign , which property we will show in further part of this section .
let us now present the results of calculations for the complete system , including the magnetic planes .
first we can follow the dependence of indirect coupling energy between the magnetic planes attached to the zigzag terminations of the sample graphene nanostructure as a function of the external electric field . the plot fig . [
fig : fig4](a ) presents the predictions of calculations for three representative nanostructures , each of width @xmath53 , but with lengths equal to @xmath54 , @xmath76 and @xmath77 .
the exchange energy is taken as @xmath78 and armchair edge deformation is included .
first it is visible that for low ( or zero ) electric field the coupling is antiferromagnetic ( af ) and its magnitude decreases with the increase of the structure size .
when the field increases , @xmath79 indicates first some plateau , but then becomes weaker and passes through zero , changing its sign to ferromagnetic ( f ) .
this is the effect we expect on the basis of the analysis of spin polarization . for the nanostructure with @xmath53 and @xmath54 , the field at which the sign change occurs coincides with the onset of spin polarization visible in fig .
[ fig : fig3](a ) .
the coupling energy reaches a maximum and the drops again to antiferromagnetic values when @xmath22 increases further .
a qualitatively similar behaviour of @xmath79 as a function of electric field is seen for the longer nanostructures , with @xmath80 and @xmath81 .
however , it that cases the ferromagnetic maximum is much reduced in magnitude and is shifted to weaker fields @xmath22 .
what is more , the oscillatory character of changes of @xmath51 for higher fields becomes more pronounced for larger structures .
finally , let us observe that for larger lengths the low - field ferromagnetic maximum tends to vanish .
an analogous analysis is presented in fig .
[ fig : fig4](b ) , but that time for structures sharing the same length @xmath54 and different in width : @xmath82 , @xmath83 or @xmath84 . for the narrowest of the nanostructures ,
the magnitudes of indirect coupling are generally reduced and for increasing @xmath22 the interaction energy slowly tends to zero , however , without visible ferromagnetic maximum .
when @xmath0 increases to @xmath83 , a behaviour described in the above paragraph emerges . if @xmath85 , still a ferromagnetic maximum occurs ( like that for @xmath53 ) , but the strongest coupling is reduced and the peak is present at weaker fields .
the results described above allow to state that for some sizes of the studied nanostructures , an electric field - induced , continuous switching between antiferro- and ferromagnetic indirect coupling can be achieved .
it is further interesting to verify the stability of the effect with respect to the value of exchange energy parameter @xmath27 .
dependencies of @xmath51 on electric field @xmath22 are plotted in fig .
[ fig : fig5 ] for four selected values of exchange energy @xmath27 , ranging from @xmath86 up to @xmath87 .
the selected nanostructure is the one which provides the most pronounced field - induced ferromagnetic peak in indirect interaction , i.e. @xmath53 and @xmath54 .
it is visible that the ferromagnetic peak occurs for all the values of @xmath88 and its location is similar in all the cases . on the other hand , for the antiferromagnetic ranges surrounding this peak , @xmath51 is very sensitive to exchange coupling energy @xmath27 .
especially , a very deep minimum can emerge at around @xmath89 when @xmath88 is strong enough .
the values of the electric field at which @xmath51 crosses zero ( the boundaries of the ferromagnetic range ) can be studied in more detailed way as a function of exchange energy @xmath88 for various sizes of the nanostructure . in view of the potentially wide range of parameters @xmath27 describing the exchange coupling between magnetic planes and spins of the charge carriers ,
it is highly interesting to determine the behaviour of the critical electric field @xmath90 for which the indirect coupling crosses the zero value as a function of exchange energy @xmath88 .
such results are presented in fig .
[ fig : fig9 ] for various sizes of a nanoribbon piece mediating the coupling .
filled symbols correspond to the field of transition between low - field af interaction and higher - field f interaction . on the other hand
, empty symbols denote the electric field at which f - interaction switches back to af , what takes place for higher fields . in fig .
[ fig : fig9](a ) structures of width @xmath53 and three lengths : @xmath91 are considered .
it is visible that the dependence of the normalized electric fields @xmath92 on @xmath88 is not very pronounced ( unless @xmath93 ) , the main feature being a shallow extremum in the vicinity of @xmath94 .
therefore , around this value of @xmath27 the ferromagnetic interaction covers the widest range of electric fields .
the tendencies of variation of both characteristic fields ( for af - f and f - af transition ) as a function of @xmath27 are opposite .
what is crucial , the effect of changing the coupling @xmath51 from af to f is present in the whole studied range of exchange potentials @xmath27 .
the situation is slightly different in fig .
[ fig : fig9](b ) , which presents analogous data for structures with constant length @xmath54 and three different widths @xmath95 . when the width increases , the minimum in dependence of @xmath90 for af - f transition on @xmath27 shifts significantly to lower values of @xmath27
. moreover , for @xmath96 an additional maximum appears ( also moving to lower @xmath88 when @xmath1 increases ) .
the overall dynamics of the critical fields @xmath90 is more significant for wider pieces of nanoribbons .
the field of f - af transition ( the higher one ) shows in this case a very similar behaviour like in fig .
[ fig : fig9](a ) .
also for the set of nanostructures analysed here , the significant range of field - induced f interaction is present for every considered value of @xmath27 .
the critical fields for switching between ferro- and antiferromagnetic indirect coupling can also be followed as a function of width of the mediating graphene nanostructure .
the results of such calculations are presented in fig .
[ fig : fig6b ] for fixed length @xmath54 and for @xmath78 . like in fig .
[ fig : fig9 ] , the filled symbols correspond to af - f switching when electric field increases , while the empty ones denote further af - f transition . when the width of the nanostructure increases , a monotonous decrease of the critical fields is visible , together with significant reduction of the range of fields in which f indirect coupling is present ; however , such a range was always found .
the inset in fig .
[ fig : fig6b ] presents also the indirect coupling values achieved either at zero field ( strongest af interaction ) or at maximum of f coupling ( at the field in between critical fields plotted in fig .
[ fig : fig6b ] ) . for these quantities ,
a sort of non - monotonous behaviour is found , with reduction of magnitude of f coupling when @xmath0 increases , while af interaction shows rather an increase with some oscillatory tendency .
nevertheless , a considerably high sensitivity of @xmath51 to external field is present in the whole range of studied widths , and the widest structures offer rather weak critical field for switching , still with significant amplitude of both af and f interaction .
let us now focus the attention on the indirect interaction energy at the extremal values , i.e. for @xmath97 ( approximately the ferromagnetic peak ) , for @xmath98 ( the subsequent antiferromagnetic minimum ) as well as at @xmath99 .
the dependence of @xmath51 at these field values on exchange energy @xmath27 is presented in fig .
[ fig : fig6](a ) in linear scale and fig .
[ fig : fig6](b ) in double logarithmic scale ( where the absolute values are plotted ) .
it is seen that in absence of electric field @xmath51 is proportional to @xmath100 for @xmath101 , what indicates that the origin of the indirect coupling is the second - order perturbative mechanism .
the similar property is observed for coupling at @xmath98 , i.e. at the antiferromagnetic minimum . for stronger exchange @xmath27 , the antiferromagnetic coupling changes abruptly the slope . for @xmath102 , after some range of weak dependence on @xmath27 , it starts again to become stronger in the vicinity of @xmath103 , but with reduced slope in comparison to the range of small @xmath27 .
the behaviour of @xmath51 at ferromagnetic maximum ( @xmath104 ) is , however , different . for weak
@xmath105 it varies linearly with @xmath27 , what proves that the leading contribution to the indirect coupling comes from first - order perturbative process .
such a situation is expected on the grounds of the predicted spin polarization emerging in the range of electric fields close to @xmath104 [ see fig . [
fig : fig3](a ) ] . for stronger exchange parametrized by @xmath27 ,
the slope of @xmath51 is reduced and it varies with @xmath27 approximately proportionally to @xmath106 . normalized indirect coupling as a function of normalized electric field in nanostructures of varying size : ( a ) for width @xmath53 and three lengths : @xmath91 ; ( b ) for length @xmath54 and three widths : @xmath107 . ] normalized indirect coupling as a function of normalized electric field in nanostructure characterized by @xmath53 and @xmath54 , for four different values of exchange energy @xmath27 . ]
critical normalized electric field of transition between ferro- and antiferromagnetic coupling between magnetic planes as a function of exchange energy @xmath27 in nanostructures of varying size : ( a ) for width @xmath53 and three lengths : @xmath91 ; ( b ) for length @xmath54 and three widths : @xmath108 .
filled symbols denote the lower field , at which af coupling turns into af one , while filled symbols correspond to a higher field at which opposite transition takes place . ] critical normalized electric field of transition between ferro- and antiferromagnetic coupling between magnetic planes as a function of nanostructure width for fixed length @xmath54 .
filled symbols denote the lower field , at which af coupling turns into f one , while filled symbols correspond to a higher field at which opposite transition takes place .
inset presents normalized indirect coupling energy at zero field ( filled rectangles ) and at the first ferromagnetic maximum ( filled triangles ) as a function of nanostructure width for fixed length @xmath54 . ] normalized indirect coupling as a function of exchange energy @xmath27 for three values of electric field , corresponding approximately to extremum magnitudes of indirect coupling .
the results concern the nanostructure with @xmath53 and @xmath54 .
( a ) indirect coupling in linear scale ; ( b ) indirect coupling in double logarithmic scale . ] the first - order perturbative contribution to indirect coupling can be described quantitatively by referring to the corrections to total energy of the system of charge carriers introduced by interaction with magnetic planes .
this can be done in a very similar way as in ref . .
therefore , @xmath109 , where @xmath110 is the first - order perturbative correction to the ground state energy of the charge carriers due to their interaction with magnetic planes ( with either f or af orientation of magnetic moments ) .
this correction is directly proportional to the total spin density at the appropriate sites where the contact interaction occurs . in fig .
[ fig : fig7 ] we present the dependence of the total spin density at sites interacting with magnetic planes ( separately for both nanostructure terminations ) on exchange energy @xmath27 in the presence of electric field @xmath97 .
the outermost sites are schematically marked with frames in the insets .
the cases of plane magnetizations oriented ferromagnetically and antiferromagnetically are considered separately , as indicated in the insets .
if magnetizations of both planes are parallel , the total spin density at both nanostructure terminations takes identical values .
it is visible that this value increases monotonously with @xmath27 ; the rise is very slow for low @xmath88 and for @xmath111 the slope increases and @xmath112 .
the situation is somehow different when magnetizations of both planes are antiparallel . in such circumstances the total spins at both terminations are different . for @xmath105 , at one termination the spin rises with @xmath27 , while at the other one it decreases , while both spins keep parallel orientation .
however , at @xmath113 the spins become antiparallel and their absolute values become identical ( a kind of metamagnetic transition takes place ) . therefore , the total spin density at outermost atoms of both zigzag terminations is equal to 0 .
it can be concluded that for @xmath105 the spin density is only weakly influenced by exchange interaction with magnetic planes . on the other hand , for @xmath114 , the leading contribution to indirect coupling originates only from the ferromagnetic state of magnetization of both planes .
therefore , since spin density approximately increases according to @xmath112 , also @xmath51 follows this proportionality .
let us state that this is no longer a first - order contribution to indirect coupling , since the spin density is modified by exchange interaction between planar magnetic moments and spins of charge carriers and such a modified value enters the formula for indirect exchange .
the second one of the mentioned ranges seems to contain the most important values of @xmath27 for practical uses .
total spin density at outermost sites of zigzag terminations of the nanostructure with @xmath53 and @xmath54 as a function of exchange energy @xmath27 , for ferromagnetic ( solid line ) and antiferromagnetic ( dashed and dotted lines ) orientation of magnetic moments in both planes attached to the nanostructure .
electric field of @xmath97 is present . ] in order to identify the crucial factors shaping the response of the indirect coupling to the electric field , let us supplement the previous results with those shown in fig .
[ fig : fig8 ] . in that plot ,
@xmath51 is shown as a function of normalized electric field for the representative and most interesting nanostructure with @xmath53 and @xmath54 .
different lines correspond to the results in which armchair edge deformation ( parametrized by @xmath115 ) and coulombic interactions ( parametrized by hubbard @xmath20 parameter ) are separately included or excluded .
first it is noticeable that the weakest antiferromagnetic indirect exchange at zero electric field occurs for the absence of coulombic interactions and edge deformation .
emergence of the edge bonds shortening for @xmath116 is only weakly reflected in @xmath51 , while switching on @xmath20 results in considerable strengthening of indirect coupling .
for the range of electric fields for which switching to ferromagnetic interaction happens , the similar situation can be seen .
namely , the most crucial factor influencing the coupling is coulombic interaction ( positive @xmath20 value ) , while @xmath115 has only slight effect on @xmath51 .
if the electric field becomes stronger , the important influence of edge deformation can be noticed mainly in presence of @xmath117 for the deep antiferromagnetic minimum .
it can be generally concluded that the presence of coulombic interaction tends to boost the magnitude of indirect coupling , conserving its sign .
this factor also extends significantly the range of ferromagnetic interaction peak present at @xmath118 , shifting the field for which sign changes back to antiferromagnetic one towards higher values .
therefore , coulombic interactions promote the features which are desirable from the point of view of obtaining a clearly field - controlled indirect coupling .
let us also observe that the electric field switching of @xmath51 between antiferro- and ferromagnetic values is observable for every choice of parameters in fig .
[ fig : fig8 ] , what allows to state that this feature is considerably robust . indirect coupling between the planes as a function of normalized external electric field for a nanostructure with @xmath53 and @xmath54 , in presence and in absence of coulombic interactions ( parametrized by hubbard parameter @xmath20 ) and armchair edge deformation of the nanostructure .
in our study we have demonstrated the prediction of indirect coupling between magnetic planes mediated by charge carriers in a graphene nanostructure of particular size and geometry .
the crucial characteristic of this interaction is its sensitivity to external electric field in plane of the nanostructure .
namely , the coupling which is antiferromagnetic below certain critical electric field can be continuously switched to ferromagnetic sign for field strong enough .
further increase of the field brings the coupling back to antiferromagnetic regime .
the effect is present for a wide range of exchange energies between attached magnetic planes and spins of the charge carriers .
moreover , it is also robust against deformation of the armchair edge of the mediating nanostructure and boosted by coulombic interactions .
the origin of the effect is connected with field - induced appearance of the nonzero spin density distributed along the zigzag termination , not necessarily connected with net magnetic polarization of the structure as a whole .
the extreme values of ferro- and antiferromagnetic indirect interaction are found to be of comparable magnitude .
let us emphasize that the nanostructures mediating the interaction ( in the form considered in the present study ) appear experimentally accessible and their magnetic properties can be examined @xcite .
moreover , there is also a constant progress in achieving well - controllable graphene nanoribbons , including the development of bottom - up methods utilizing molecular precursors ( see e.g. @xcite ) .
this encourages studies of geometrically well - defined graphene structures , focusing on their usefulness in spintronics .
moreover , in presence of magnetic planes attached to the zigzag edges of the nanostructure , the conditions for emergence of zigzag edge spin polarization should be more favourable than those in free - standing graphene , since the magnetic planes serve as a source of exchange field and magnetic anisotropy stabilizing the ordering against thermal excitations .
this factor , together with recent results concerning zigzag edge magnetism @xcite , supports interest in such systems
. the described effect can find application in construction of electric - field controlled spintronic devices based on graphene .
the further developments may include search for analogous field - tunable effects among structures of various geometry as well as , for example , study of simultaneous effect of electric and magnetic field on the described phenomenon .
moreover , the effect of charge doping might be worthy of investigations .
however , let us mention that the properties of the smallest mediating nanostructures appear the most promising , since the increase in size causes inevitably the screening effects to reduce the influence of external electric field on the desirable properties @xcite .
let us also comment that for the purpose of increasing the energy of indirect coupling between the magnetic planes , the device would be imagined with numerous identical mediating nanostructures attached to the planes .
these structures would be placed at equal distance one from another , possibly each structure accompanied with its own pair of electric gates .
in such a way the coupling would be enhanced .
the computational support on hugo cluster at laboratory of theoretical aspects of quantum magnetism and statistical physics , p. j. afrik university in koice is gratefully acknowledged .
this work has been supported by polish ministry of science and higher education on a special purpose grant to fund the research and development activities and tasks associated with them , serving the development of young scientists and doctoral students .
91ifxundefined [ 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.1102896 [ * * , ( ) ] link:\doibase 10.1038/nature04233 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.76.323 [ * * , ( ) ] http://stacks.iop.org/0022-3727/47/i=9/a=094011 [ * * , ( ) ] link:\doibase 10.1038/nmat3305 [ * * , ( ) ] link:\doibase 10.1557/mrs.2012.277 [ * * , ( ) ] http://stacks.iop.org/1367-2630/11/i=9/a=095005 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.245415 [ * * , ( ) ] link:\doibase 10.1038/nphys547 [ * * , ( ) ] link:\doibase 10.1038/nnano.2008.163 [ * * , ( ) ] `` , '' in link:\doibase 10.1002/9781118678107.ch23 [ _ _ ] , ( , ) pp .
http://stacks.iop.org/0957-4484/21/i=38/a=385201 [ * * , ( ) ] link:\doibase 10.1063/1.4766914 [ * * , ( ) ] http://stacks.iop.org/0034-4885/73/i=5/a=056501 [ * * , ( ) ]
link:\doibase 10.1103/physrevlett.99.177204 [ * * , ( ) ] link:\doibase 10.1021/nl072548a [ * * , ( ) ] link:\doibase 10.1021/nl073364z [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.157201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.081402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.205306 [ * * , ( ) ] \doibase doi:10.1166/jnn.2012.6532 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.075425 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.075431 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.115433 [ * * , ( ) ] link:\doibase 10.1016/j.ssc.2013.09.023 [ * * , ( ) ] link:\doibase 10.1016/j.spmi.2013.09.015 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.035403 [ * * , ( ) ] link:\doibase 10.1016/j.physleta.2013.02.029 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.035425 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.246805 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.035425 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.115416 [ * * , ( ) ] link:\doibase 10.1166/jnn.2012.5350 [ * * , ( ) ] link:\doibase 10.1063/1.4858378 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.201411 [ * * , ( ) ] http://stacks.iop.org/0953-8984/26/i=5/a=055007 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.205409 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.045422 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.88.155429 [ * * , ( ) ] link:\doibase 10.7567/apex.5.063007 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4823592 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2362971 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.085315 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.88.207204 [ * * , ( ) ] http://stacks.iop.org/1367-2630/11/i=2/a=023008 [ * * , ( ) ] link:\doibase 10.1038/nphys1455 [ * * , ( ) ] link:\doibase 10.1038/35050040 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.113302 [ * * , ( ) ] link:\doibase 10.1143/jjap.43.l691 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2011.02.002 [ * * , ( ) ] link:\doibase 10.1016/j.jmmm.2014.03.045 [ * * , ( ) ] link:\doibase 10.1016/j.jmmm.2013.09.018 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.144410 [ * * , ( ) ] link:\doibase 10.1038/ncomms3010 [ * * , ( ) ] link:\doibase 10.1063/1.4821954 [ * * , ( ) ] link:\doibase 10.1021/nn4037877 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.205427 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.165446 [ * * , ( ) ] link:\doibase 10.1021/jp310463k [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.165429 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/23/20/205202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.085414 [ * * , ( ) ] http://stacks.iop.org/1742-6596/200/i=6/a=062015 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.115445 [ * * , ( ) ] link:\doibase 10.1021/jp205841z [ * * , ( ) ] http://stacks.iop.org/1367-2630/13/i=5/a=053043 [ * * , ( ) ] link:\doibase 10.1126/science.1130681 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.155433 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.235111 [ * * , ( ) ] http://stacks.iop.org/0953-8984/25/i=16/a=166001 [ * * , ( ) ] link:\doibase 10.1016/j.ssc.2013.07.005 [ * * , ( ) ] link:\doibase 10.1021/nn303730v [ * * , ( ) ] link:\doibase 10.1021/ar3001487 [ * * , ( ) ] link:\doibase 10.1038/nature09211 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.073409 [ * * , ( ) ] link:\doibase 10.12693/aphyspola.126.236 [ * * , ( ) ] link:\doibase 10.1016/j.physe.2013.03.017 [ * * , ( ) ] link:\doibase 10.1016/j.ssc.2014.07.006 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.216803 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.81.109 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.036601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.116602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.186604 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.184201 [ * * , ( ) ] @noop _ _ , ed .
( , , ) link:\doibase 10.7567/jjap.53.06jd05 [ * * , ( ) ] link:\doibase 10.3390/cryst3010049 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.245428 [ * * , ( ) ] link:\doibase 10.1021/ja309599 m
[ * * , ( ) ] link:\doibase 10.1143/apex.5.095101 [ * * , ( ) ] | in the paper we consider the system composed of two magnetic planes attached to zigzag terminations of the graphene nanostructure being an ultrashort fragment of the armchair nanoribbon .
we investigate theoretically an indirect coupling between these magnetic planes mediated by charge carriers as a function of external in - plane electric field .
the calculations are based on a tight - binding model supplemented with hubbard term to account for coulombic interactions . for selected sizes of the graphene nanostructure ,
particularly high sensitivity of the coupling to the electric field is found .
this leads to the possibility of control over coupling magnitude and continuous switching of its sign between antiferromagnetic and ferromagnetic one .
such a phenomenon is demonstrated in the numerical results and its origin is analysed .
the robustness of this effect against armchair edge deformation and variation of exchange energy between magnetic planes and spins of charge carriers is discussed in detailed way . |
supernova remnants ( snrs ) are believed to be the main sources of galactic cosmic rays , charged particles with energies up to @xmath8 ev that are thought to gain their energy mainly through the mechanism of first order fermi acceleration in the shocks of snrs @xcite and its non - linear modification ( e.g. , * ? ? ?
however , since these particles do not travel directly from their sources due to the presence of galactic magnetic fields that change their directions of motion , the cosmic ray content of snrs is studied indirectly by analyzing the properties of their non - thermal photon spectrum .
high - energy electrons produce synchrotron emission that is seen mainly at radio wavelengths and up to x - ray energies ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for examples of x - ray spectra of non - thermal filaments in snrs ) .
these electrons may also produce leptonic gamma - ray photons through inverse compton ( ic ) up - scattering of low energy ambient photons and bremsstrahlung emission ( e.g. , * ? ? ?
gamma - rays may also result from hadronic interactions through the decay of neutral @xmath9mesons produced in inelastic collisions between cosmic - ray protons ( or ions ) and ambient nuclei .
the detection of a characteristic pion decay signature in the gamma - ray spectrum of snrs would therefore constitute direct evidence for the existence of high - energy cosmic rays in these sources .
an important window of observations , with improved energy and angular resolution , mainly in the range from @xmath10 100 mev to 300 gev , has been opened with the launch of the _ fermi _ observatory @xcite .
however , most studies of gamma - ray observations of snrs have yielded ambiguous results regarding the nature of the emission .
snrs that are relatively bright in gamma - rays and interact with molecular clouds , such as w28 , w49b , w51c , g8.7 - 0.1 and , possibly , rcw 103 , show spectra that seem to favour a hadronic origin @xcite .
recently , it has been claimed that , below @xmath10 100 mev , the spectrum of the snrs ic 443 and w44 can not be accounted for by leptonic mechanisms and shows a typical hadronic behaviour @xcite .
other examples of gev snrs include the lower luminosity extended sources cygnus loop @xcite , s147 @xcite and g296.5 + 10.0 @xcite ; and younger snrs , such as cas a @xcite , tycho s snr @xcite , rcw 86 @xcite , rx j1713.7 - 3946 @xcite and rx j0852.0 - 4622 @xcite . in this paper , cumulative observations of the region around the snr hb9 ( g160.9 + 2.6 ) by the large area telescope ( lat )
onboard _ fermi _ are analyzed .
hb9 is a large diameter ( radio extension @xmath11 ) , evolved , shell - type snr showing thermal x - ray emission in the interior @xcite and fragmented radio shells @xcite .
based on x - ray observations , a distance of 1.5 kpc and an age of @xmath12 kyr are estimated @xcite , while a distance of @xmath13 kpc is obtained from a likely association with hi structures @xcite .
there are several other estimates of these parameters in the literature , for example , the evaporative cooling model developed by white & long @xcite give an age of 4 - 7 kyr and an explosion energy of @xmath14 erg .
an explosion energy of @xmath15 erg , an age of 8 kyr and a distance of 1 kpc are adopted in this work .
the large extension of the snr has allowed radio spectral studies in different regions of the source . after subtracting background radio sources ,
flux measurements have yielded a radio spectral index @xmath16 ( @xmath17 ) between 408 and 1420 mhz @xcite and @xmath18 @xcite for the remnant .
spatial variations in spectral index seen @xcite imply that the radio spectra of regions near the rim of hb9 are harder than those of regions towards the interior of the remnant , which is explained in terms of a higher compression and therefore higher magnetic field values at larger distances from the center of the snr , which is also consistent with the higher radio brightness seen in the outer shell compared to the interior @xcite .
hb9 is believed to be located near the molecular clouds sh217 and sh219 @xcite , but no definitive interaction with a molecular cloud has been claimed .
as has been mentioned , ambient gas interacting with snrs might provide target material for cosmic - ray collisions , which is thought to be responsible for an enhancement of gamma - ray emission seen in some sources .
several compact sources lie in the field of view of hb9 ( e.g. , * ? ? ?
the snr might be associated to the pulsar psr b0458 + 46 ( e.g. , * ? ? ?
perhaps the most prominent source seen in the radio images of the region is the extragalactic radio source 0503 + 467 near the eastern edge of hb9 ( source flux density of @xmath19 jy at 1.4 ghz , spangler et al .
this source has been observed with vlbi by spangler et al .
( 1986 , 1987 ) in search for predicted angular broadening by turbulence outside the shock of the snr .
the authors give a position for this source of ra ( j1950)= 5@xmath2003@[email protected] , dec ( j1950)= + 46@xmath2341@[email protected] and its radio spectrum is reported ( spangler et al .
1987 ) , which is seen to flatten due to synchrotron self - absorption at frequencies of 1 - 2 ghz .
the gamma - ray observation by the lat of a region containing hb9 is described in section [ observations ] .
the spatial and spectral properties of the observed emission are analyzed and its nature explored .
the overall sed is modeled in section [ model ] .
the discussion of results and conclusions are given in section [ discussion ] .
_ fermi _ lat data between the beginning of the mission , 04 august 2008 , and 28 february 2014 were analyzed with the standard software _ sciencetools _
version v9r32p5 released october 24 , 2013 with the latest reprocessed pass7 photon and spacecraft data and the instrument response functions p7rep_source_v15 .
standard selection criteria are applied to the data selecting _ source _ class events and a reconstructed zenith angle less than 100@xmath26 to avoid contamination from gamma rays from earth s limb .
recommended time intervals for data analysis are selected with the standard criteria ( data_qual==1 ) & & ( lat_config==1 ) & & abs(rock_angle)@xmath2752 .
the spectral analysis is restricted above 200 mev due to uncertainties in the effective area and broad psf at low energies , and below 100 gev due to limited statistics .
events within a square region of 14@xmath2814@xmath26 of the catalogued position of hb9 , ra ( j2000)= 5@xmath2001@xmath2100@xmath22 , dec ( j2000)= + 46@xmath2340 , are used in the analysis to account for the large psf .
the emission model from this region includes the positions and spectral shapes of the point sources reported in the lat 2-year source catalog @xcite .
the recently released galactic diffuse emission model ( _ gll_iem_v05_rev1.fit _ ) is used to account for the local extended background .
this model corrects for anomalously low molecular hydrogen emissivities used in previous models , which improves the quality of the analysis mostly for galactic sources in the longitude range @xmath29 to @xmath30 . with the new galactic background model , a considerable level of background emission that was previously unaccounted for
was found to be particularly important near and at the position of hb9 .
the isotropic extragalactic file _
iso_source_v05.txt _ is also used in the background model .
spectral parameters of sources located beyond 10@xmath23 of the position of hb9 are kept fixed to the values reported in the catalog .
the fit is performed with the optimizer newminuit until convergence is achieved .
the spatial and spectral properties of lat sources are obtained by means of a maximum likelihood analysis using the tool _ _ gtlike _ _ , which estimates the probability of obtaining the data given a source model and fits the parameters to maximize this probability .
the previously detected _ fermi _ source 2fgl j0503.2 + 4643 which is seen towards the interior of the snr is not included in the analysis .
the resulting model is referred to as the null hypothesis .
another nearby gamma - ray source , 2fgl j0503.3 + 4517 , lies at @xmath31 from the center of hb9 , outside of the edge of the radio image .
it has been associated with a background galaxy @xcite , and thus it is treated here as a background source . the significance of a source is estimated by the test statistic ( ts ) defined as @xmath32 log(@xmath33 ) , where @xmath34 and @xmath35 are the maximum likelihood values for the null hypothesis and for a model including the additional source , respectively @xcite .
for large number of counts the detection significance of the source is given by @xmath36 ( however see * ? ? ?
* for some caveats ) .
in order to probe for new gamma - ray sources in the data , a ` residuals map ' is obtained above 1 gev to take advantage of the narrower psf at higher energies .
the map is obtained by subtracting the observed counts map from the model map , which is constructed with the resulting likelihood fit to the data , and it is shown in fig .
[ fig1 ] .
extended excess emission is apparent in fig .
[ fig1 ] , which also shows the radio and x - ray contours of the source obtained from a 4850 mhz green bank catalogue ( gb6 ) image @xcite and a rosat observation , respectively . using events above 1 gev
, several hypotheses for the morphology of the residual gamma - ray emission are tested : spatial templates obtained from ( a ) the radio gb6 observation , ( b ) a rosat pspc ( x - ray ) observation of hb9 and ( c ) a uniform disk template . the x - ray emission from hb9 is concentrated towards the center of the snr @xcite , as shown in fig .
[ fig1 ] .
the center of the disk template , and its radius , are independently and systematically changed throughout the region of hb9 and the resulting ts value calculated .
the highest ts for the disk is obtained for a template with a radius of @xmath37 @xmath38 and whose center is located within @xmath39 of the position ra ( j2000)= 5@xmath2001@xmath2117@xmath22 , dec ( j2000)= + 46@xmath2324@xmath2449@xmath25 .
the quoted errors in the radius and position are at the 3@xmath0 level .
the results of the morphological analysis are shown in table [ table1 ] .
the ts values with respect to the null hypothesis for the spatial morphologies above 1 gev , assuming a power - law spectral shape above this energy , are ( a ) 91 , ( b ) 64 , and ( c ) 96.8 ( in the same order as before ) .
although the uniform disk shows the highest ts value , the improvement to the fit is not significant compared to the following best - fit scenario ( the radio template ) , since it has 3 additional degrees of freedom .
thus , the radio morphology is adopted here as a better description of the gamma - ray excess for the rest of the analysis .
the photon spectrum in the range 0.2 - 100 gev is now obtained for the radio template .
different spectral shapes are explored and the results are shown in table [ table2 ] .
the spectral shapes probed are : a power - law ( @xmath40 ) , a power - law with an exponential cutoff ( @xmath41exp@xmath42 ) and a `` log - parabolic '' spectrum ( @xmath43 ) .
the resulting ts values are , respectively , 230 , 255 and 261 .
evidence of spectral curvature is observed given the fit improvement using a log - parabola over a simple power - law .
the highest ts value ( 261 ) corresponds to a spectral curvature significance of @xmath44 ( see table [ table2 ] ) .
the fit parameters are @xmath45 and @xmath46 .
the overall source significance for the best - fit spectral shape and spatial morphology is thus @xmath47 , and its integrated photon flux and statistical error @xmath48 @xmath2s@xmath3 ( 0.2 - 100 gev ) . above 1 gev ,
the corresponding gamma - ray luminosity is @xmath4 erg s@xmath3 for a source distance of 1 kpc , about 40% higher than the _ fermi _ lat luminosity of the cygnus loop snr above 1 gev @xcite , one of the dimmest gamma - ray snrs detected so far , or about half the luminosity of g296.5 + 10.0 in the same energy range @xcite .
the model parameters can be seen in table [ table2 ] .
the gamma - ray spectral energy distribution ( sed ) for the observed excess emission coincident with hb9 is obtained using the radio template to describe the morphology with a maximum likelihood analysis for 15 equally spaced logarithmic intervals from 0.2 to 100 gev . if the significance of a detection within a bin is less than @xmath49 a 95% confidence level upper limit for the flux is derived .
two sources of systematic errors are considered for each energy bin : the uncertainty of the galactic diffuse emission which is estimated by varying the best - fit value of the normalization of the galactic level by @xmath50 ( abdo et al .
2009a ) and the uncertainty in the effective area , which is energy - dependent and given by @xmath51 at 100 mev , @xmath52 at 560 mev and @xmath53 at 10 gev @xcite .
[ fig3 ] shows the seds with the data points obtained and statistical and systematic errors added in quadrature .
some pulsars are sources of gamma - ray emission ( e.g. , * ? ? ?
however , no point - like emission is seen at the position of the pulsar psr b0458 + 46 , as seen in fig .
[ fig1 ] , so this possibility for the origin of the emission is discarded here .
the extragalactic compact radio source 0503 + 467 lies at the boundary of the eastern edge of the radio image of hb9 .
radio galaxies are also known gamma - ray sources in the _ fermi _ band , and some of them show variability .
a lightcurve was obtained here for a point source at the position of 0503 + 467 performing a binned likelihood analysis for events divided in time bins of @xmath54 months each ( for such a time bin , the significance of the source is typically just above 3@xmath0 ) .
no signs of flux variability in these time scales were found . a more detailed analysis , which is beyond the scope of this study ,
should be carried out in order to confirm or discard the association of part of the gamma - ray emission seen in the region of hb9 to compact sources in the field . in
what follows , an association to hb9 is assumed based on the likelihood results that show the emission is best described by an extended morphology .
the radio sed is obtained from the literature @xcite .
the radio spectrum can be described by synchrotron emission from a power - law energy ( @xmath55 ) distribution of electrons varying as @xmath56 .
the distribution extends up to a maximum electron energy , @xmath57 , and the particles are uniformly distributed across the volume of the radio shell of the snr ( which is approximated as a sphere of radius 18 pc ) . the measured x - ray ( 0.1 - 2.5 kev ) flux from hb9 @xcite , emitted by the hot gas inside the shell , is used as an upper limit for the synchrotron emission in this energy band .
gamma - ray emission mechanisms include ic up - scattering of cosmic microwave background ( cmb ) photons , bremsstrahlung emission from the synchrotron - emitting electrons interacting with ambient plasma ( assumed to consist only of thermal electrons , protons and ionized helium with standard abundances and proton density @xmath58 , see * ? ? ?
* ) and neutral pion decay produced in interactions of cosmic ray protons with ambient protons ( with density @xmath58 ) .
any contribution to ic emission from other photon fields such as infrared emission from dust , or interstellar radiation fields , is not included in this model .
hadronic gamma - ray emission is calculated according to aharonian & atoyan @xcite for a particle density distribution in the form of a power - law in _ momentum _ ( @xmath59 ) of the form @xmath60 , as predicted by the theory of diffusive shock acceleration .
note that a pure power - law in momentum results in a slightly `` curved '' particle flux in kinetic energy ( @xmath61 ) of the form @xmath62 @xcite , which is more significant at lower particle energies but should be taken into account in the calculations if one works with a distribution in kinetic energies .
the gamma - ray emission can be explained with a simple power - law electron energy distribution , as can be seen in fig .
the electron distribution extends up to @xmath63 tev .
the magnetic field and total electron energy are given , respectively , by 8 @xmath64 g and @xmath65 erg , or 2.7% of the total explosion energy of @xmath66 erg .
previous x - ray observations @xcite imply low ambient densities . the model shown in fig .
[ fig3 ] for this scenario shows the corresponding bremsstrahlung flux with an ambient density @xmath67 @xmath68 . since the gamma - ray spectrum steepens above @xmath101 - 2 gev , a proton spectrum with a power - law momentum distribution with index @xmath69 is able to fit the sed .
however , such a particle distribution is rather steep , and a more physically plausible scenario is obtained with a particle distribution of the form @xmath70 exp[@xmath71 , which is also shown in fig .
the break in the particle momentum distribution is required by the observations . a lower index below the break of 2 , like that of electrons , results in gamma - ray emission that is too hard to account for the _ fermi _ lat fluxes .
the total cosmic ray energy , @xmath72 , and ambient density in this case are given by @xmath73 erg .
if , for example , @xmath1020% of the available energy is in these cosmic rays , the required density is @xmath741 @xmath68 .
this scenario is shown in fig .
[ fig3 ] , where the leptonic component is calculated with a magnetic field of 40 @xmath64 g , a power - law electron distribution with index 2 extending up to 460 gev and a total electron energy of @xmath75 erg .
an ambient density of 1 @xmath68 is used for the bremsstrahlung flux also . the gamma - ray flux upper limits constrain the maximum electron energy in this scenario , which should not exceed 12 gev .
together with an ambient density of 1 @xmath68 , a magnetic field of 6 @xmath64 g and total electron energy content of @xmath76 erg , the modeled sed is also shown in fig .
excess gamma - ray emission in the region of the large snr hb9 is revealed by analysis of cumulative observations by _ fermi _ lat .
the morphology of the gamma - rays is best described by a the radio morphology of the source ( see table [ table1 ] ) . in the energy band
explored ( above 0.2 gev ) the photon spectrum is curved and can be best described by a log - parabola of the form @xmath77 .
the extension of the snr offers an exciting opportunity to carry out studies on spatially - dependent particle acceleration .
for example , a spatial study of the radio emission in the band 408 - 1420 mhz has shown that synchrotron emission from the center of the shell is softer than that from the outer shell @xcite . however , the current statistics of the gamma - ray observation is too low to draw definitive conclusions regarding any possible variation of the flux level or of the spectral parameters across the source .
the gamma - ray spectrum is consistent with an inverse compton scenario where cmb photons are up - scattered by the synchrotron - emitting electrons responsible for the radio emission .
the physical parameters derived are reasonable ( magnetic field 8 @xmath64 g , total electron energy of 2.7% of the available energy ) and consistent with observations in the x - ray band ( low ambient density , @xmath78 @xmath68 ) .
the gamma - ray data observed above 0.2 gev corresponds , in this scenario , to the high - energy tail of the ic emission .
the electron distribution that is required to account for the lat observation is a simple power - law in particle energy with an index of 2 and maximum electron energy of 0.5 tev .
the maximum energy in the electron spectrum contains information on the details of the acceleration process . in the absence of escape
, the maximum energy can be determined by radiative losses after a sufficiently long time .
synchrotron losses will dominate over ic - cmb losses for magnetic fields above @xmath103.27 @xmath64 g , for which the magnetic energy density is the same as the cmb energy density @xcite . the maximum electron energy in this case is then given by @xcite @xmath79 , where @xmath80 is the upstream magnetic field in gauss
, @xmath81 is the shock speed in units of @xmath82 km / s , @xmath83 is the ratio of the particle mean free path to its gyroradius ( also a measure of the amount of magnetic turbulence ) , and @xmath84 is a factor depending on the shock obliquity , the shock compression ratio and @xmath83 . for a snr age of 8 kyr ,
explosion energy of @xmath85 erg and ism density of 0.1 @xmath68 , the sedov - taylor solution gives a shock speed @xmath86 .
if the ism magnetic field is @xmath87 g then , for a parallel shock with @xmath88 , the maximum electron energy found in this work through a fit of the gamma - ray sed in the ic - dominated scenario gives @xmath89 , which would imply very low resonant magnetic turbulence ( @xmath90 ) .
a hadronic interpretation might seem less likely due to the low ambient densities observed .
the required mean density across the snr is around 1 @xmath68 for a total cosmic ray energy of 20% of the available energy of @xmath91 erg ( or higher density for a corresponding lower cosmic ray energy content ) . a power - law with exponential cutoff momentum distribution with index @xmath92 and cutoff momentum 80 gev / c for the protons can account for the gamma - ray sed .
in the bremsstrahlung scenario , a similar value for the density is used ( 1 @xmath68 ) .
the electron distribution in this case must extend up to a particle energy of only 12 gev , which is close to the lower limit for the maximum electron energy required to explain the radio data . extending
the radio observations of this object above 3 ghz could discard this scenario if the same power - law synchrotron emission is present at higher radio frequencies .
the high density requirement can be relaxed by increasing the total energy in the electrons and lowering the magnetic field ( the value used is 6 @xmath64 g ) .
however , the magnetic field is likely not lower than at least the typical galactic value ( @xmath103 @xmath64 g ) . for both the hadronic and bremsstrahlung scenarios
the density requirements could be easily achieved if the snr interacts with clumps of high density material in its environment .
new observations should be carried out to confirm or discard this possibility .
finally , it is noted that a better fit to the sed might be obtained with more complex scenarios such as those involving several particle populations .
mixed scenarios involving comparable fluxes from leptonic and hadronic emission mechanisms are also certainly plausible and have not been explored here . this research has made use of nasa s astrophysical data system and of the simbad database . financial support from universidad de costa rica , through research grant number 829-b2 - 171 from vicerrectora de investigacin , and from the escuela de fsica , is acknowledged .
comments made by the anonymous referee were highly appreciated and helped improved the quality of this work substantially . 1 abdo , a. a. , et al .
2009a , apj , 706 , l1 abdo , a. a. , et al .
2009b , physical review letters , 103 , 251101 abdo , a. a. , et al .
, 2010a , apj , 722 , l1303 abdo , a. a. , et al . , 2010b , apj , 718 , 348 abdo , a. a. , et al .
, 2010c , apj , 710 , l92 abdo , a. a. , et al .
, 2013 , astrophysical journal supplement 208 , 17 ackermann , m. , et al .
2013 , science , 339 ( 6121 ) , 807 aharonian , f. a. & atoyan , a. m. 2000 , a&a , 362 , 937 ajello , m. , et al .
2012 , apj , 744 , 80 araya , m. & cui , w. 2010 , apj , 720 , 20 araya , m. 2013 , mnras , 434 , 2202 atwood , w. b. , et al .
( the fermi lat collaboration ) , 2009 , apj , 697 , 1071 bell , a. r. , 1978 , mon . not .
soc . , 182 , 147 berezhko , e. g. , & vlk , h. j. 2004 , a&a , 419 , l27 berezhko , e. g. , ksenofontov , l. t. , & vlk , h. j. 2002 , a&a , 395 , 943 blandford , r. d. , & eichler , d. , 1987 , physics reports , 154 , no . 1
, 1 - 75 gaisser , t. k. , protheroe , r. j. & stanev , t. 1998 , apj , 492 , 219 caprioli , d. 2012 , jcap , 07 , 038 condon , j. j. , broderick , j. j. , seielstad , g. a. , douglas , k. , gregory , p. c. 1994 , the astronomical journal , 107(5 ) , 1829 dermer , c. d. 2012 , physical review letters , 109 , 091101 dwarakanath , k. s. , shevgaonkar , r. k. & sastry , ch .
v. 1982 , j. astrophys .
astr . , 3 , 207 ellison , d. c. , patnaude , d. j. , slane , p. , & raymond , j. 2010 , apj , 712 , 287 giordano , f. , naumann - godo , m. , ballet , j. , bechtol , k. , funk , s. , lande , j. , mazziotta , m. n. , rain , s. , tanaka , t. , tibolla , o. , & uchiyama , y. 2012 , apj , 744 , l2 gotthelf , e. v. , koralesky , b. , rudnick , l. , jones , t. w. , hwang , u. , petre , r. 2001 , apj , 552 , l39 hwang , u. , decourchelle , a. , holt , s. s. , petre , r. , 2002 , apj , 581 , 1101 katagiri , h. , tibaldo , l. , ballet , j. , giordano , f. , grenier , i. a. , porter , t. a. , roth , m. , tibolla , o. , uchiyama , y. & yamazaki , r. 2011 , apj katsuta , j. , uchiyama , y. , tanaka , t. , tajima , h. , bechtol , k. , funk , s. , lande , j. , ballet , j. , hanabata , y. , lemoine - goumard , m. & takahashi , t. , 2012 , apj , 752 , 135 kothes , r. , fedotov , k. , foster , t. j. & b. uyaniker , b. 2006 , a&a , 457 , 1081 leahy , d. & roger , r. 1996 , a&a ss , 115 , 345 leahy , d. & aschenbach , b. 1995 , a&a , 293 , 853 leahy , d. & roger , r. 1991 , astron . j. , 101(3 ) , 1033 leahy , d. & tian , w. 2007 a&a , 461 , 1013 long , k. s. , reynolds , s. p. , raymond , j. c. , winkler , p. f. , dyer , k. k. , & petre , r. 2003 , apj , 586 , 1162 nolan , p. l. , et al .
2012 , apjs , 199 , 31 mattox , j. r. , et al .
, 1996 , apj , 461 , 396 protassov , r. , van dyk , d. a. , connors , a. , kashyap , v. l. , & siemiginowska , a. 2002 , apj , 571 , 545 reich , w. , zhang , x. & frst , e. 2003 , a&a , 408 , 961 reynolds , s. p. 1998 , apj , 493 , 375 rho , j. , dyer , k.k . , borkowski , k.j .
, reynolds , s.p . , 2002 , apj , 581 , 1116 spangler , s. r. , mutel , r. l. , benson , j. m. & cordes , , j. m. 1986 , apj , 301 , 312 spangler , s. r. , fey , a. l. & cordes , , j. m. 1987 , apj , 322 , 909 tanaka , t. , et al .
2011 , apj , 740 , l51 vink , j. & laming , j. m. 2003 , apj , 584 , 758 white , r. & long , k. 1991 , apj , 373 , 543 xing , y. , wang , z. , zhang , x. & chen , y. 2014 , apj , 781 , 64 yuan , q. , huang , x. , liu , s. & zhang , b. 2014 , apj , 785 , l22 yuan , q. , liu , s. , fan , z. , bi , x. , & fryer , c. l. 2011 , apj , 735 , 120 ( @xmath2s@xmath3 ) & parameter values & ts + @xmath94 & @xmath95 & @xmath96 & 230 + & & & + @xmath97 & @xmath98 & @xmath99 & 255 + & & @xmath100 & + & & & + @xmath101 & @xmath48 & @xmath102 & 261 + & & @xmath103 & + ( similar to the lat psf above 1 gev ) .
the ` x ' marks the position of the source 2fgl j0503.2 + 4643 , which is not included in the model for hb9 , the cross is the position of the pulsar psr b0458 + 46 and the circle shows the position of the source 2fgl j0503.3 + 4517 , associated with a background galaxy which is included in the model .
the green contours are obtained from a gb6 4850 mhz image smoothed using a gaussian kernel with @xmath104 ; contours are at 0.6 , 1.2 , 1.8 and 2.4 kjy / sr .
the magenta contours are obtained from a rosat image smoothed using a gaussian kernel with @xmath105 ; contour levels at 0.0001 , 0.0002 and 0.0003 cts / s / pixel.[fig1],width=472,height=340 ] | a 5.5-year _ fermi _ lat gamma - ray observation shows significant extended emission at the position of the supernova remnant hb9 ( g160.9 + 2.6 ) .
the significance of the detection above the background for photon energies above 0.2 gev is 16@xmath0 .
the gamma - ray flux above 0.2 gev is @xmath1 photons @xmath2s@xmath3 , and the corresponding luminosity above 1 gev is @xmath4 erg s@xmath3 ( for a source distance of 1 kpc ) .
the spectrum of the source is best described by curved power - law ( log - parabola , @xmath5 with @xmath6 and @xmath7 ) .
the gamma - ray spectrum of the source is consistent with both leptonic and hadronic models , and the relevant physical parameters in each case are derived .
more studies on the ambient density in the region of hb9 should be carried out to rule out or confirm hadronic and non - thermal bremsstrahlung scenarios for the gamma - ray emission . |
galaxy collisions and interactions play an important role in the local universe in driving powerful star formation activity in galaxies , and the generation of powerful emission in the infrared @xcite .
multi - wavelength studies of local galaxy populations have shown that nearly all ultraluminous infrared galaxies ( ulirgs with l@xmath10 @xmath9 10@xmath11 l@xmath6 ) involve mergers of gas - rich galaxy pairs @xcite .
similarly , nearby luminous ir galaxies ( lirgs with l@xmath10 @xmath9 10@xmath12 l@xmath6 , see * ? ? ?
* ) , often involve strongly interacting and sometimes merging systems .
numerical models @xcite have emphasized the importance of strong tidal torques , which can act to drive gas into the centers of galaxies , initiating nuclear enhancements in star formation . despite two decades of study of these galaxies , there are still many uncertainties about how to realistically model the complex astrophysical environment of a powerful major merger @xcite .
many questions remain .
for example , what is the the dependence of the star formation triggering on the physical conditions in the gas , especially the effects of feedback from young stars , active galactic nuclei ( agn ) winds , or other processes ( see * ? ? ?
do all deeply penetrating gas - rich mergers inevitably lead to powerful star formation and luminous infrared activity ?
there are examples of major mergers where the gas has been disturbed significantly , but is still in a pre-(u)lirg phase .
the ` antennae ' system ( ngc 4038/39 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) is only just classified as a lirg , and yet clearly contains large amounts of gas in the overlap region which have not fully developed their star forming potential .
clues to this are seen in the recent high - resolution study of co(32 ) with alma by @xcite , where peculiar velocities and compact linear filaments of co gas are seen : likely the result of turbulence . in the current paper
, we discuss another possible example of a galaxy pair , the taffy ( total l@xmath13 = 4.5 @xmath1 10@xmath14 l@xmath6 ) , that has recently experienced a head - on collision but has yet to display strong global star formation enhancements .
the taffy galaxies , ugc 12914 and ugc 12915 ( vv254 , kpg603 ) were discovered to be connected by a bridge of radio continuum emission ( figure 1a ) and neutral hydrogen gas @xcite .
it was suggested that the two galaxies have recently interpenetrated , creating a common magnetic field between them as they moved apart after the collision . in this picture ,
cosmic rays ( crs ) generated within the galaxies , would flow into the common halo generating radio synchrotron radiation .
@xcite showed that the logarithmic ratio of the far - infrared to radio continuum flux @xmath15 is 1.94 , unusually low for normal galaxies ( @xmath16 = 2.3 ) .
@xcite suggested that _ in situ _ shock acceleration of cosmic rays in the bridge can explain the strength of the radio emission and radio spectral index .
in this view , the galaxies do not need to be the source of cr electrons , because the crs can be accelerated within the highly turbulent conditions in the bridge @xcite .
further evidence that shocks may play an important role in the material in the bridge was presented by @xcite , who showed that the bridge contains @xmath9 4.5 @xmath1 10@xmath17 m@xmath6 of warm ( 150 @xmath18 t @xmath18 175 k ) molecular hydrogen .
the excitation of this warm intergalactic gas was found to have a strong heating contribution from shocks and turbulence .
the detection of broad co ( 10 ) lines in the bridge @xcite is consistent with turbulent gas and kinetic energy dissipation in this colder component .
this could provide a significant fraction of the heating needed to maintain the warm h@xmath7 for the @xmath825 myrs since the collision @xcite .
while @xcite were able to demonstrate that the warm h@xmath7 emission was likely the result of turbulent or shock heating of the bridge , they could not rule out heating of the molecular gas by x - rays , since , at that time , no observations of this kind had been made .
one of the consequences of a head - on collision between the two galaxies is the possibility of the sweeping of gas out of the galaxies .
the large amounts of gas found in the bridge ( @xmath87 @xmath1 10@xmath19 m@xmath6 , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) has to survive the violence of the collision @xcite .
nevertheless , in the gas and stellar dynamical modeling of the taffy collision by @xcite , it was shown that momentum transfer between the gas clouds in the galaxies during the collision was sufficient to transport large quantities of gas into the bridge .
the model also provided a possible explanation for the spatial offset in the distribution of the dense molecular gas , and the diffuse hi in the bridge @xcite .
this was explained as a consequence of the two galaxies colliding slightly off - center , leading to an asymmetry in the density structure of the resulting bridge .
the models did not include the thermal balance of the different phases of the interstellar medium ( ism ) in the galaxies , which is especially relevant to x - ray observations .
hot x - ray emitting gas can extend outside galaxy disks through both internal and external processes .
in starburst galaxies , supernova explosions can create hot bubbles of x - ray emitting gas that can break - out of their host molecular disks , and push gas into the halo @xcite .
large scale shocks can also heat gas to x - ray temperatures in the intergalactic diffuse environment . in the 40 kpc - long filament in the stephan s quintet galaxy group ,
x - ray emission was detected in association with a giant shock structure @xcite .
the filament also contains copious amounts of warm and cold molecular gas @xcite .
the emission is consistent with the model of @xcite in which a large - scale shock is propagated into a multi - phase medium . in this model ,
the low density regions of this medium become heated to x - ray temperatures , whereas the higher density clumps collapse and can survive the initial disturbance and radiate strongly in the mid - ir h@xmath7 lines .
shocked regions can also radiate strongly in diffuse [ cii ] emission , as observed @xcite .
similarly powerful mid - ir h@xmath7 , [ cii ] , and [ oi ] emission have been seen in the taffy system ( peterson & appleton , in preparation ) , providing further motivation for the current study of the taffy system in x - rays . in this paper
, we detect faint soft diffuse x - rays from the taffy galaxy disks and bridge region , as well as from a number of luminous discrete sources .
we explore various explanations for the observed x - ray emission , favoring shock heating as the most likely explanation for the diffuse emission .
we assume a distance to the galaxies of 62 mpc based on the heliocentric velocities for ugc 12914 and ugc 12915 of @xmath20 and @xmath21 km
s@xmath3 respectively , and a hubble constant of 70 km s@xmath3mpc@xmath3 .
the taffy galaxies were observed for 39.5ks on 2013 february 3 ( obsid 14898 ; pi p. appleton ) with the back - illuminated ccd chip , s3 , of the _ chandra _ advanced ccd imaging spectrometer ( acis ; * ? ? ? * ) .
we processed the observation using ciao version 4.6 ( calibration database 4.6.4 ) to create a level2 events file , following the software threads from the _ chandra _ x - ray center ( cxc ) .
most of the images , covering a 4.1 x 4.1 arcminute@xmath22 field with a pixel scale of 0.492 arcsec pixel@xmath3 are shown for the @xmath23@xmath24kev range , but we also examined smaller energy ranges , including soft ( @xmath23@xmath25kev and @xmath23@xmath26kev ) , medium ( @xmath25@xmath26kev ) , and hard ( @xmath26@xmath24kev ) bands .
we created smoothed images in two ways .
first , we smoothed with a simple @xmath27 gaussian kernel with the saoimager ds9 @xcite .
second , we used the ciao task dmimgadapt to adaptively smooth images .
dmimgadapt smooths each pixel on the scales necessary to reach the desired count threshold ( selected to be 10 counts ) under its convolution kernel or the maximum kernel size , which we set to 5@xmath28 .
we also created images of the diffuse emission by removing point sources , dividing by the exposure map , and then adaptively smoothing the cleaned image ( section 3.2 ) .
point source identification was done using the ciao task wavdetect @xcite , a mexican - hat wavelet source detection program .
these sources , except for the plausible centers of the galaxies , were then cleaned using the task dmfilth , which replaces source pixel values with values determined by interpolating the surrounding background .
we used the specextract task to extract x - ray spectra in the 0.38.0kev range for several extraction apertures in the taffy system , including the two galaxies and the bridge , after excluding the point sources .
a background spectrum was simultaneously extracted in the same chip in a sourceless region , and automatically scaled based on the ratio of the source - to - background areas .
we filtered the events based on the energy range and then grouped it to a minimum of 20 counts per bin prior to modeling the spectrum .
for the two bridge sub - regions as well as the extragalactic hii complex ( see section 3 ) , we estimate the flux and luminosities based on the count rates , assuming the model parameters from the total bridge . the _ spitzer _
24@xmath29 m observations shown in this paper were obtained from the spitzer heritage archive ( pi : j. r. houck ; observid 21 ) and had been processed through the s18 ssc pipeline .
a constant background was removed from the image before extracting fluxes from the extended regions .
the _ herschel _ photometry data at 70 , 100 and 160@xmath29 m was de - archived from the herschel science archive ( pi e. sturm , proposal name kpgt@xmath30esturm@xmath301 , obsids 1342212798/9 ) as level 1 data , which are fully calibrated frames . these scan and cross - scan data were then further processed through the mapping package scanamorphos ( roussel 2013 ) , kindly made available to us by the author . additionally , we obtained _ herschel _ [ cii]157.7@xmath29 m line data as part of our ot1 program pacs spectroscopy program ( pi : p. appleton , proposal ot1@xmath30pappleto@xmath301 , obsid 1342238415 ) .
these data were processed on a large virtual machine through a version of hipe 13 @xcite available at the nasa herschel science center .
the integrated line map shown here was created in hipe after first removing baselines from all the individual spaxels in the image , and then integrating over the range of velocities in which signal was detected .
full details of the reduction of the [ cii ] data will be provided in a separate paper devoted to the pacs spectroscopy of the taffy system ( peterson & appleton , in preparation ) .
we show in figure 1b and 1c two different representations of the observations . in figure 1b , we show the compact x - ray structure present in the 0.58 kev image of these data .
we detect nine bright point - like sources at high significance above background , and numerous fainter ones .
an additional extended source , is coincident with the peak of the radio continuum emission in the nucleus of the south - western galaxy ugc 12914 ( and 2mass k - band emission - not shown ) .
in contrast , the north - eastern galaxy , ugc 12915 , contains several point - like x - ray sources , but none of them seem to be directly associated with the galaxy s radio continuum and near - ir light peak .
1c shows clearly the detection of extended emission in the bridge region , but it is somewhat offset from the radio continuum bridge to the north - west . to quantitatively describe the x - ray properties of the various features seen in the x - ray images , we define in figure 2a several extended polygon - shaped regions , as well as identifying the point sources .
for the extended regions , polygons define both galaxies and the bridge , which is divided into two parts .
the x - ray properties of the galaxies are given in table 1 , along with coordinates , aperture description , net counts ( after taking into account the background model ) , and total x - ray flux ( 0.58 kev ) and luminosity .
the luminosities are derived from model fits to the x - ray spectra ( section 3.3 ) .
the x - ray luminosities ( 0.58 kev ) of ugc 12914 and ucg 12915 ( excluding the point sources ) are 1.66 and 1.15 @xmath110@xmath31 erg s@xmath3 respectively .
if we add the contribution of the compact sources within their disks ( see next section ) , ugc 12914 and ugc 12915 have total luminosities of 3.2 and 2.2 @xmath110@xmath31 erg s@xmath3 respectively .
therefore , the compact sources contribute significantly to the total x - ray luminosities of the galaxies . in table [ soft ] , we present the observed soft ( 0.52 kev ) x - ray fluxes for the extended polygon regions .
we measure a total x - ray luminosity over the the full bridge of l@xmath32 = 4.7 @xmath1 10@xmath2 ergs s@xmath3 .
most of the soft x - ray emission from the bridge comes from region bridge-1 , with a surface x - ray brightness of 5.7 @xmath1 10@xmath33 erg s@xmath3 @xmath34 arcsec@xmath35 .
this is only a little less that the average surface brightness of the soft x - ray emission from the two galaxies .
the region called `` hii '' , which we consider separately from the bridge ( figure 2a ) , is dominated by the compact source b1 , but may also include some extended emission . figure
2a and figure 2b shows two kinds of labeled point sources .
those with yellow circles e1-e5 ( in ugc 12915 ) , w1-w3 ( in ugc 12914 ) , and b1 in the bridge , are all detected at a 3-sigma level of significance .
the level of detection significance is based on poisson noise estimates , taking into account the estimated ( generally low ) background counts detected in addition to the sources .
these sources , which are strongly detected , are listed individually in table 1 .
we further identify 15 sources ( labelled as red circles in figure 3 ) which are detected at a lower level of significance ( 2 - 3@xmath36 ) .
these sources have net counts in the 1.5 - 2 arcsecs extraction apertures of between 4.5 and 10 .
formally they represent significant detections , but their x - ray luminosities are poorly determined because of poisson noise .
we list the properties of these sources in table [ compact ] .
how many of the fainter sources would we expect in the area of the bridge ( @xmath81 arcmin@xmath35 ) by chance ? based on an exploration of larger adjacent fields on the chip , and by simulating an additional component representing the added ( randomly distributed ) count rate from the diffuse taffy - bridge emission , we would expect 0.3 sources arcmin@xmath35 with a count rate @xmath9 4 in a 1.5 arcsec extraction aperture .
this number is consistent with the expected number of agn ( 0.5 sources arcmin@xmath35 ) based on a detection threshold of close to 1.5 @xmath1 10@xmath37 erg s@xmath3 @xmath34 @xcite . considering the larger area of the whole taffy system
, we conclude that no more than one of the fainter sources is a chance association : most appear to be real detections . to investigate how their properties might influence our estimate of the extended soft x - ray emission from the bridge , we identified them by eye and flagged them for removal in the smoothed images .
as can be seen by comparing figure 2c with figure 2d , the effects of the removal of these additional sources is small , reducing the net flux in the soft band by only an additional 20@xmath4 . in what follows , we present results based on the extended emission with only the more reliable 3@xmath36 sources removed .
we estimate the flux of each 3@xmath36-detected source assuming a power law with a photon index of @xmath38=1.7 @xcite .
their luminosities ( table 1 ) put them in the range of ultra - luminous x - ray sources ( ulxs ; commonly associated with star forming galaxies * ? ? ?
* ; * ? ? ?
such sources are usually assumed to be massive accreting black - holes ( m@xmath39 @xmath40 20 m@xmath6 ) associated with young star clusters ( see * ? ? ?
* ; * ? ? ?
this is consistent with the majority of the compact sources being associated with the disks of both galaxies , as shown in figure 3 .
several are concentrated around the inner region of ugc 12915 , and some of the fainter sources seem loosely associated with the bright spiral arm in ugc 12914 , seen in visible light .
the numbers of brighter sources are also consistent with the average number of ulx s as a function of star formation rate ( sfr ) .
@xcite estimate that for gas - rich galaxies , there is , on average , 0.5 ulxs for every 1 m@xmath6 yr@xmath3 of sfr .
based on the sfrs for the two galaxies discussed in section 5.1 , we estimate that ugc 12914 and ugc 12915 should contain 2 and 5 ulxs ( l@xmath41 @xmath9 10@xmath2 erg s@xmath3 ) , which is close to the actual number of 3@xmath36 sources found in the two disks . unlike @xcite , who studied a sample of interacting galaxies , we do not see a significant enhancement in ulxs compared with normal spirals in this system .
the brightest ulx source ( b1 ) lies outside the disk , to the south - west of ugc 12915 , and has an x - ray luminosity of l@xmath0 = 3.0 @xmath110@xmath2 erg s@xmath3 .
this bridge source is associated with a massive molecular rich cloud - complex ( see * ? ? ?
* ) , and is the only region outside the galaxies detected in h@xmath42 and pa@xmath42 images @xcite .
we estimate that this extragalactic hii region complex ( labelled as `` hii '' in figure 2a ) has a sfr @xmath8 0.25 m@xmath6 yr@xmath3 ( section 5.1 of this paper ) based on its ir emission .
it also lies close to a peak in the radio emission in the bridge seen in figure 1a . as such ,
the discovery of a single ulx in this complex is interesting , as it lies outside either of the two galaxy disks and lies near the densest part of the molecular bridge ( * ? ? ?
* and braine et al . in preparation ) .
the failure to detect the source in hard x - rays probably also rules out a chance association with a background agn .
although no age estimate is available for the source ( it is highly obscured in visible and uv light ) , the escape of emission of h@xmath42 light suggests a young stellar population is present .
if this `` extragalactic hii region '' was formed by a compression of gas in the bridge , it would only have had 1020 myrs to form
. however , this is feasible , since berghea et al .
( 2013 ) have shown that the ages of typical massive star clusters associated with ulxs are typically 10 myrs old .
a more detailed study of the correlation between the position of taffy b1 and the host molecular cloud and associated star - formation would be worthwhile .
in addition to b1 , several fainter sources lie in the bridge ( b2 - 5 ) .
although we can not completely rule of a chance association with a background source , the concentration of these fainter sources is interesting .
as figure 3 shows , the sources fall in the region where faint dusty filaments are seen in the optical image suggesting heavy obscuration , and they fall well within the main region of the molecular bridge ( section 4 ) .
it is therefore possible that these sources represent fainter versions of b1 , and are associated with obscured star formation within the bridge . unlike b1 , there is no clear association of these sources with optical counterparts .
the slightly extended source `` w@xmath43 '' lies in the nucleus of ugc 12914 , and contains a significant fraction of the x - ray luminosity of the whole galaxy .
this source is the only source detected in the 28 kev band , with a hard x - ray luminosity l@xmath44 = [email protected] @xmath1 10@xmath31 erg s@xmath3 ( luminosity calculation assumes an x - ray power - law of slope @xmath38 = 1.7 ) .
previous observations @xcite of the galaxies with the _ spitzer _ infrared spectrograph ( irs ) failed to detect spectroscopic evidence of an agn from either of the taffy galaxies .
formally , it has a hardness ratio ( hr ) of [email protected] , the most positive of any source in the system ( all the others have upper limits ) , but lower than typical agn ( hardness ratios in the range 0.70.9 )
. the elevated hardness ratio of w@xmath43 may indicate that ugc 12914 contains a low luminosity agn , accreting at a sub - eddington rate @xcite .
despite having only 90 net counts , we have simulated how the hr varies over a range of power - law slopes and intrinsic hydrogen column densities . within the range allowed by a 2@xmath36 uncertainty in hr , a typical source with 1.7 @xmath18 @xmath38 @xmath18 1.9 , would require a range of possible absorbing columns of 0.3 - 1 @xmath1 10@xmath46 @xmath34 .
this is consistent with an agn with moderate intrinsic absorption .
however , given the small number of detected counts at higher - energy , we can not completely rule out a highly - obscured , compton - thick agn ( e.g. * ? ? ?
although such a source in isolation would exhibit a much larger hardness ratio than observed , the hr might be lowered by the addition of a thermal component from extended nuclear star formation , or shocks which would soften the spectrum as the agn became more obscured . a conclusive test of whether ugc 12914 contains a compton - thick agn would require deep observations from a telescope sensitive to x - rays beyond 10 kev .
it is reasonable to ask whether the compact sources are consistent with the x - ray luminosity function ( xlf ) of populations of compact x - ray sources found in other nearby systems .
for example , @xcite performed an in - depth study of @xmath8120 compact x - ray sources discovered in the ` antennae ' system @xcite , and was able to determine the xlf down to limits of 10@xmath47 erg s@xmath3 , a factor of @xmath830 times fainter than the sources discovered in the taffy .
their results were consistent with the brighter sources ( ulxs ) being an extension to high mass of a fainter population of x - ray binaries accreting close to their eddington limits .
figure 4 shows the cumulative xlf determined for the compact sources ( cyan line bracketed by black uncertainty boundaries ) .
the luminosity function of the cumulative source number counts can be approximated by a power - law of the form n(@xmath9l ) = al@xmath48 ( e.g. * ? ? ?
n(@xmath9l ) is the number of sources with luminosity greater than l , a is a normalization factor , and @xmath42 is the slope of the assumed relationship .
following @xcite , we binned the luminosities of the 2 and 3-sigma sources in intervals of the natural unit of counts ( one count converts closely to a unit of 1.25 @xmath1 10@xmath49 ergs s@xmath3 at a distance to taffy of 62 mpc ) .
we calculated the luminosities of the fainter compact sources in the same manner as the brighter sources ( assuming a power - law spectral index of 1.7 ) .
the xlf shows a smooth transition from the brighter ( 3-sigma sources ) to the fainter ( 2-sigma ) sources , suggesting that they form a continuous population . also plotted , to provide additional information , are the differential counts as green and red squares in figure 4 .
given the small numbers of total sources ( we only have 25 in all , and only 9 above 3@xmath36excluding w@xmath43 ) , we did not attempt to fit the slope of the xlf , nor did we attempt to correct for completeness in the fainter sources .
rather we simply compare our data with results from the deeper survey of zezas et al .
( 2006 ) for the average @xmath42 = 0.55 determined for the full deep ` antennae ' survey ( purple line ) .
this slope is dominated by the many fainter sources ( l@xmath50 @xmath18 10@xmath49 erg s@xmath3 ) in the ` antennae ' , which is characteristic of the massive x - ray binary population .
it is clear from the comparison that taffy sources appear to follow a steeper slope .
however , although it might be tempting to speculate that the taffy source may be powered by something other than massive x - ray binaries , the steepness is very likely a consequence of small number statistics in our sample .
indeed the same effect was found by zezas et al .
when they fit these same ` antennae ' data with a split power law .
these authors showed that sources brighter than l@xmath50 @xmath9 7.5 @xmath1 10@xmath47 erg s@xmath3 ( this would be the case for all our sources ) could be fit with an @xmath42 = 1.35 ( red line ) .
this slope is much more consistent with the taffy sources . as zezas et al .
describe in their paper , the steeper slope is most likely an artifact poor statistical sampling at the bright end of the xlf , and the potentially rapid variability of fluxes found predominantly in the higher - luminosity sources .
given , in our case , the additional uncertainty in the fluxes of the fainter sources , we conclude that our data are consistent with the sources being the tip of the luminosity function of high - mass x - ray binaries .
for the galaxy extractions and the bridge ( the sum of extracted regions bridge-1 and bridge-2 of figure 2a ) , we have enough counts to spectrally model these regions .
spectral modeling was done using the sherpa packages of ciao , having first removed the 3@xmath36 point sources . for the diffuse emission in the galaxies and bridge , we fit combinations of apec thermal models ( astrophysical plasma emission code ; * ? ? ?
* ) and a power - law .
solar metallicity is assumed .
apec uses a model database of more than a million atomic spectral lines , plus free - free continuum emission for a hot plasma , to find the best fitting parameters consistent with these data
. a fixed galactic foreground absorption @xmath51 @xcite is assumed .
we also tried adding intrinsic absorption as a free parameter .
the results are given in table [ fits ] .
in addition to the name of the region ( column 1 ) and the model type ( column 2 ) , table [ fits ] also provides the apec normalization factor n@xmath52 ( column 3 ; proportional to the volume emissivity of the apec modeled plasma ) , thermal temperature kt ( column 4 ) , power - law flux density at 1 kev ( column 5 ) , power - law index @xmath38 ( column 6 ) , best - fitting intrinsic hi absorption column density if allowed to be non - zero ( column 7 ) , and the normalized minimum @xmath53 of the final fit ( @xmath53/n ) , where n is the number of degrees of freedom in the model . in figure 5
, we plot x - ray spectra from the observations , and over - plot the models .
the spectra of the diffuse emission in ugc 12914 and ugc 12915 ( figure 5a and b ) were best fit by a combined apec and power - law fit .
for ugc 12914 , we fitted a thermal component at 0.71(@xmath450.25)kev , and a power law model with a photon index of @xmath38 = 2.3(@xmath451.5 ) .
similarly , for ugc 12915 , the spectrum is best fit with a thermal model at 0.72(@xmath450.16)kev , and a power law with a photon index of @xmath54 ) .
we note that the differences in the normalized minimum @xmath53 for the fit , with and without a power - law , is not large .
nevertheless , we argue that inclusion of a power - law component in the modeling is physically reasonable , because an extrapolation of the xlf of the observed compact sources to fainter flux levels implies the existence of population of faint power - law sources which would appear as a diffuse component when averaged together = 10@xmath49 erg s@xmath3 and find that the expected unobserved luminosity from massive x - ray binaries is within a factor of two of the power - law component of the two galaxies measured from the spectral modeling . ] . to improve the signal to noise of the modeling ,
we have also summed the signal from both galaxies , and re - fit the resulting spectrum ( figure 5c ) . as table [ fits ] shows ,
a significantly improved fit was found for the summed galaxies .
we consider the apec+power - law fit to be the most reasonable with a new power - law index that is closer to that expected for the brighter point sources .
although formally the apec+absorption provides an equally good fit , it is less compelling because it does not include the expected compact source contributions .
the pure apec model results in a fit of lower significance .
in contrast , the spectrum of the emission from the entire bridge ( figure 5d ) is best fit with a single thermal model of 0.63(@xmath450.17)kev with some intrinsic absorption .
the amount of intrinsic absorption is plausible , given the gas - rich nature of the bridge .
fits with and without absorption are shown in figure 5d .
the inclusion of intrinsic absorption increases the 0.58kev x - ray luminosity of the bridge a factor of 2.6 . in conclusion , the temperature of the thermal component for the galaxies and the bridge are quite similar within the uncertainties , with values ranging between 0.7 and 0.8 kev without absorption , and 0.50.7kev if intrinsic absorption is present .
table [ model ] present the derived properties of the hot gas .
the mass of hot x - ray gas derived from the observations in the bridge is estimated to be @xmath80.81.3 @xmath1 10@xmath5 ( f@xmath55)@xmath56 m@xmath6 , where f@xmath55 is the filling factor of the x - rays . assumed a filling factor of unity for the x - ray emitting gas , this is 11@xmath4 of the total warm h@xmath7 mass of 9@xmath57 @xmath1 10@xmath17 m@xmath6 , extrapolated to the whole bridge by peterson et al .
( 2012 ) . by comparison ,
the mass of dust estimated from the spectral energy distribution ( sed ) for region bridge-1 + bridge-2 is 7 @xmath1 10@xmath58 m@xmath6 , or a total gas mass of 7 @xmath1 10@xmath59 m@xmath6 assuming a gas to dust ratio of 100 .
this is close to to sum of the hi and h@xmath7 masses in the bridge estimated by @xcite .
the x - ray gas therefore constitutes @xmath81@xmath4 of the total gas mass in the bridge .
as shown in figure 1b and c , the x - ray emission only correlates approximately with the radio continuum emission in the two galaxy disks , but less so in the bridge .
the radio continuum emission in ugc 12915 correlates best with the collection of compact sources and the bright nucleus of ugc 12914 . in the bridge , however , the x - ray emission appears to avoid the main radio continuum ridge and is distributed further to the north and west .
it is possible that the magnetic field is more strongly coupled to the denser molecular gas which also has a different distribution from the x - ray emitting gas ( see below ) .
the radio emission at @xmath606 cm is strongly polarized and quite regular across the whole bridge @xcite and shows no sign of faraday depolarization in the direction of the x - rays , as commonly seen in radio galaxies immersed in hot x - ray gas @xcite . in figures 6a-6c ,
we show the extended x - ray emission compared with images of dust and gas from _ spitzer _ and _ herschel_. the x - ray data have been color - coded according to energy , using adaptive smoothing after removal of the bright compact 3@xmath36 sources listed in table 1 ( see also section 3.2 ) .
figure 6a shows contours of the 24@xmath29 m emission from _ spitzer _ of the galaxies superimposed on the x - ray color image . in figure 6b and c ,
we similarly over - plot [ cii]157.7@xmath29 m line emission , and 100@xmath29 m dust emission from the _ herschel _ pacs spectroscopy and photometry data . the [ cii ] line map was made by summing over the full line profile derived from a small map made with the pacs integral field unit spectrometer ( peterson & appleton in prep . ) .
the extended soft x - rays seem more closely associated with the ir continuum emission than the brighter ridge of radio continuum .
the [ cii ] emission , which represents the diffuse neutral medium unshielded from ambient uv light , is also quite well correlated with the x - rays .
figure 7a and 7b shows the soft x - ray emission superimposed on the integrated bima co ( 10 ) molecular gas map and vla hi map respectively .
it is striking that the co ( 10 ) molecular gas in the taffy bridge seems anti - correlated with the soft x - rays , whereas the hi , like the dust and [ cii ] emission , is better correlated . in the models of @xcite , the offset between the diffuse hi and the denser co emitting gas
was explained in terms of the impact parameter in the collision .
the x - ray emission seems to follow this trend , being more correlated with the diffuse components of the bridge than the denser colder molecular gas .
another interpretation of the anti - correlation of the soft x - ray emission with the co ( 10 ) distribution may have to do with selective absorption of x - rays across the bridge by dense gas . in table
[ soft ] , we show that bridge-2 is not detected in 0.52 kev x - rays to a 3@xmath36 level .
could heavy absorption in that part of the bridge extinguish the x - ray signature ? based on the intensity of the co ( 10 ) emission from @xcite in that region , and using the unusually low value of x@xmath61 = 2.6 @xmath1 10@xmath62 @xmath34 ( k km s@xmath3)@xmath3 measured by @xcite in the bridge , we estimate that the h@xmath7 column density would be @xmath9 1 @xmath1 10@xmath63 @xmath34 in bridge-2 .
if we assumed that bridge-2 is as bright as bridge-1 in x - rays , we have performed simulations that allow us to estimate that we would need @xmath82 @xmath1 10@xmath63 @xmath34 of molecular hydrogen to extinguish the x - rays to the 3@xmath36 level observed .
it is therefore plausible that _ if x - rays occupied the bridge-2 region at a similar intensity to that observed in bridge-1 _ , then we would not detect that emission in the presence of moderate intrinsic absorption .
if true , then the total x - ray emission from the entire bridge could be a factor of 2 larger than observed .
higher resolution co observations would be needed to determine more precisely how clumpy the co emission is , relative to the x - rays , and whether this explanation is viable , since our estimates of the h@xmath7 column are based on low - resolution ( 9.8 arcsecs ) bima observations of @xcite .
an important missing piece of information is the filling factor , between the dense molecular and the x - ray emitting gas .
the above argument assumes a 100@xmath4 filling factor , which is probably not realistic because it would imply far more molecular gas than is observed filling factor with a column density of 2 @xmath1 10@xmath63 cm @xmath35 implies a molecular mass in the filament(1 x 0.3 arcmin@xmath22 ) of 1.2 @xmath1 10@xmath14 m@xmath6 .
this is approximately five times the estimated h@xmath7 mass in the bridge ( warm and cold h@xmath7 , * ? ? ?
* ; * ? ? ?
* ) , and a factor of two larger than the total gas mass , including hi . a lower filling factor would be required to bring these into approximate agreement , and this would significantly reduce the effect of the absorption of the x - rays . ] .
the existence of soft x - rays in galaxies is a natural consequence of supernovae and winds from regions of star formation which create the hot thermal component which pervades the ism @xcite . with advances in x - ray imaging
, it was soon discovered that this hot gas can escape into the haloes of galaxies , especially in starburst systems where ionized gas and x - rays were seen along the minor axis of such galaxies @xcite .
the vertical extent , surface brightness , and other properties of x - ray gas in edge - on normal and starburst galaxies was studied in detail by @xcite .
these studies were able to explore the relationship between disk and halo x - ray gas and properties , and found that , in many cases , the vertical distribution of x - ray gas in these disk systems was an exponential with scale - heights of typically 24 kpc . in low star - forming systems ( with sfrs @xmath18 0.5 m@xmath6 yr@xmath3 ) , gas was not found in the haloes , whereas in galaxies with significantly higher star formation activity ( 3 @xmath18 sfr ( m@xmath6 yr@xmath3 ) @xmath18 10 ) , all had x - ray haloes .
the surface brightness of the halos was typically a factor of 10 lower than in the disks , as was the local ratio of x - ray to far - ir luminosity in the haloes . with the advent of multi - wavelength archival data from the uv to the far ir , we are now in a position to accurately measure the star formation rates of the taffy galaxies and bridge .
we can then ask how normal is the observed x - ray emission in the galaxies , and specifically whether star formation alone could explain the existence of soft x - ray emission in the bridge ? to measure the star formation rates for the galaxies and the bridge region , we use an sed modeling code magphys @xcite .
this code fits the observed seds with sets of model templates , and estimates the likelihood distributions of physical parameters of galaxies , including sfrs , stellar mass , and ir luminosities .
the seds were constructed for the two galaxies and bridge region ( defined in figure 2 ) , utilizing all the available multi - wavelength imaging from uv ( _ galex _ ) to sub - mm ( jcmt ) data available from the archives .
figure 8 presents these seds for the two galaxies and for bridge-1 ( of figure 2 ) , along with fits using magphys .
the derived properties of the two galaxies and the bridge regions , which are well fit by the models , are presented in table [ ir ] .
the taffy galaxies were previously known to be relatively weak ir emitters @xcite compared with other interacting systems .
the pair was detected ( but barely resolved ) by iras , and @xcite quote a far - ir luminosity of 6.5 @xmath1 10@xmath14 l@xmath6 .
our results , using the full sed , are consistent with this .
we find that ugc 12914 and ugc 12915 have far - ir luminosities of 1.0 and 3.5 @xmath1 10@xmath14 l@xmath6 , and global sfrs of 1.05@xmath64 and 2.6@xmath65 m@xmath6 yr@xmath3 respectively , where the uncertainty is derived from the range of allowable values found by fitting different templates in magphys .
@xcite estimate the star formation rate for the entire system to be @xmath824 m@xmath6 yr@xmath3 , which is very similar to our values .
yr@xmath3 for ucg 12914/5 respectively ) obtained by @xcite using pa@xmath42 emission - line fluxes .
their measurements assume a very high optical extinction , and it is possible that the measured pa@xmath42 fluxes are biased upwards by shock excitation of the gas in the disks ( see * ? ? ? * ) .
indeed we find that the values quoted by @xcite do not follow the usual pa@xmath42 to 24@xmath29 m relationship observed by @xcite , perhaps supporting the idea that the pa@xmath42 is not all coming from star formation . ]
it is known that the x - ray luminosity of a galaxy is correlated with the star formation rate @xcite .
measuring the star formation rates of the taffy galaxies using the soft x - ray luminosity calibrated for normal galaxies by @xcite of l@xmath66 erg s@xmath3 = 8.3 @xmath1 10@xmath49 @xmath1 sfr m@xmath6 yr@xmath3 , yields much higher sfrs than those determined from the sed / ir methods .
based on the apec luminosities in the 0.5 - 2kev range , we estimate sfr@xmath67 = 10@xmath68 and 4.7@xmath69 m@xmath6 yr@xmath3 respectively for ugc 12914 and ugc 12915 . these rates ( even taking into account the large uncertainty of 0.34 dex in the relationship ) are not only higher than the sed derived sfrs , but if taken at face value , imply the sfr is larger in ugc 12914 than ugc 12915 .
this is contrary to all the other previous published work .
we believe that the most likely explanation for the discrepancy is that unlike normal galaxies , the taffy galaxies contain significant quantities of gas shock - heated in the collision , rather than all originating from supernova heating .
we already noted the odd distribution of hot gas in ugc 12914 , suggesting that much of the emission may be from bridge - gas , and this may explain why this system is more luminous in soft x - rays than its companion .
this picture is supported by the compact sources ( section 3.2 ) , where the number of very luminous x - ray point sources is larger in ugc 12915 than ugc12914 , suggesting that the former , not the latter galaxy , is the more prolific of the pair(@xmath9 10@xmath49 erg s@xmath3 ) @xmath8 3.22 sfr ( m@xmath6 yr@xmath3 ) from @xcite , we find sfrs of [email protected] and [email protected] m@xmath6 yr@xmath3 for ugc 12914 and ugc 12915 respectively .
however , within the considerable uncertainties in this relationship , the higher sfr for ugc 12914 over than of ugc 12915 derived from fainter sources is not significant . ] . adopting our sed
derived star formation rates , we compare the results with other systems that have supernovae - driven outflows . only ucg 12915 has a global star formation rate comparable with galaxies like ngc 4631 and ngc 3628 , which have known x - ray halos @xcite . in figure 9 , we show two cross - sections through the taffy system approximately perpendicular to the major axes of both galaxies .
section a in figure 9 , passes through the nuclear regions of both galaxies , whereas section b is offset from the centers but passes through the brightest parts of the bridge .
these cross - sections show that the soft x - ray emission has a complex distribution , and does not follow a simple exponential profile expected if both galaxies had winds .
ugc 12915 , the galaxy one might expect to have a starburst wind , has a distribution that is strongly biased towards the bridge region in both cross - sections . in section
a , ugc 12914 has an x - ray distribution that falls symmetrically about its nucleus , until it reaches the bridge where it remains strong . in section b ,
clumps of soft emission are seen in either side of the disk of ugc 12914 , but the emission again runs into a distinct bridge structure .
thus the bridge x - ray emission is clearly not a simple symmetrical starburst - driven halo .
returning to the bridge , we also find the average star formation rate across region bridge-1 m emission alone and using the method of calzetti et al .
this provided a sfr of 0.24 m@xmath6 yr@xmath3 , which is very consistent with the full sed approach . ] to be 0.26@xmath70 m@xmath6 yr@xmath3 , and a slightly lower value of 0.19@xmath71 m@xmath6 yr@xmath3 for bridge-2 .
these values are more than 10 times lower than the star formation rates in the disk of ugc 12915 .
given that in normal galaxies , the x - ray luminosity is expected to scale roughly as sfr@xmath72 @xcite , then if the bridge x - rays were created in a normal galactic environment , they should be at least ten times fainter .
however , as table [ soft ] shows , the soft x - ray surface brightness in the bridge-1 region is only a factor of two to three times less than the disk surface brightness .
alternatively , we can predict the expected star formation rate in the bridge based on the correlation derived for normal galaxies by @xcite .
based on this prediction , if star formation dominated the bridge x - ray emission , it should have a sfr of @xmath8 2.8 m@xmath6 yr@xmath3 , whereas we measure it to be ten times smaller .
therefore normal _ in situ _ star formation in the bridge is unlikely to be the origin of the x - rays .
observations with _ spitzer _ @xcite revealed the existence of an unusually large amount of warm molecular hydrogen in the bridge .
this was attributed to mechanical heating of the molecular gas by turbulent energy dissipation left over from the direct collision of the two galaxies . however , before discussing shocks further , it is worth ruling out a more mundane explanation for the abundance of warm h@xmath7 : that the molecular gas is heated because it is bathed in a strong x - ray field .
@xcite could not rule this out , since until the present work , no previous x - ray observations had been made .
we estimate the x - ray luminosity in the bridge evaluated over the same areas as those observed by the long - low slit of the _ spitzer _ irs .
these measurements are then compared with the h@xmath7 line luminosity l(h@xmath7)@xmath73 , where the ( 0,1 ) designation refers to the 0 - 0 s(0)28@xmath29 m and 0 - 0 s(1)17@xmath29 m pure rotational lines of warm molecular hydrogen .
this luminosity is almost certainly a lower limit to the total power emitted in the rotational hydrogen lines , since higher order transitions where not observed ( they fell outside the wavelength range of the long - low grating ) . summing the line luminosities over bridge regions b , c , d , e , j , k , l , m ( as defined by * ? ? ?
* ) , we find l(h@xmath7)@xmath73= 1.2 @xmath1 10@xmath74 erg s@xmath3 .
the x - ray luminosity over the same area is found to be l@xmath32 = 2.7 @xmath110@xmath2 erg s@xmath3 .
this implies that l(h@xmath7)@xmath75/l@xmath50 @xmath9 42 .
given the inherently inefficient ( @xmath18 1@xmath4 ) heating of h@xmath7 by x - rays , this high ratio rules out x - ray heating of cold molecular gas as a mechanism to explain the high luminosity of warm h@xmath7 .
the interpretation of the _ spitzer _ observations in terms of turbulence and shocks by @xcite is therefore upheld .
if the x - ray emitting gas is heated by shocks , what does the temperature of the bridge gas imply about the shock velocity ? based on the spectral fitting of the bridge , our best model for the bridge-1 region yields a temperature kt = 0.63 kev ( t=7.3 @xmath1
10@xmath76 k ) with absorption , or 0.77 kev ( 8.9 @xmath1 10@xmath76 k ) in the absence of significant absorption . for an idealized radiationless j - shock driven perpendicularly into a low - density perfect gas with high mach number , the temperature of the post - shock gas , t@xmath77= ( 3@xmath29m@xmath78/16k ) @xmath79@xmath22 , where @xmath29 is the mean molecular weight , m@xmath78 is the mass of the hydrogen atom , k is the bolztmann constant , and @xmath79 is the shock velocity .
for the range of temperatures we obtained from the _ chandra _ observations , shocks with velocities of 430 @xmath18 @xmath79 @xmath18 570 km s@xmath3 would be required if the pre - shocked medium was neutral .
such velocities are not out of the question , as it has been estimated that the taffy galaxies collided with an impact velocity of @xmath8 600 km s@xmath3 @xcite . when combining this with the high rotational velocity retrograde collision @xcite , the impact velocities at the shock interface could have been as high as 800 km s@xmath3 .
the lower - mass ugc 12915 could collide at velocities
@xmath9500 km s@xmath3 in the center of mass frame . despite the uncertainties in the actual relative velocities of the two galaxies (
most of the motion is now in the plane of the sky ) , the heating of the lower - density ism in these galaxies to high temperatures is quite plausible .
hot gas heated in the original impact exceeding 5 @xmath1 10@xmath80 k ( 0.43 kev ) , will cool extremely slowly ( few @xmath1 10@xmath81 to 10@xmath5 years ; * ? ? ?
* ) ) are of order 1gyr , but unlike guillard et al .
( 2009 ) , do not include the important cooling effects of the spluttering of small grains which considerably reduces the x - ray gas cooling time .
effects of heat conduction in supersonic turbulence ( not included ) would further reduce the cooling time . ] : longer than the time since collision of the two disks , based on radio synchrotron aging in the bridge , is 2 @xmath1 10@xmath81 yr @xcite .
it is therefore likely that the x - ray emitting bridge gas is a cooling remnant of the pre - shocked low - density material , splashed out into the bridge that has not yet cooled down .
evidence that gas has been splashed out may be seen not just in the bridge , but also in the distribution of hot gas in the galaxies .
although in ugc 12914 , the soft x - rays are well correlated with the nucleus and the northern half of the galaxy , they show a significant absence of x - rays in the southern half of the disk . for a normal galaxy
, this would be unexpected , since that region of the galaxy is populated with both warm and cool dust , and diffuse gas ( figure 6 ) .
a natural explanation is the x - ray emission from ugc 12914 is entirely dominated by shocked gas in the bridge ( seen in projection against the northern disk ) , and the intrinsic soft x - ray emission from ugc 12914 is too faint to be detected .
although ugc 12915 shows a closer correlation between the x - ray emission and the brighter parts of the molecular disk , there are strong deviations along the north - western part of the disk where the co emission is weak but x - rays are strong .
this may again indicate that the disk gas is strong disturbed in the direction of the bridge emission . the reality of the physical conditions in the taffy bridge are likely to be much more complex than that of an idealized perfect gas experiencing shocks .
numerical hydrodynamic simulations of a head - on collision of this kind by struck ( 1993 ) , predicted that shortly after the collision , a `` splash bridge '' should form containing a mix of hot gas from the original collision , with cooler gas reforming rapidly .
@xcite showed that with the right distribution of molecular cloud sizes and collisional impact parameters , significant amounts of both dense and diffuse gas can be stripped from the galaxies into the bridge .
their model was able to reproduce many of the observed morphological and kinematic features of the bridge .
it also provided an explanation for the offset spatial distributions of dense co - emitting gas , and the diffuse hi . in the model , this resulted from ugc 12914 striking the disk of ugc 12915 significantly off
center , leading to an enhancement of diffuse hi in the western part of the bridge compared with the denser eastern part .
this is consistent with our suggestion that the x - ray emitting gas is tracing the path of the lower - density material in the pre - collisional disk of ugc 12914 , drawn - out and offset from the denser gas as a result of the collision geometry .
although the @xcite models seem to reproduce the basic distribution and kinematics of the gas very well , this `` sticky - particle '' modeling approach , like most current hydrodynamic models , is unable to properly deal with the obvious multi - phase nature of the gas in the bridge .
the fact that gas can be both ejected from the galaxies , and can also rapidly cool in the aftermath of the collision , means that a proper description of the relative masses in each phase must take into account transformations ( and mixing ) between the phases , and detailed gas cooling in shocks and turbulence . by analogy with the taffy bridge
, @xcite presented a multi - phase model to explain x - ray ( trinchieri et al . 2003 ; 2005 ,
osullivan 2009 ) and warm h@xmath7 emission ( appleton et al .
2006 , cluver et al .
2010 ) in the intergalactic shock in stephan s quintet .
this model was designed to treat the case of a high - speed ( 1000 km s@xmath3 ) collision of an intruder galaxy , ngc 7318b with an large 40 kpc - long hi tidal filament .
the model showed that the initial high - speed shock moving through multi - phase hi filament would shock heat the low - density pockets to x - ray temperatures , rapidly spluttering the dust in those regions , while the denser pre - shock gas could survive because it is shocked to lower velocities . like
stephan s quintet , the pressure ( p / k ) of the hot diffuse x - ray emitting gas in the taffy bridge ( see table [ model ] ) @xmath810@xmath82 k @xmath83 , is fifty times less than the pressure in the warm , dense h@xmath7 phase of bridge ( @xmath86@xmath110@xmath84 k @xmath83 for t= 165 k , and n(h@xmath7)=10@xmath85 ; * ? ? ?
* ) . as discussed by guillard et al .
( 2009 ) , supersonic turbulence leads to significant pressure differences in the post - shock gas , with no expectation of pressure equilibrium .
indeed , the cooling time for the warm h@xmath7 is so short that warm h@xmath7 must be continuously created , as individual dense regions dissolve as they expand into the ambient hot gas .
this implies rapid cycling of material through different ism phases .
the ratio of the warm h@xmath7 to x - ray luminosity l(h@xmath7)/l@xmath50 in stephan s quintet ( @xmath8510 ) is much smaller than that seen in the taffy bridge , l(h@xmath7)/l@xmath50 @xmath9 42 .
the l(h@xmath7)/l@xmath50 ratio for taffy is also particularly large compared with other shocked systems ( see figure 9 of * ? ? ?
part of the explanation may relate to the difference in thermal energy imparted by the faster collision in the quintet ( the difference in imparted kinetic energy could be a factor of 5 ) compared with taffy .
the difference could also relate the distribution of pre - shock densities . in the quintet s shock
, it is believed that the intruder collided with a diffuse hi filament , not a complete galaxy disk , thus increasing the amount of low - density material available for x - ray heating . in the taffy
, there may have been a much larger fraction of dense gas present in both colliders , which would enhance the warm h@xmath7 fraction relative to the hot x - rays .
previous observations of the taffy system have shown that the bridge contains @xmath87@xmath1 10@xmath59 m@xmath6 of molecular and atomic gas @xcite .
the existence of the x - ray emission adds to the multi - phase nature of the bridge , and supports the view that the bridge material was gas stripped away from the galaxies in the head - on collision .
the stripping of gas from the galaxies , and the turbulent nature of the bridge gas , may partly explain why the galaxies have not yet evolved into lirg ( the combined ir luminosity of the taffy galaxies , l@xmath13 = 4.5 @xmath1 10@xmath14 l@xmath6 ; see table [ ir ] ) . for the bridge , the depletion timescale , @xmath86 , defined as the total gas mass divided by the star formation rate , is @xmath830 gyrs , assuming an average sfr of 0.225 m@xmath6 yr@xmath3 ( see table [ ir ] ) . if only molecular gas ( cold and warm ) is considered in the estimate ( m(h@xmath7 ) = 2.3 @xmath1 10@xmath19 m@xmath6 for d = 62 mpc , * ? ? ?
* ; * ? ? ?
* ) , then @xmath87 @xmath810 gyrs .
these values ( the ratio of n@xmath88 to @xmath89 ) for the bridge , used by zhu et al .
( 2007 ) of 2.6 @xmath1 10@xmath90 @xmath34 k km s@xmath3 , is very conservative , being a factor of ten lower than that adopted for the galaxy , based on their lvg excitation analysis .
thus the depletion times could be underestimated if the total gas mass is larger than assumed here . ] , which are comparable with , or greater than the age of the universe , suggest very inefficient star formation in the bridge ( the star formation efficiency is the inverse of @xmath87 ) .
a similar conclusion , that the bridge exhibits unusually low star formation for its gas mass , was found by @xcite using _ spitzer _ data alone .
long depletion times have been found in a sample of warm h@xmath7rich hickson compact group galaxies ( alatalo et al . submitted ) , and in ngc 1266 @xcite . in both situations , shock - heating of the gas was implicated in the star formation suppression . in taffy ,
the depletion times are likely to be lower limits , because the star formation rate in the bridge is almost certainly overestimated , since the sed fitting assumes the uv radiation from young stars is the main heating mechanism for the dust .
we can also estimate the gas depletion times for the individual galaxies . adopting the total h@xmath7 + hi masses for ugc 12914 and ugc 12915 of 9.2 and 8.2 @xmath1 10@xmath59 m@xmath6 @xcite , and the star formation rates from table [ ir ] , we find t@xmath91 = 8.8 and 3.2 gyrs respectively .
if we only include the h@xmath7 masses , the depletion times drop to 4 and 1.6 gyrs respectively ( or log@xmath92(@xmath87 ) = 9.6 and 9.2 ) .
@xcite have shown that , for normal galaxies , there is a good correlation between h@xmath7derived log(@xmath87 ) and stellar mass log(m@xmath93 ) .
if we adopt the values of stellar mass from the sed fitting , the correlation of @xcite would predict depletion times of log(t@xmath91 ) = 9.1 and 9.0 respectively for ugc 12914 and ugc 12915 with a 1-@xmath36 scatter of dex 0.27 .
the observed value of log(@xmath87 ) for ugc 12915 is therefore well within the main distribution for normal galaxies , and ugc12914 deviates by only 1.9@xmath36 from the mean relation ( on the star formation deficient side ) .
we conclude that , unlike the bridge , the taffy galaxies have depletion times ( and star formation efficiencies ) that are within the normal range for their gas content . nevertheless , their @xmath87 values are an order of magnitude longer than those of lirgs @xcite .
the most likely reason that the taffy galaxies have not yet become strong ir emitters , despite having recently undergone a major gas - rich collision , is the geometry of the head - on collision .
this has strongly disrupted the gas , splashing it @xcite into the bridge where much of it is still turbulent .
the bridge gas still contains large quantities of warm h@xmath7 , ( see * ? ? ?
* ) , as well as a significant component of hot x - ray emitting gas .
in addition , there has not been enough time since the collision ( 25 myrs ) for the bridge gas to fall back onto the disks , which might eventually lead to more activity .
@xcite pointed out the existence of a ring structure in ugc 12914 , probably arising from the same process that leads to collisional ring galaxies ( see * ? ? ?
the geometry of the taffy system is such that both galaxies might be expected to form radially expanding rings within their disks in this nearly face - on collision .
it is possible that the ulxs in ugc 12915 are part of a forming gas - rich ring , seen edge - on .
for example , in @xcite , more than a dozen ulxs were discovered associated with the main expanding star forming ring of the cartwheel collisional galaxy .
as @xcite showed , collisional ring galaxies are rarely bright enough to be lirgs , partly because the star formation is confined to a narrow ring ( or multiple rings ) , and generally does not form a strong nuclear starburst .
we conclude that the peculiar near - central collisional geometry , combined with strong gas stripping , has predisposed the taffy galaxies to be less active than a more typical major merger .
the x - ray observations of the taffy galaxies and bridge , combined with new _ herschel _
data , have shown the following main results : * x - ray emission is detected from both galaxies and from the bridge regions of the taffy system .
the galaxies themselves contain both compact point - like sources and extended soft x - ray emission .
* nine ulx ( l@xmath50 @xmath8 10@xmath2 erg s@xmath3 ) , and potentially more fainter point - like sources , are found in the taffy system , and the brightest ulx ( l@xmath0 = 2.8 @xmath110@xmath2 erg s@xmath3 ) lies in the bridge , close to the position of a giant extragalactic hii region which is associated with a peak in the molecular gas density in the bridge .
an additional , slightly extended source with a harder x - ray spectrum , is found associated with the nucleus of ugc 12914 , perhaps hinting at the existence of a buried agn .
the number of ulxs in the two galaxies scales roughly with the sfr , as found in previous studies , suggesting that these x - ray sources are generated in massive x - ray binary systems , with an evolution that is intimately tied to massive star formation .
this is consistent with the shape of the xlf for the compact sources . *
extended x - ray emission , l@xmath0 = 5.4 @xmath1 10@xmath2 erg s@xmath3 , is found primarily in the north - western part of the bridge .
the total mass in hot gas is ( 0.81.3 ) @xmath1 10@xmath5 m@xmath6 , which is approximately 1@xmath4 of the total ( hi + h@xmath7 ) gas mass in the bridge , and @xmath811@xmath4 of the mass of warm molecular hydrogen discovered by _
spitzer_. it is more closely associated with far - ir emitting diffuse gas and dust , than with either the brightest ridge of radio continuum emission seen in the bridge or with the distribution of denser gas tracers in the bridge . the lack of obvious correlation with the brighter radio continuum emission in the bridge rules out a direct connection between the x - rays and the synchrotron emission in the bridge caused by cosmic rays . *
we show , using full sed fitting including _ herschel _ far - ir data , that the star formation rate ( 0.24 m@xmath6 yr@xmath3 ) , is too low in the bridge to explain the relatively high x - ray surface brightness detected there .
furthermore , only ugc 12915 has a global star formation rate comparable with previously studied minor - axis x - ray outflow galaxies , and the x - ray surface brightness profiles are quite dissimilar to galaxies with known x - ray outflows .
we favor shock heating from the galaxy collision as the main source of x - rays .
the offset distribution of the diffuse gas tracers ( including the x - ray emission ) from the dense gas tracers in the bridge , may be the results of an off - center collision between the two galaxies .
* we suggest that the shock - heating of low - density material in the precursor ism of both galaxies has been splashed out ( see * ? ? ?
* ) into the bridge along with other diffuse gas leaving a highly turbulent bridge unable to support much star formation .
the properties of the x - ray bridge implies that the gas is only just beginning to cool significantly in the 25 myrs since the collision when it was likely first shocked .
the temperature of the extended soft x - ray emitting gas is consistent with heating by shocks of over a range of 430 to 570 km@xmath3 depending on the assumed intrinsic absorption in the x - ray spectral modeling .
these values are similar to the impact velocities assumed for the head - on collision that has led to the bridge formation .
the pressure in the hot gas is found to be much lower than the implied pressure in the warm h@xmath7 , as expected from models of supersonic turbulence which predict rapid cycling of molecular material through many different phases in the post - shocked region . *
the luminosity of the x - ray emission is 42 times lower than the mid - ir line luminosity of warm h@xmath7 emission seen in the bridge by _
spitzer _ in previous observations .
the new _ chandra _ observations thus rule out the x - rays themselves as being the heat - source for the 10@xmath59 m@xmath6 of warm molecular gas in energetic grounds . exploring
the importance of x - ray heating of the molecular gas in the bridge was an important motivation for this study . *
the combination of _ herschel _ photometry , with full sed fitting , implies normal gas depletion times for the two taffy galaxies ( implying normal star formation efficiencies ) but an extremely long depletion time ( @xmath910 gyrs ) for the gas bridge .
star formation therefore seems suppressed in the bridge where warm h@xmath7 and x - ray emitting gas have not yet cooled , and velocity dispersions remain high in the molecular gas .
we suggest that the reason the taffy system is not a strong ir - emittor ( as with many other major gas - rich mergers ) is related to the collisional geometry , which has stripped and disrupted a large amount of the gas into the bridge .
we thank an anonymous referee for very helpful comments that improved the manuscript .
the scientific results reported in this article are based on observations made by the chandra x - ray observatory .
support for this work was provided by the national aeronautics and space administration through chandra award number go3 - 14087x , issued by the chandra x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of the national aeronautics space administration under contract nas8 - 03060 . the work is also based , in part , on observations made with _
, a european space agency cornerstone mission with significant participation by nasa .
partial support for ll , pa & ka was also provided for _ herschel _ observations , through a contract issued by the jet propulsion laboratory , california institute of technology under a contract with nasa .
this research has made use of the nasa/ ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
partial support for ka was also provided by nasa through hubble fellowship grant awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas5 - 26555 .
j.w . acknowledges support from nsfc grants 11443003 and 11473021 .
alatalo , k. , lacy , m. , lanz , l. , et al .
2015 , , 798 , 31 appleton , p. n. , & struck - marcell , c. 1987 , , 312 , 566 appleton , p. n. , & struck - marcell , c. 1996 , fund . of cosmic physics , 16 , 111 appleton , p. n. , xu ,
k. c. , reach , w. , et al .
2006 , , 639 , l51 appleton , p. n. , guillard , p. , boulanger , f. , et al . 2013 , , 777 , 66 armus , l. , heckman , t. m. , & miley , g. 1987 , , 94 , 831 armus , l. , heckman , t. m. , weaver , k. a. , & lehnert . , m. d. , 445 , 666 barnes , j. e. , & hernquist , l. 1996 , , 471 , 115 braine , j. , davoust , e. , zhu , m. , lisenfeld , u. , motch , c. , & seaquist , e. r. 2003 , , 408 , l13 bushouse , h. a. 1987 , , 320 , 49 calzetti , d. , kennicutt , r. c. , engelbracht , c. w. , et al .
2007 , , 666 , 870 cox , d. p. 1979
, , 234 , 863 cluver , m. e. , appleton , p. n. , boulanger , f. , et al .
2010 , , 710 , 248 condon , j. j. , helou , g. , sanders , d. b. , & soifer , b. t. 1993 , , 105 , 1730 cox , d. p. 1981 , , 245 , 534 da cunha , e. , charlot , s. , & elbaz , d. 2008 , , 388 , 1595 da cunha , e. , charmandaris , v. , daz - santos , t. , et al .
2010 , , 523 , aa78 dahlem , m. , weaver , k. a. , & heckman , t. m. 1998 , , 118 , 401 drzazga , r. t. , chyy , k. t. , jurusik , w. , & wirkiewicz , k. 2011 , , 533 , a22 elbaz , d. , cesarsky , c. j. , chanial , p. , et al .
2002 , , 384 , 848 fabbiano , g. , & trinchieri , g. 1984 , , 286 , 491 fabbiano , g. , & trinchieri , g. 1985 , , 296 , 430 fabbiano , g. 1988 , , 330 , 672 fabbiano , g. 1989 , , 27 , 87 fabbiano , g. , heckman , t. , & keel , w. c. 1990 , , 355 , 442 fabbiano , , g. zezas , a. , king , a. r. , et al .
2003 , , 584 , l5 freeman , p. e. , kashyap , v. , rosner , r. , & lamb , d. q. 2002 , , 138 , 185 garrington , s. t. , & conway , r. g. 1991 , , 250 , 198 gao , y. , zhu , m. , & seaquist , e. r. 2003a , , 126 , 2171 gao , y. , wang , q .- d . ,
appleton , p. n. , & lucas , r .- a . 2003b , , 596 , l171 guillard , p. , boulonger , f. , pineau des forts , g. , & appleton , p. n. , 2009 , , 502 , 515 guillard , p. , boulanger , f. , pineau des forts , g. , et al .
2012 , , 749 , 158 hardcastle , m. j. , & krause , m. g. h. 2014 , , 443 , 1482 harrison , f. a. , eckart , m. e. , mao , p. h. , helfand , d. j. , & stern , d. 2003 , , 596 , 944 hayward , c. c. , lanz , l. , ashby , m. l. n. , et al .
2014 , , 445 , 1598 heckman , t. m. , armus , l. , & miley , g. k. 1990 , , 74 , 833 heckman , t. m. 2002 , extragalactic gas at low redshift , 254 , 292 huang , m .-
, & kauffmann , g. 2014 , mnras , 443 , 1329 ho , l. c. 2009 , , 699 , 626 hopkins , p. f. , cox , t. j. , younger , j. d. , & hernquist , l. 2009 , , 691 , 1168 hopkins , p. f. , kere , d. , oorbe , j. , et al . 2014 , , 445 , 581 houck , j. r. , roellig , t. l. , van cleve , j. , et al .
2004 , , 154 , 18 jarrett , t. h. , helou , g. , van buren , d. , valjavec , e. , & condon , j. j. 1999 , , 118 , 2132 joseph , r. d. , & wright , g. s. 1985 , , 214 , 87 joye , w. a. , & mandel , e. 2003 , astronomical data analysis software and systems xii , 295 , 489 kalberla , p. m. w. , burton , w. b. , hartmann , d. , et al .
2005 , a&a , 440 , 775 kilgard , r. e. , kaaret , p. , krauss , m. t. et al .
2002 , , 573 , 138 komossa , s. , burwitz , v. , hasinger , g. , et al .
2003 , , 582 , l15 komugi , s. , tateuchi , k. , motohara , k. , et al .
2012 , , 757 , 138 lanz , l. , ogle , p. m. , evans , d. , et al .
2015 , , 801 , 17 lisenfeld , u. , & vlk , h. j. 2010 , , 524 , a27 luangtip , w. , roberts , t. p. , mineo , s. , et al .
2015 , , 446 , 470 mckee , c. f. , & ostriker , j. p. 1977
, , 218 , 148 nesvadba , n. p. h. , boulanger , f. , salom , p. , et al .
2010 , , 521 , a65 mineo , s. , gilfanov , m. , & sunyaev , r. 2012 , , 419 , 2095 mineo , s. , gilfanov , m. , & sunyaev , r. 2012 , , 426 , 1870 mirabel , i. f. , vigroux , l. , charmandaris , v. , et al .
1998 , , 333 , l1 ogle , p. , antonucci , r. , appleton , p. n. , & whysong , d. 2007 , , 657 , 161 ogle , p. , boulanger , f. , guillard , p. , et al .
2010 , , 724 , 1193 oka , t. , geballe , t. r. , gioto , m. , usuda , t. , & mccall , b. j. 2005 , , 632 , 882 osullivan , e. , giacintucci , s. , vrtilek , j. m. , raychaudhury , s. , & david , l. p. 2009
, , 701 , 1560 ott , s. 2010 , astronomical data analysis software and systems xix , 434 , 139 peterson , b. w. , appleton , p. n. , helou , g. , et al .
2012 , , 751 , 11 ranalli , p. , comastri , a. , & setti , g. 2003 , , 399 , 39 read , a. m. , ponman , t. j. , & strickland , d. k. 1997 , , 286 , 626 rigopoulou , d. , kunze , d. , lutz , d. , genzel , r. , & moorwood , a. f. m. 2002 , , 389 , 374 roussel , h. , helou , g. , hollenbach , d. j. , et al .
2007 , , 669 , 959 roussel , h. 2013 , pasp , 125 , 1126 sanders , d. b. , soifer , b. t. , elias , j. h. , et al .
1988a , , 325 , 74 sanders , d. b. , soifer , b. t. , elias , j. h. , neugebauer , g. , & matthews , k. 1988b , , 328 , l35 sanders , d. b. , & mirabel , i. f. 1996 , , 34 , 749 sanders , d. b. , mazzarella , j. m. , kim , d .-
c . , surace , j. a. , & soifer , b. t. 2003 , , 126 , 1607 scoville , n. z. , sanders , d. b. , & clemens , d. p. 1986
, , 310 , l77 sivandandam , s. , rieke , m. j. , rieke , g. h. 2009 , , 717 , 147 smith , r. k. , brickhouse , n. s. , liedahl , d. a. , & raymond , j. c. 2001 , apjl , 556 , l91 smith , b. j. , swartz , d. a. , miller , o. , et al . 2012 , , 143 , 144 soifer , b. t. , sanders , d. b. , madore , b. f. , et al .
1987 , , 320 , 238 soifer , b. t. , & neugebauer , g. 1991 , , 101 , 354 soria , r. 2011 , astronomische nachrichten , 332 , 330 strickland , d. k. , heckman , t. m. , weaver , k. a. , hoopes , c. g. , & dahlem , m. 2002 , , 568 , 689 strickland , d. k. , heckman , t. m. , colbert , e. j. m. , hoopes , c. g. , & weaver , k. a. 2004 , , 606 , 829 strickland , d. k. , heckman , t. m. , colbert , e. j. m. , hoopes , c. g. , & weaver , k. a. 2004 , , 151 , 193 struck , c. 1997 , , 113 , 269 swartz , d. a. , ghosh , k. k. , tennant , a. f. , & wu , k. 2004 , , 154 , 519 swartz , d. a. , soria , r. , tennant , a. f. , & yukita , m. 2011 , , 741 , 49 trinchieri , g. , sulentic , j. , breitschwerdt , d. , & pietsch , w. 2003 , , 401 , 173 trinchieri , g. , sulentic , j. , pietsch , w. , & breitschwerdt , d. 2005 , , 444 , 697 vollmer , b. , braine , j. , & soida , m. 2012 , , 547 , a39 watson , m. g. , stanger , v. , & griffiths , r. e. 1984 , , 286 , 144 weisskopf , m. c. , tananbaum , h. d. , van speybroeck , l. p. , & odell , s. l. 2000 , proc .
spie , 4012 , 2 whitmore , b. c. , & schweizer , f. 1995 , , 109 , 960 whitmore , b. c. , brogan , c. , chandar , r. , et al .
2014 , , 795 , 156 wilson , c. d. , scoville , n. , madden , s. c. , & charmandaris , v. 2003 , , 599 , 1049 xu , c. k. , shupe , d. l. , bthermin , m. , et al .
2012 , , 760 , 72 zezas , a. , et al .
2006 , apjs , 116 , 211 zezas , a. , fabbiano , g. , baldi , a. , et al .
2007 , , 661 , 135 zhu , m. , gao , y. , seaquist , e. r. , & dunne , l. 2007 , , 134 , 118 ugc 12914 ( diffuse ) & ugc12914 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath98 & @xmath99 & 40.22 + ugc 12914 ( + point sources)@xmath100 & & & & & & & 40.51 + ugc 12915 ( diffuse only ) & ugc12915 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath101 & @xmath102 & 40.06 + ugc 12915 ( + point sources)@xmath100 & & & & & & & 40.34 + taffy bridge & bridge & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath103 & @xmath104 & 39.73 + extragalactic hii ( + b1 ) & hii & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath105 & @xmath106 & 39.47 + cxou j000139.0 + 232942 & bridge-1 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath107 & @xmath1081.7 & 39.61 + cxou j000141.4 + 232903 & bridge-2 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & polygon & @xmath1842 & @xmath185.1 & @xmath1839.37 + cxou j000138.3 + 232901 & w@xmath43 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath110 & @xmath111 & 40.06 + cxou j000138.0 + 232905 & w1 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath112 & @xmath113 & 39.12 + cxou j000139.1 + 232909 & w2 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath112 & @xmath114 & 39.12 + cxou j000137.9 + 232836 & w3 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath115 & @xmath1161.0 & 39.16 + cxou j000140.9 + 232938 & b1 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath117 & @xmath1181.5 & 39.47 + cxou j000142.9 + 232935 & e1 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath119 & @xmath120 & 39.39 + cxou j000142.3 + 232941 & e2 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath119 & @xmath121 & 39.41 + cxou j000141.2 + 232949 & e3 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath1095circle & @xmath122 & @xmath1231.2 & 39.32 + cxou j000141.6 + 232950 & e4 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 1@xmath1095circle & @xmath124 & @xmath125 & 39.12 + cxou j000141.8 + 232948 & e5 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 1@xmath1095circle & @xmath126 & @xmath127 & 39.32 ugc 12914 & 3704 & [email protected] & [email protected] & 9.0 + ugc 12915 & 2562 & [email protected] & [email protected] & 7.4 + bridge & 3771 & [email protected] & [email protected] & 3.1 + bridge-1 & 1637 & [email protected] & [email protected] & 5.7 + bridge-2 & 2127 & @xmath180.5 & @xmath180.5 & @xmath183 + hii & 356 & [email protected] & [email protected] & 17.1 lrrlccc b2 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath128 & @xmath129 & 38.95 + b3 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath128 & @xmath129 & 38.95 + b4 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath130 & @xmath131 & 38.88 + b5 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath132 & @xmath133 & 38.81 + e6 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath130 & @xmath131 & 38.88 + e7 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 1@xmath1095circle & @xmath134 & @xmath135 & 38.84 + e8 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 1@xmath1095circle & @xmath136 & @xmath137 & 38.75 + w4 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath138 & @xmath139 & 39.06 + w5 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath140 & @xmath141 & 38.95 + w6 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath138 & @xmath139 & 39.06 + w7 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath132 & @xmath133 & 38.81 + w8 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath140 & @xmath141 & 38.95 + w9 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath130 & @xmath131 & 38.88 + w10 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath142 & @xmath143 & 39.01 + w11 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath142 & @xmath143 & 39.00 + w12 & 00@[email protected]@xmath96 & + 23@[email protected]@xmath96 & 2@xmath28circle & @xmath144 & @xmath133 & 38.80 ugc 12914 & apec+pl & [email protected] & [email protected] & [email protected] & [email protected] & ... & 8.4/18 + & abs(apec ) & [email protected] & [email protected] & ... & ... & [email protected] & 11.9/19 + & apec & [email protected] & [email protected] & ... & ... & ... & 16.5/20 + ugc 12915 & apec+pl & [email protected] & [email protected] & [email protected] & [email protected] & ... & 7.6/10 + & abs(apec ) & [email protected] & [email protected] & ... & ... & [email protected] & 10.5/11 + & apec & [email protected] & [email protected] & ... & ... & ... & 14.5/12 + taffy galaxies & apec+pl & [email protected] & [email protected] & [email protected] & [email protected] & .... & 21.7/31 + ( summed ) & abs(apec ) & 397@xmath45383 & [email protected] & ... & ... & [email protected] & 23.6/32 + & apec & [email protected] & [email protected] & ... & ... & ... & 33.2/33 + bridge & apec & [email protected] & [email protected] & ... & ... & ... & 4.3/10 + & abs(apec ) & [email protected] & [email protected] & ... & ... & [email protected] & 3.5/9 lllccccccc bridge & apec & 0.77 & 3.8 & 5.2 & 1.8 & 0.8 & 1.8 & 2.5 & 1.0 + & abs(apec ) & 0.63 & 10.2 & 5.2 & 3.0 & 1.3 & 2.3 & 3.3 & 1.4 + ugc 12914 & apec+pl & 0.71 & 6.1 & 8.4 & 1.8 & 1.3 & 2.6 & 2.3 & 1.0@xmath145 + & abs(apec ) & 0.58 & 28.7 & 8.4 & 4.0 & 2.8 & 4.6 & 4.0 & 0.9 + & apec & 0.75 & 8.7 & 8.4 & 2.2 & 1.5 & 3.3 & 2.9 & 0.6 + ugc 12915 & apec+pl & 0.72 & 2.9 & 4.6 & 1.7 & 0.7 & 1.4 & 2.1 & 1.1@xmath145 + & abs(apec ) & 0.61 & 19.0 & 4.6 & 4.2 & 1.7 & 2.9 & 4.6 & 0.8 + & apec & 0.78 & 4.7 & 4.6 & 2.2 & 0.8 & 1.9 & 2.9 & 0.5 lrrrrr ugc 12914 & 10.43@xmath146 & 10.01@xmath146 & [email protected] & 10.87@xmath147 & 1.05@xmath64 + ugc 12915 & 10.81@xmath148 & 10.54@xmath148 & [email protected] & 10.62@xmath149 & 2.55@xmath65 + bridge-1 & 9.79@xmath70 & 9.28@xmath70 & [email protected] & 9.87@xmath150 & 0.26@xmath70 + bridge-2 & 9.66@xmath151 & 9.14@xmath151 & [email protected] & 9.77@xmath152 & 0.19@xmath71 + hii & 9.64@xmath71 & 9.22@xmath71 & 7.30@xmath150 & 9.04@xmath153 & 0.24@xmath154 | we present the first x - ray observations of the taffy galaxies ( ugc 12914/5 ) with the _
chandra _ observatory , and detect soft x - ray emission in the region of the gas - rich , radio - continuum - emitting taffy bridge .
the results are compared to _
herschel _ observations of dust and diffuse [ cii ] line - emitting gas .
the diffuse component of the taffy bridge has an x - ray luminosity of l@xmath0 = 5.4 @xmath1 10@xmath2 erg s@xmath3 , which accounts for 19@xmath4 of the luminosity of the sum for the two galaxies .
the total mass in hot gas is ( 0.81.3 ) @xmath1 10@xmath5 m@xmath6 , which is approximately 1@xmath4 of the total ( hi + h@xmath7 ) gas mass in the bridge , and @xmath811@xmath4 of the mass of warm molecular hydrogen discovered by _ spitzer_. the soft x - ray and dense co - emitting gas in the bridge
have offset distributions , with the x - rays peaking along the north - western side of the bridge in the region where stronger far - ir dust and diffuse [ cii ] gas is observed by _
herschel_. we detect nine ultra luminous x - ray sources ( ulxs ) in the system , the brightest of which is found in the bridge , associated with an extragalactic hii region .
we suggest that the x - ray emitting gas has been shocked heated to high temperatures and `` splashed '' into the bridge by the collision .
the large amount of gas stripped from the galaxies into the bridge and its very long gas depletion timescale ( @xmath910 gyr ) may explain why this system , unlike most major mergers , is not a powerful ir emitter .
accepted for publication in apj |
the standard model ( sm ) of interactions among elementary particles is one of the best verified physics theories up to now but there are many open fundamental questions remain unanswered within the scope of the sm .
high energy physics experiments are designed to address the open questions through the search of new physics ( np ) using two complementary approaches .
one is to discover the new particles at the high energy large hadron collider ( lhc ) . the other is to search for the effects of np through measurements of flavor physics reactions at lower energy scales and evidence of a deviation from the sm prediction .
after the observation of the rare radiative decay @xmath14
@xcite , the flavor - changing neutral current ( fcnc ) transitions became more attractive and since then many works about rare radiative , leptonic and semileptonic decays have been intensively done in the @xmath15 system @xcite . among these decays ,
semileptonic decay channels are significant because their branching ratios are relatively larger .
these works will be more perfect if similar studies for @xmath16 , observed in 1998 by cdf collaboration @xcite , are also included .
the charmed @xmath16 meson is a ground state of two heavy quarks @xmath17 and @xmath18 . because of the two heavy quarks , the decays of the @xmath16 meson are rather different from @xmath19 mesons .
physicists therefore believe that the @xmath16 physics must be very rich compared to the other @xmath20 mesons if the statistics reaches high level @xcite . at lhc , around @xmath21 @xmath16 events per year
are expected @xcite .
the expected number events are motivating to work on the @xmath16 phenomenology and this possibility will provide facilities to study the observables of rare @xmath16 decays such as branching ratios , forward - backward asymmetry and polarization asymmetries .
the rare @xmath22 decays are proceeded by fcnc transition of @xmath23 , which are forbidden at the tree level in the sm , and play an important role in the precision test of the sm .
meanwhile , they offer a valuable possibility of an indirect search of np for their sensitivity to the gauge structure and new contributions . up to now , the possible new physics contributions to @xmath22 decays have been studied extensively , for example , by using model independent effective hamiltonian @xcite , in supersymmetric models @xcite , with fourth generation effects @xcite , and in single universal extra dimension @xcite .
when concentrating on the exclusive @xmath22 decays , one needs to know the form factors . as for @xmath24 transition
, the form factors have been calculated using different approaches , such as light front constituent quark models @xcite , a relativistic constituent quark model @xcite , a relativistic quark model @xcite , the ward identities @xcite , in light cone qcd @xcite , and qcd sum rules @xcite . in this work
, we will adopt the form factors calculated in the three - point qcd sum rules @xcite to study the @xmath25 effects on the observables for @xmath26 decay .
the general framework for non - universal @xmath1 model has been developed in ref .
@xcite . in this model ,
@xmath25 gauge boson could be naturally derived by adding additional @xmath27 gauge symmetry .
non - universal @xmath25 couplings can induce fcnc @xmath28 and @xmath29 transitions at tree level .
its effects on @xmath30 transition have received great attention and been widely studied in the literature . the previous works in a family
non - universal @xmath25 model boson redound to resolve many puzzles , such as `` @xmath31 puzzle '' @xcite , anomalous @xmath32 mixing phase @xcite and mismatch in @xmath33 spectrum at low @xmath34 region @xcite .
motivated by this , we will study the effects of the @xmath25 boson on the rare decay @xmath26 .
this paper is organized as follows . in section 2
, we present the effective hamiltonian responsible for the @xmath35 transition in both the sm and the family non - universal @xmath1 model . in this section
we also present the matrix element , and the expressions of various physical observables in @xmath1 model . in section 3 ,
we show the numerical results of the observables for @xmath36 decay in the sm and @xmath1 model .
the final section is the summary .
at quark level , the rare semileptonic decay @xmath35 can be described in terms of the effective hamiltonian which is given by @xcite @xmath37 here the explicit expressions of @xmath38 can be found in ref .
@xcite , in which @xmath39 the wilson coefficients @xmath40 can be expanded perturbatively @xcite .
the effective coefficients @xmath41 , can be written as @xcite @xmath42 where the perturbative part @xmath43 stands for the matrix element of four - quark operators and is given by @xmath44 with @xmath45 , @xmath46 .
we have neglected the resonance contribution . for the detailed discussion of such resonance effects
, we refer to refs .
@xcite .
exclusive decay @xmath36 is described in terms of matrix elements of the quark operators in the effective hamiltonian over meson states , which can be parameterized in terms of form factors .
the matrix elements of @xmath47 transition are given by @xcite d_s^*(p ) | ( v - a)_| b_c(p_b_c ) & = & -i ^*_(m_b_c+m_d_s^ * ) a_0(s ) + i ( p_b_c+p)_(^*p_b_c ) + [ eq : ff3 ] and d_s^*(p ) | |s _
q^(1+_5 ) b
| b_c(p_b_c ) & = & i _ ^ * p_b_c^p^ 2 t_1(s ) + & & + t_2(s ) \ { ^ * _ ( m_b_c^2-m_d_s^*^2 ) - ( ^ * p_b_c ) ( p_b_c+p ) _ } + & & + t_3(s ) ( ^ * p_b_c )
\ { q_- ( p_b_c+p)_}[eq : t ] here @xmath48 , @xmath49 , and @xmath50 is polarization vector of the vector meson @xmath51 .
the dilepton invariant mass spectrum for @xmath22 decays can be expressed by @xcite = | v_ts^v_tb |^2 ( ) d where the explicit expression of @xmath52 is d & = & ( 1 + 2 ) + |e|^2 . + & & + . . + & & + . . + & & - . .
2 + |h|^2 , [ eq : dwbvll ] with @xmath53 , and @xmath54 .
the kinematic variables @xmath55 and @xmath56 are the same as ref .
the auxiliary functions @xmath57 and @xmath58 which are combinations of the effective wilson coefficients in eq .
( [ eq : effwc ] ) and the form factors of @xmath59 transition can be found in refs .
@xcite . for the convenience of the reader , we present these functions in the appendix a.
the normalized forward - backward asymmetry ( fba ) is defined as _
fb()=d .according to this definition ,
the explicit expression of fba is : [ eqafb ] d & = & ( ) [ re(be^ * ) + re(af^ * ) ] .
the lepton polarization can be defined as : = ( ) _ 0[1 + ( p_l_l+p_n_n+p_t_t ) ] where the subscript @xmath60 stands for the unpolarized decay case . @xmath6 and @xmath11 are the longitudinal and transverse polarization asymmetries in the decay plane respectively , and @xmath61 is the normal polarization asymmetry in the direction perpendicular to the decay plane
. the lepton polarization asymmetry @xmath62 can be derived by p_i()= the results are p_l d&=&\ { re(ae^*)+re(bf^ * ) .
+ & & .-re(bg^*+cf^ * ) + re(cg^*)},[plks ] + p_n d&=&\ { .
+ & & .+2_d_s^*_l[im(be^*)+im(af^*)]},[pnks ] + p_t d&= & \ { 4re(ab^*)+ .
+ & & .+[re(cf^*)-(1-^2_d_s^*)re(cg^ * ) -re(ch^*)]}. [ ptks ] in the family non - universal @xmath25 model , the flavor neutral currents arise even at tree level owing to non - diagonal chiral coupling matrix . postulating that the couplings of right - handed quark flavors with @xmath1 boson are diagonal , the @xmath1 part of the effective hamiltonian for @xmath35 transition is described by @xcite @xmath63+{\rm h.c.}\,.\ ] ] to extract the @xmath1 corrections to the wilson coefficients
, one can reformulate eq .
( [ zphbsll ] ) as @xmath64+{\rm h.c.}\,,\ ] ] with @xmath65 where @xmath66 , @xmath67 with @xmath68 and @xmath69 referring to the effective chiral @xmath1 couplings to quarks and leptons , respectively .
the off - diagonal element @xmath68 contains a new weak phase and can be written as @xmath70 .
when we include the @xmath1 contributions with the assumption of no significant rg running effects between @xmath71 and @xmath72 scales , the wilson coefficients can be written as c_9,10^sm(m_w)c_9,10^sm(m_w)+c_9,10^(m_w ) .
after inclusion of the new contributions from @xmath1 boson , the rg evolution of the wilson coefficients down to low scale is exactly the same as in the sm .
in this section , we focus on the numerical calculations of the branching ratios , forward - backward asymmetry and polarization asymmetries for @xmath26 decay .
the input parameters which are related to our analysis are summarized in table [ ipp ] .
0.7pt 0.1 in .default values of inputs parameters used in our numerical calculations . [ cols="<,<,<,<",options="header " , ] in the family non - universal @xmath1 model ,
the @xmath1 contributions rely on four parameters @xmath73 , @xmath74 , @xmath75 and @xmath76 .
these parameters have been constrained from the well measured decays by many groups @xcite . @xmath73 and @xmath74
have been strictly constrained by @xmath32 mixing , @xmath77 and @xmath78 decays . after taking into account constraints from @xmath79 , @xmath80 and @xmath81 , as well as
@xmath82 decays , the bounds on @xmath75 and @xmath76 are also obtained . for the sake of convenience , we recollect their numerical results in table [ nppara_value ] , with s1 and s2 corresponding to two fitting results of utfit collaboration for @xmath32 mixing @xcite . recently , cdf , d0 , and lhcb collaborations @xcite have updated the cp violation parameter @xmath83 in @xmath84 system .
these precise measurements will suppress the magnitude of @xmath85 coupling by about @xmath86 , and have no effect on the new weak phase @xmath74 .
however , the weak phase can be constrained by the data of @xmath77 and @xmath78 decays and the results are consistent with the previous refs .
indeed , the quantity that is directly related to the decay studied here is the product of the couplings of @xmath85 and @xmath87 , and the updated experimental data of @xmath84 mixing have less effect on it . according to the above analysis
, we will adopt the inputs parameters for the @xmath1 couplings as in table [ nppara_value ] in our theoretical calculation .
meanwhile , we also choose the extreme values of s1 which are named extreme limit values ( elv ) to show the maximal effects of @xmath1 contributions , and the elv are [ elv ] using the input parameters given above , we obtain the results of the branching ratios both in the sm and the family non - universal @xmath1 model without resonance contributions .
br(b_c d_s^ * ^+^- ) & = & \ { ll 2.32^+0.27_-0.2610 ^ -7 & ( sm ) , + 3.36^+0.38_-0.3510 ^ -7 & ( s1 ) , + 2.80^+0.32_-0.3010 ^ -7 & ( s2 ) , + 5.21^+0.57_-0.5410 ^ -7 & ( elv ) . .
[ br ] the theoretical errors are induced by the uncertainties of form factors . from the numerical results , one can see that branching ratio for decay @xmath0 is sensitive to the @xmath1 contributions .
with respect to the central value of the sm prediction , the new physics contributions in the family non - universal @xmath1 model can provide an enhancement about @xmath88 , @xmath89 , and @xmath90 for the case of s1 , s2 , and elv , respectively .
[ fig : dbranching ] shows the @xmath55 dependence of the differential decay rates for decay @xmath0 both in the sm and the family non - universal @xmath1 model using the central values of the input parameters .
the solid line refers to the sm prediction , while the dashed , dash - dotted , short - dashed curves correspond to the theoretical results of s1 , s2 , and elv , respectively . the @xmath1 enhancements to the differential decay rate are significant in almost the whole region of @xmath55 and strongly depend on the variation of np parameters . the @xmath55 dependence of forward - backward asymmetry for @xmath0 decay is presented in fig . [ fig : fb ] . compared to the sm results , when including the np effects from @xmath1 boson , the fba can be increased by about @xmath2 , @xmath3 , and @xmath4 at most in s1 , s2 , and elv , respectively .
it is easy to see that the zero crossing in @xmath91 also exists and @xmath1 corrections can shift @xmath92 in the sm to @xmath93 in s1 , and @xmath94 in s2 , respectively . as for the case of elv , the @xmath1 effects on @xmath91 are more significant and can lead zero crossing to vanish . in fig . [
fig : pl ] , we plot the longitudinal lepton polarization asymmetry of decay @xmath0 as a function @xmath55 both in the sm and the family non - universal @xmath1 model . after inclusion of the @xmath1 contributions , there are also apparent deviations in the values of the @xmath95 for all the cases in @xmath1 model from that of the sm predictions .
when @xmath5 , the value of the longitudinal polarization asymmetry can be changed from @xmath7 in the sm to @xmath8 in s1 , and @xmath9 in s2 , respectively . in the extreme case , the @xmath1 effects could flip the sign of the sm predictions when @xmath96 and the theoretical values might be close to zero in large momentum region .
the transverse lepton polarization asymmetry of decay @xmath0 as a function @xmath55 both in the sm and the family non - universal @xmath1 model is given in fig .
[ fig : pt ] .
the new physics corrections from @xmath1 boson are small , and will decrease the sm prediction about @xmath12 for the cases of s1 and s2 in low @xmath55 region
. however , for the case of elv , the decrease could be rather large and reach @xmath13 of the sm predictions .
in addition , the sign of @xmath11 will be changed in low momentum region and its values approach to zero when @xmath97 .
in this paper , we calculated the @xmath1 contributions to the branching ratio , forward - backward asymmetry and polarization asymmetries for @xmath0 decay in the family non - universal @xmath1 model by employing the effective hamiltonian with the form factors calculated in the three - point qcd sum rules . in section 2
, we presented the theoretical framework of @xmath35 transition including the effective hamiltonian , matrix element and the physical observables . in section 3 , we showed the numerical results of the observables and made phenomenological analysis for @xmath36 decay in the sm and the family non - universal @xmath1 model .
as expected , the @xmath1 contributions to the observables for @xmath0 decay could be significant in size . from the numerical results
, we found that : * with respect to the sm prediction , the @xmath1 contributions to the differential decay rates are significant in almost the whole region of @xmath55 and strongly depend on the variation of np parameters . * the new physics enhancements to fba could be large , and reach @xmath2 , @xmath3 , and @xmath4 at most in s1 , s2 , and elv , respectively .
the zero crossing could be shifted from @xmath92 in the sm to @xmath93 in s1 , and @xmath94 in s2 , respectively . as for the case of elv
, the @xmath1 effects could lead zero crossing to vanish . *
the values of @xmath6 deviated apparently from that of the sm predictions for all the cases in @xmath1 model . in high @xmath55 region , the values of @xmath6
could be changed from @xmath7 in the sm to @xmath8 in s1 , @xmath9 in s2 , and 0 in elv , respectively .
* the new physics corrections to @xmath11 would decrease the sm prediction about @xmath12 for the cases of s1 and s2 in low @xmath55 region .
however , for the case of elv , the decrease could be rather large and reach @xmath13 of the sm predictions .
one of the authors lin - xia l would like to thank prof .
zhen - jun xiao for his valuable help .
the work is supported by the national natural science foundation of china under grant no .
10947020 and 11147004 , and natural science foundation of henan province under grant no .
the auxiliary functions are given as follows @xcite : a()&= & _ 9^eff()a_v ( ) + c_7^eff t_1(),[ahs ] + b()&=&(1+_d_s^*)_9^eff()a_0 ( ) + ( 1-^2_d_s^*)c_7^efft_2 ( ) , [ bhs ] + c()&=&_9^eff()a_+ ( ) + c_7^eff(t_3()+ t_2()),[chs ] + e()&= & c_10^effa_v(),[ehs ] + f()&=&(1+_d_s^*)c_10^effa_0(),[fhs ] + g()&=&c_10^effa_+(),[ghs ] + h()&=&c_10^effa_-(),[hhs ] | using the form factors calculated in the three - point qcd sum rules , we calculate the new physics contributions to the physical observables of @xmath0 decay in a family non - universal @xmath1 model . under the consideration of three cases of the new physics parameters
, we find that : ( a ) the @xmath1 boson can provide large contributions to the differential decay rates ; ( b ) the forward - backward asymmetry ( fba ) can be increased by about @xmath2 , @xmath3 , and @xmath4 at most in s1 , s2 , and extreme limit values ( elv ) , respectively .
in addition , the zero crossing can be shifted in all the cases ; ( c ) when @xmath5 , the value of @xmath6 can be changed from @xmath7 in the standard model ( sm ) to @xmath8 in s1 , @xmath9 in s2 , and @xmath10 in extreme limit values , respectively ; ( d ) the new physics corrections to @xmath11 will decrease the sm prediction about @xmath12 for the cases of s1 and s2 , @xmath13 for the case of elv .
* key words : * rare decays , standard model , a family non - universal @xmath1 model |
herbig ae / be stars are young intermediate - mass ( 210 m@xmath1 ) stars which are still surrounded by gas and dust from their natal cloud .
many possess circumstellar disks which are believed to be the site of on - going planet formation .
the dust in these circumstellar disks , heated by the central star and possibly by viscous heating of material that is being accreted , shows up as excess emission above photospheric levels at infrared to sub - mm wavelengths .
the circumstellar gas can be traced in spectral lines , of which h@xmath0 is most prominent .
large - amplitude ( @xmath2 1 ) variations in optical brightness are seen in some herbig ae stars , and are commonly ascribed to circumstellar dust clouds moving in and out of our line of sight towards the central star .
although the optical to sub - mm energy distribution of herbig ae / be stars has been well explored by previous authors ( e.g. hillenbrand et al .
1992 ) , the chemical and mineralogical composition of the dust remained poorly studied until the 1995 launch of the _ infrared space observatory _ ( iso ; kessler et al .
this first possibility to study the complete infrared spectrum of these objects in detail revealed a large variety in dust properties , from small aromatic hydrocarbons to silicate dust .
moreover , some sources were shown to contain partially crystalline dust grains , similar to those found in comets in our own solar system ( malfait et al .
1998 , 1999 ; van den ancker et al . 1999 ; meeus et al . , these proceedings ) . here ,
the first inventory of solid - state features in all herbig stars observed by the _ infrared space observatory _ is presented and we investigate their correlation with more traditional tracers of circumstellar material .
an inspection of the iso data archive revealed the presence of spectroscopic data on 46 herbig ae / be stars , obtained with the short - wavelength spectrometer ( sws ) and the spectroscopic mode of the photometer ( isophot ) .
spectra were retrieved and reduced , after which they were inspected for the following features : ( a ) the emission bands at 3.3 , 3.4 , 6.2 , 7.6 , 7.8 , 8.6 , 11.3 and 12.7 @xmath3 m , often attributed to polycyclic aromatic hydrocarbons ( pahs ) , ( b ) the broad band around 10 @xmath3 m due to amorphous silicates , and ( c ) sharper emission bands at 10.2 , 11.4 , 16.5 , 19.8 , 23.8 , 27.9 and 33.7 @xmath3 m due to crystalline silicates . using these data
, we investigated the correlation of infrared spectral features with parameters of the systems from literature ( @xmath4 , @xmath5 , level of optical variability , h@xmath0 profile , and dust masses as traced by sub - mm fluxes ) .
the strongest correlation found is between spectral type of the central star and silicate emission : herbig stars of spectral type earlier than b9 show silicate absorption , whereas a large majority ( @xmath670% ) of the herbig ae stars of later type show silicate emission .
since strong optical variability due to variable circumstellar extinction is also only found in herbig stars with spectral type of b9 or later ( van den ancker et al .
1998 ) , both phenomena may be related .
however , no correlation between the level of optical variability and silicate emission could be found , perhaps due to our limited sample size . in most sources with absorption due to amorphous silicates ,
we also observe the absorption bands due to h@xmath7o and co@xmath7 ice , with a relative strength comparable to that in the interstellar medium .
however , two sources ( z cma and v645 cyg ) show strong silicate absorption , but no evidence for water or co@xmath7 ice bands , demonstrating the chemical evolution that has taken place in the circumstellar environment of these objects .
no strong correlation between spectral type and pah emission could be found .
this is surprising , since the excitation of pah molecules is thought to require intense ultraviolet radiation fields .
this means that on average the particles responsible for the pah emission must be closer to the central star , and hence suffer less geometric dilution of the stellar radiation field , in herbig ae stars than in herbig be stars .
both the differences in silicate and in pah behaviour can be explained by assuming that the infrared spectrum of herbig be stars is in general dominated by their circumstellar envelope rather than a disk .
in contrast , the more slowly evolving herbig ae stars have time to disrupt their envelope and their spectrum may be dominated by thermal emission from the protoplanetary disk .
crystalline silicates , as are also found in comets in our own solar system , are visible in 15% of the late - type herbig stars , all of systems that are relatively isolated and appear to be relatively old ( a few million years )
. therefore also in young stars longevity appears to be a prerequisite for the annealing process .
these systems form a close analog to the young solar system and may provide the strongest clue to date that the same processes that have led to rocky planets in our own solar system are also taking place around other stars .
bernasconi , p.a .
1996 , a&as 120 , 57 hillenbrand , l.a . , strom , s.e . ,
vrba , f.j . , keene , j. 1992 , apj 397 , 613 kessler , m.f .
, steinz , j.a . , anderegg , m.e .
, et al . , 1996 , a&a 315 , l27 malfait , k. , waelkens , c. , waters , l.b.f.m . , vandenbussche , b. , huygen , e. , de graauw , m.s .
1998 , a&a 332 , l25 malfait , k. , waelkens , c. , bouwman , j. , de koter , a. , waters , l.b.f.m .
1999 , a&a 345 , 181 van den ancker , m.e .
, de winter , d. , tjin a djie , h.r.e .
1998 , a&a 330 , 145 van den ancker , m.e .
, bouwman , j. , wesselius , p.r . , waters , l.b.f.m . , dougherty , s.m .
, van dishoeck , e.f .
1999 , a&a 357 , 325 | an analysis of solid - state features in infrared spectra of 46 herbig ae / be stars is presented .
the presence of solid - state emission bands is compared to other indicators of circumstellar material , such as h@xmath0 emission , optical variability and sub - mm continuum fluxes .
the correlation between these different indicators is weak , if present at all , in our sample .
however , a strong dependence on spectral type of the central star seems to be present : stars with spectral type earlier than b9 show either amorphous silicate in absorption or infrared spectra dominated by pah emission , whereas more than 70% of the stars of later spectral type show silicate emission .
we conclude that the infrared spectrum of herbig be stars is in general dominated by emission from the circumstellar envelope , whereas the lower - mass herbig ae stars show a spectrum that is dominated by a disk that is passively heated by the central star . |
_ boundary alignment _ is a critical feature of meshes in many applications . in a boundary
aligned mesh the boundary , or some other line , is traced by the sides of high - quality cells , see fig .
[ fig : bound_align ] . the definition of a well - shaped cell may be application dependent , but in many cases , cells similar in shape to squares ( for quadrilateral cells ) or equilateral triangles ( for triangles ) are preferred . characteristics of the entire mesh are also important , such as smooth cells - size and cell - shape transitions .
[ ptb ] the problem of producing boundary aligned meshes with well - shaped cells has been the subject of extensive research @xcite . still ,
many popular algorithms are heuristic in nature , and a more general understanding of the subject is called for , especially when quadrilateral meshes are considered .
a key difficulty is the problem s _ global _ character : the shape and position of every cell in the mesh is , at least in principle , related to that of any other cell . in a previous work @xcite , we described a relation between the problem of two - dimensional unstructured mesh generation , on both planar and curved surfaces , and another well - known problem , namely the _ inverse poisson _ ( ip ) problem .
the ip problem is concerned with reconstructing a _ source distribution _
@xmath0 of the poisson equation @xmath1 , from information on the potential @xmath2 at the boundaries . in that work
, the mesh was assumed to be _ conformal _ away from the irregular vertices ( vertices whose degree is different than four ) , like a grid mapped by a conformal mapping .
such grids have the property of having square cells in the limit of an increasingly finer grid . under this assumption ,
the problem of mesh generation was then shown to reduce to an ip problem .
the irregular vertices of the mesh correspond to point sources ( delta functions ) of @xmath0 , and @xmath2 is interpreted as the logarithm of the local resolution .
this theoretical framework turns the focus to the irregular vertices of the mesh : once their distribution is fixed , the continuum properties of the mesh - local resolution and directionality - are known .
note that it is also an explicitly global formulation , since the resolution at any given point is affected , via the function @xmath2 , by the locations of all irregular vertices in the mesh
. in this paper the generation of planar quadrilateral meshes is discussed .
resting on the results of @xcite , a new ip algorithm is presented , designed to construct source distributions of the appropriate type , which approximate the resolution and mesh directionality inputs at user - specified points , such as at the boundaries .
the ip algorithm is then incorporated into a complete mesh generation scheme , which also includes a technique for generating the final mesh .
an implementation is described , and shown in example cases to generate boundary aligned meshes , where well - placed point sources create smooth cell transitions and high quality cells .
a similar procedure is probably applicable to triangular meshes , but is not discussed in the present work .
remeshing of curved surfaces has recently attracted considerable attention ; for a review see @xcite .
many of the algorithms receive an input mesh directionality throughout the surface , usually the principal curvature directions of the surface .
this setting presents different challenges than those addressed here , since the mesh structure is determined , to a large extent , when mesh directionality is given everywhere in the domain .
for example , the locations of critical irregular vertices are dictated by the mesh directionality in the _ vicinity _ of these points .
another related subject is surface parameterization , concerned with creating mapping of surfaces to the plane .
conformal surface parameterizations are created in @xcite , but boundary alignment is not addressed .
the paper is organized as follows : section [ sec : ip_review ] shortly reviews the relevant theoretical background , with emphasis on the relation between the ip problem and unstructured mesh generation .
sections [ sec : algorithm ] and [ sec : create_final ] describe the proposed ip algorithm , and the mesh generation scheme . section [ sec : implementation ] describes an implementation of the algorithm , and section [ sec : examples ] gives examples of meshes generated . conclusions and possible directions for future research
are discussed in section [ sec : conclusions ] .
in this section an overview of the background theory will be given .
section [ sec : motivation ] explains the rationale underlying the mathematical formulation . in section [ sec : definitions ] , some key conformal geometry concepts are discussed , together with their relevance to the present problem .
section [ sec : ip_background ] summarizes the results developed in @xcite , relating mesh generation with the ip problem .
the exposition is limited to planar two - dimensional mesh generation ; a detailed account , in the more general setting of curved surfaces , can be found in @xcite .
we start by considering mesh generation using conformal mappings , which can be viewed as a special case of the theory to be described .
mapping techniques , in general , construct a function from one domain , for which a mesh already exists , to a second domain , which is being meshed . the mapping function
is then used to map the mesh into the second domain .
a key idea is that continuum properties of the mapping function control the shapes of the cells of the new mesh , at least for small enough cells .
for example , if the mapping is conformal , i.e. angle preserving , a cell with right inner angles ( rectangle ) will be mapped to a target cell with approximately right inner angles . in unstructured mesh generation
the connectivity of the mesh is not known in advance , and a more general framework is called for . in what follows ,
the interplay between the two domains which serves as a paradigm of mapping techniques , is replaced with an interplay between two definitions of distances on the input domain . instead of imagining the mesh as being the image of a mesh on a different domain , we _ redefine the distances _ on the domain to be meshed , and _
fix the cell edge length_. thus , the local mesh resolution will be proportional to the new local distance definition : a large new distance between two given points will mean more cells will be placed in between , hence a higher local resolution .
distances are redefined using the concept of a _ metric _ , known from riemannian geometry . since we focus on cells which are squares in the limit of an increasingly finer mesh , we need to define only two local properties : the resolution ( inverse of cell - size ) , and the direction of the cell - edges .
the new resolution , as noted before , is controlled by a new distance definition .
we use a new distance which is localy a scaling of the old distances : the new distance between a point and other nearby points is equal to the old distance , multiplied by a scalar factor which is independent of the direction .
thus , a small square measured in one distance definition is also ( approximately ) a square according to the new distance definition .
in riemannian geometry terminology , the old and new metrics are said to be _ conformally related_. the mesh directionality is related to the resolution , as is described in the following section .
it is convenient to work with a function @xmath3 , defined as the _ logarithm _ of the local scaling factor .
that is , a small square of side length @xmath4 measured with the original , cartesian distance definition , is a square with side length @xmath5 as measured with the new distance definition , see fig .
[ fig : lay_grid1],a .. if we imagine that the domain is covered with many small squares , all with the same side length @xmath6 in the new distance definition , the size of these cells in the original distance definition will be @xmath7 . the ( local ) resolution is the inverse of the local - cell - size , hence @xmath8 where the proportionality constant is a single number for the whole mesh .
the second continuum property is the local directionality of the cell edges .
the edges should ideally meet at right angles everywhere , except at _ irregular vertices _ , where the number of edges incident on the vertex is different than four .
it is therefore natural to assign to every point a set of four directions , mutually parallel or perpendicular .
this concept was expressed by many authors , and given various names , such as _ mesh directionality _
4-symmetry direction field _
@xcite , and _ frame field _ , although the last may refer to a structure which also holds cell - size information @xcite .
graphically , this object can be represented by a cross at every point , see fig .
[ fig : lay_grid1],a . , and here will be called a _ cross - field _ @xcite . on the plane
, the cross direction can be measured by the angle @xmath9 from the @xmath10-axis to one of the directions of the cross .
this angle is fixed up to an addition of @xmath11 radians , i.e. @xmath12 represent the same cross iff @xmath13 , for some integer @xmath14 .
[ ptb ] the function @xmath2 and the cross - field are not unrelated ; due to conformality , lines that trace the edge directions bend towards the side with smaller cells , see fig . [
fig : lay_grid1],a .. in the continuum theory , these lines are known as _
geodesics _ of the manifold .
geodesics are a generalization of the concept of straight lines to non - euclidian geometries . in the original , cartesian coordinates , the geodesic obeys the following differential equation : @xmath15 where @xmath16 is the curvature of the geodesic in the cartesian coordinate system , and @xmath17 is the derivative of the function @xmath2 in the direction normal to the tangent , see fig .
[ fig : lay_grid1],b .. eq . ( [ eq : kappa_dphi_dn ] ) allows one to calculate the change in the direction of the cross - field between two points connected by a geodesic .
there is also a direct way of calculating the change in the direction of a cross - field between any two points , along any curve connecting the two points , known as _ parallel - transport_. let @xmath18 be some curve from point @xmath19 to point @xmath20 , and let @xmath21 be the angles of the crosses at points @xmath19,@xmath20 respectively ( as noted above , the angles are defined up to an addition of a @xmath11 radians ) .
then @xmath22 where the integration denotes a line integral along the curve @xmath18 , according to the length parameter @xmath23 on @xmath18 , as measured in the original coordinate system .
the differential formulation of this equation is @xmath24 where @xmath25 are a pair of perpendicular vectors that form a right hand system .
where the function @xmath2 is defined , that is , at any point in the domain that is not a singularity , @xmath2 can be shown to _
harmonic _ , that is to obey the laplace equation : @xmath26 where , again , the derivatives are taken with the original coordinate system .
( [ eq : kappa_dphi_dn],[eq : phi_laplace ] ) are well - known results of conformal geometry @xcite,@xcite .
( [ eq : theta_change_par_tranport ] ) is derived in @xcite equations ( [ eq : kappa_dphi_dn],[eq : phi_laplace ] ) fully describe the relations between the cross - field and the function @xmath2 , at any regular point of the domain , that is , any point that is not a singularity of the function @xmath2 .
the singularities of the function @xmath2 are a key ingredient of the theory , since they correspond to the irregular vertices of the mesh , and unstructured meshes are those which contain irregular vertices . a detailed analysis of the possible of singularities of the harmonic function @xmath2 , and their effect on the resulting mesh , was carried out in @xcite , and shows that the only type of singularity that corresponds to a mesh with a finite number of cells is of the type @xmath27 , where @xmath28 is the location of the singularity . in the ip literature , such a sigularity is known as a _ point source_. furthermore , the prefactor of the logarithm is directly related to the degree of the irregular vertex in the final mesh . more specifically , suppose there are @xmath29 singularities in a domain @xmath30 , at points @xmath31 , then the function @xmath2 can be written as * condition 1 : * @xmath32 where @xmath33 is a harmonic function , @xmath34 , and @xmath35 .
the numbers @xmath36 are known as _ charges _ , so condition 1 states that the charges are integer multiples of @xmath11 .
we will refer to the numbers @xmath37 simply as the @xmath38 -values of the singularities .
the degree ( number of incident edges ) of the irregular vertex corresponding to the source at @xmath39 is equal to @xmath40 .
so , for example , irregular vertices of degrees 3 and 5 will correspond to a singularities with @xmath41 and @xmath42 , respectively . in a geometrical context
, a logarithmic singularity of @xmath2 represents a _ cone - point _ - the tip of a cone - in the new distance definition .
the charge corresponds to the angle deficit of the tip of the cone .
related subjects where manifolds with cone - points are considered include the study of disclinations in elastic media @xcite , and surface parameterizations @xcite .
an unstructured mesh is required to be aligned with the boundary .
namely , that one of the cross directions be parallel to the tangent to the boundary ( where a tangent exists ) : * boundary alignment definition : * @xmath43 where @xmath44 is a curve tracing the boundary , @xmath45 and @xmath46 .
since @xmath9 is fixed up to an addition of a multiple of @xmath11 radians , this definition depends only on the cross - field itself , and not on the particular choice of @xmath9 . requiring boundary alignment , as defined in eq .
( [ eq : bound_align ] ) can be shown to be equivalent to three more conditions on the function @xmath2 @xcite .
the following , condition 2 , applies to a point on a smooth section of the boundary , @xmath47 .
it roughly states that a smooth section of the boundary is a geodesic .
this provides information on the derivative of @xmath2 normal to the boundary : * condition 2 : * @xmath48 @xmath16 is the boundary curvature at @xmath49 .
the next condition is concerned with _ junction points _ of the boundary , where the tangent to the boundary is discontinuous .
it assures that the cross - field will be aligned with the boundary on both sides of the junction point .
let @xmath18 be a curve from point @xmath19 to point @xmath20 on both sides of a point @xmath50 of the boundary , see fig .
[ fig : cond_3_4],b .. then * condition 3 : * @xmath51 where @xmath52 is the angle of the boundary , and @xmath46 .
this means that for some inner angles @xmath2 will contain a singularity at @xmath50 ; at a distance @xmath53 from @xmath50 , the singularity will be of type @xmath54 .
[ ptb ] the final condition is concerned with the relation between two different boundary components .
it is just a restatement of eq .
( [ eq : theta_change_par_tranport ] ) for two points on two different boundary segments .
* condition 4 : * let @xmath19,@xmath20 be two points on two boundary curves @xmath55 .
let @xmath18 be a curve from @xmath19 to @xmath20 .
then @xmath56 conditions 1 - 4 form a complete set of conditions on @xmath2 , for a cross - field that is boundary aligned to exist .
that is , conditions 1 - 3 , and condition 4 on a number of selected curves , one from a selected point to each boundary component , are sufficient for a boundary aligned cross - field to exist , see @xcite .
the mesh generation scheme consists of the following steps : 1 .
setup and solution of the ip problem . finds the sources locations and charges . described in sections [ sec : input]-[sec : calc_locations ] .
solution of the direct poisson problem , to obtain the functions @xmath57 throughout the domain , see section [ sec : restore_harmonic ] .
generation of the final mesh , see section [ sec : create_final ] .
suppose that we are given an open domain @xmath30 , with boundary @xmath47 , and are given the required resolution function on @xmath47 .
( [ eq : phi_is_log_res ] ) , this is readily translated to the required value of @xmath2 on the boundary : @xmath58 condition 1 ( eq . ( [ eq : cond1 ] ) ) is a solution of the poisson equation @xmath1 , with point sources , whose locations and charges are yet unknown .
the problem is to find a distribution of sources ( location and charge ) adhering to the boundary alignment definition , eq .
( [ eq : bound_align ] ) , or alternatively to conditions 2 - 4 ( eq .
( [ eq : cond2])-([eq : cond4 ] ) ) , as well as to eq .
( [ eq : phi_from_res ] ) .
such a problem is known as an _ inverse poisson _ ( ip ) problem .
the ip problem may be compared with the _ direct poisson _
problem , where the source distribution is given , as well as some boundary information ( e.g. , dirichlet or newmann boundary conditions ) , and the value of the function @xmath2 is to be found . in the ip problem ,
the source distribution is unknown , and a source distribution adhering to the known boundary information is to be found .
ip problems have important applications in various areas of science and engineering @xcite,@xcite,@xcite,@xcite , @xcite . by its nature , the ip problem is ill - posed , and the solution might not be unique , and may be sensitive to small changes of the input , such as small changes in boundary conditions . in delicate problems of this type , any prior information on the source distribution may greatly affect the applicability of a specific solution procedure . in the present problem , we seek a point source distribution
. a number of algorithms for solving an ip problem with point sources appear in the literature . in @xcite , an inverse problem where all sources have the same charge is solved . in our implementation , however , at least two different charge values must be incorporated ( @xmath59 ) . in @xcite,@xcite,@xcite an inverse problem where both the locations of sources and their charges are unknown , and
are reconstructed .
this gives more freedom in the reconstruction than we can allow , since for the present purposes the charges must be multiples of @xmath11 .
another important aspect of the present problem is that the domain of the ip problem may be of any shape and topology ( i.e. , may contain holes ) , whereas the above works only deal with a simply connected region ( a circle , usually ) .
it is important to note that in the present application , in contrast to other standard applications , the input to the ip problem is not generated by some existing source distribution ( perhaps with some added noise ) , but by the domain s shape and input resolution .
the existence of a source distribution which reconstructs the input data , at least approximately , is therefore not obvious .
this is an interesting and important subject , but is beyond the scope of the present work .
as is well known , the real and imaginary parts of a complex analytic function are harmonic functions @xcite .
this correspondence has been utilized in ip algorithms @xcite,@xcite,@xcite , and will be used here as well .
we define the complex - valued function @xmath60 where @xmath61 , with @xmath62 the components of @xmath63 , that were defined in condition 1 .
@xmath4 is a function on @xmath30 , such that @xmath64 then , recalling that @xmath65 , it follows that @xmath66 with @xmath3 defined in condition 1 , and @xmath67 , with @xmath68 . the functions @xmath69 and @xmath70 for some @xmath71 are analytic in a neighborhood of any point that is not a singularity of @xmath2 .
however , as functions over the entire domain @xmath30 , they may be _ multi - valued_. multi - valued functions accept many - values at a point , depending on the path taken to that point .
it is well known that the complex function @xmath72 is multi - valued . defining @xmath73 as @xmath74 , where @xmath75 is a path from @xmath76 to @xmath77 , the imaginary part of @xmath72
is only fixed up to an addition of multiples of @xmath78 , depending on the path taken .
if the domain @xmath30 is not simply - connected ( i.e. , contains holes ) , then the function @xmath69 may also be multi - valued , since by following different paths around the holes , different values of @xmath79 may be obtained .
the function @xmath80 , being a sum of multi - valued functions , may also be multi - valued . according to eq .
( [ eq : re_f_is_phi ] ) , the real part of @xmath81 is single - valued , as it is equal to @xmath2 at that point .
we thus turn to examine the imaginary part of @xmath81 .
the real and imaginary parts of a complex function are related by the cauchy - reimann ( cr ) equations @xcite .
recall that for a complex differentiable function @xmath82 , where @xmath83 are real - valued , the cauchy - reimann equations read @xmath84 , and @xmath85 .
this allows one to recover the imaginary part of a analytic function if its real part is given .
writing eq .
( [ eq : phi_theta_right_hand_relation ] ) in the two right hand coordinate systems @xmath86 , gives the relations @xmath87 and @xmath88 .
these are precisely the cr equations for the complex function @xmath89 , where @xmath90 is any real constant , hence @xmath91 the constant @xmath90 is arbitrary , and will be taken to be zero .
the multi - valued nature of @xmath81 is explicit in the integral formulation in eq .
( [ eq : f_is_phi_i_theta ] ) . the right - hand - side ( rhs ) of eq .
( [ eq : f_is_phi_i_theta ] ) shows that the since the imaginary part of @xmath81 is @xmath92 , and @xmath9 must be fixed up to additions of @xmath11 , then the integral @xmath93 must also be fixed up to additions of @xmath11 . indeed ,
this was shown to hold if conditions 1 - 3 ( eq .
( [ eq : cond1])-([eq : cond3 ] ) ) hold , see @xcite . to summarize ,
the problem is restated as finding a multi - valued complex function @xmath81 , given by eq .
( [ eq : f_def ] ) , and whose value at points on the boundary of @xmath30 is : @xmath94 in the following sections , an algorithm for solving the problem as it was here restated is described .
the ip algorithm first removes the contribution of @xmath69 from the boundary conditions , paying special attention to the junction points , and then constructs the source distribution .
the function @xmath2 , and hence the function @xmath81 , may be singular at junction points .
special procedures for addressing this behavior are taken as part of two different steps of the algorithm : a. : : choosing the input resolution near junction points .
b. : : special treatment of junction points when @xmath4 is removed from the boudary conditions .
the two subjects are discussed in the following subsections , [ sec : input_near_junc ] and [ sec : remove_harmonic_junc ] . except in the special cases when @xmath52 is a multiple of @xmath11 , the behavior of @xmath2 at a small neighborhood in @xmath30 of a junction point is singular , with the singularity at the junction point ,
see condition 3 ( eq . ( [ eq : cond3 ] ) ) .
if the input resolution in a small neighbourhood of the junction point does not match this singular behavior , solution of the ip problem will feature sources at any distance from the junction , no matter how small , resulting in a distribution with an infinite number of sources . in practice , this means that an ip algorithm will cluster many sources near the junction point , in a futile attempt to reconstruct the boundary conditions there . to avoid this problem ,
we adjust the _ input _ @xmath2 at a neighborhood of the junction point
. this adjustment should vanish rapidly at a distance larger than 1 - 2 cell - sizes , so as to have little effect on the original required resolution .
condition 3 states that the gradient of @xmath2 near a junction point must diverge as @xmath95 , where @xmath53 is the distance from the junction point .
such a flux is formed by a logarithmic singularity at the junction point ( see also @xcite , section 7.1 ) .
consider a singular source term of the form @xmath96 , where @xmath53 is the distance from the junction point .
then the flux through a circular arc @xmath97 , at distance @xmath53 from the junction point is ( see also fig . [
fig : cond_3_4],a . ) : @xmath98 where @xmath52 is the junction inner angle .
then by condition 3 , @xmath99 , hence @xmath100 @xmath101 is integer , which must be positive , otherwise a mesh with an infinite number of cells will result ( @xcite , section 7.1 ) .
in fact , @xmath101 has a simple interpretation : it is the number of cells incident upon the junction point in the resulting mesh ) nearby cone - points may be joined , and cone - points may be shifted to the boundary , depending on the final cell - size .
in such a case , the number of cells incident on a junction point , as well as on other boundary points , may change . ] .
a reasonable choice for @xmath101 is therefore @xmath102 for which the inner angles of the cells incident on the junction point are closest to @xmath11 . in order to restrict the effect of this correction of @xmath2 to a small region in @xmath30 , a source term with the opposite charge @xmath103
can be placed _ outside _
@xmath30 , at a distance of about 1 - 2 edge lengths .
see for example the second and third examples in section [ sec : examples ] . in section [ sec :
remove_harmonic ] a method will be described for calculating and removing the contribution of @xmath69 to the boundary conditions . at junction points ,
however , the function @xmath79 may be singular : as @xmath30 is an open set , boundary singularities are part of @xmath69 , so care should be taken when preforming the calculations described in section [ sec : remove_harmonic ] , especially in a numerical implementation of the technique . to avoid these problems
, we subtract the junction singularities from the boundary value of @xmath80 before proceeding with removing @xmath69 .
let @xmath104 be the locations of the junction points , and @xmath105 the charges as given by eq .
( [ eq : junc_charge ] ) , ( [ eq : junc_incidence ] ) .
the complex function corresponding to @xmath106 is @xmath72 , so the contribution of the junctions is @xmath107 if a singularity is located on an inner boundary component , the logarithms must contain a branch cut somewhere in @xmath30 . in order to avoid adding a cut , we add another term to every hole , equal to minus the total charge in each hole .
denote the inner boundaries by @xmath108 , and let @xmath109 be arbitrary points inside the respective holes .
the following terms are therefore added to @xmath80 : @xmath110 in this way , the additional terms form an analytic function in @xmath30 , with no branch cut , and are a part of @xmath69 .
since the contribution of @xmath69 is removed from @xmath80 as described in the next section , the exact form of this additional term , i.e. the choice of points @xmath111 , does not affect the results . in this section a technique for calculating @xmath69
is presented .
after @xmath69 is calculated , its contribution to @xmath80 at the boundary can be subtracted . to simplify notation ,
let @xmath112 denote the sum of logarithms in @xmath113 : @xmath114 in the context of ip algorithms , the idea that the contribution of the harmonic part to the solution can be removed from the boundary information was suggested in @xcite .
there , an ip problem in the unit disk @xmath115 is discussed .
the value of the analytic part of @xmath80 on the boundary of @xmath115 can then be calculated by taking the fourier transform of @xmath80 , and leaving only the positive frequency components , as can be shown by considering the lourent series of @xmath80 in the unit disk . by its construction
, this technique applies to functions defined on the unit disk .
it can be extended to other domains , if a conformal mapping of the domain to the unit disk ( which exists according to the riemann mapping theorem ) is calculated . for our present purposes , since the domain @xmath30 in every problem input is different , using this technique would require constructing a mapping to the unit disk for each domain being meshed separately . furthermore , the domains in our problem may contain holes , which further complicates the matter .
we therefore use an alternative approach , which is now described .
it is based on the cauchy integral theorem , and is similar to the sum factorization step in the weiner - hopf technique @xcite , applied to a bounded domain ( @xmath116 , defined below ) in place of an infinite strip . as in @xcite , we assume there is a positive distance @xmath117 between the boundary of @xmath30 and the source closest to it , as is the case , for example , when the number of sources is finite . let @xmath116 be the set of points at a distance smaller than @xmath117 to the boundary , see fig .
[ fig : cauchy],a .. as will become apparent later , the value of @xmath117 does not enter the calculation and is irrelevant , as long as there exists some @xmath118 as required .
[ ptb ] we first consider a simply - connected domain @xmath30 , with boundary @xmath47 .
define the flux through any curve @xmath18 : @xmath119 the flux @xmath120 through the boundary can be calculated : the value of @xmath17 at smooth boundary points is known from condition 2 ( eq .
( [ eq : cond2 ] ) ) , and the flux at junction points is calculated from condition 3 , as is explained in section [ sec : input_near_junc ] above .
this flux can be shown to be a multiple of @xmath11 .
[ remark : flux_calc]using conditions 2,3 and the fact that the rotation of a tangent in a simple curve is @xmath121 , the total flux through a boundary component @xmath122 can be shown to be equal to @xmath123 , where @xmath124 are the junction points , @xmath125 for each junction is given by eq .
( [ eq : junc_incidence ] ) , and @xmath126 for inner and outer boundary components , respectively .
see also @xcite . to proceed with the decomposition calculation
, we assume that @xmath127 , as calculated from eq .
( [ eq : bound_flux_def ] ) above .
if this is not the case , following @xcite , we subtract a source term to the boundary conditions , centered at some point @xmath128 inside the domain : @xmath129 such that @xmath120 is zero .
as will be shown below , the choice of @xmath128 does not affect the results . since @xmath120 was equal to a multiple of @xmath11 before , @xmath130 . when @xmath131 , the function @xmath80 is single - valued and analytic in @xmath116 , as follows from the integral representation of @xmath80 in eq .
( [ eq : f_is_phi_i_theta ] ) , along with @xmath131 .
we now use the cauchy s theorem , stating that for every point @xmath132 : @xmath133 the boundary @xmath134 of @xmath116 has two boundary connectivity components : the outer boundary ( which is also the boundary of @xmath30 ) , and the inner boundary . denote them by @xmath135 and @xmath136 respectively , see fig .
[ fig : cauchy],a .. eq . ( [ eq : cauchy_d_simply_con ] ) now reads @xmath137 the function @xmath138 is harmonic on @xmath116 .
according to a decomposition theorem ( see @xcite , chapter 9 ) its decomposition into @xmath139 and @xmath140 , is also unique .
it follows that the decomposition of the multi - valued complex function @xmath81 into @xmath141 and @xmath4 is unique , up to the arbitrary constant @xmath90 in eq .
( [ eq : f_is_phi_i_theta ] ) , chosen before to be zero .
the two integrals expressions on the rhs of eq .
( [ eq : cauchy_on_omega ] ) correspond exactly to the two components of the ( unique ) decomposition of @xmath80 in @xmath116 as described in @xcite , hence @xmath142 where we have used the fact that @xmath143 .
we would like to find @xmath141 on @xmath47 , but @xmath136 is not known , nor is @xmath80 on @xmath136 , so @xmath112 can not be computed directly from the second equation .
the first equation can however be used , since @xmath80 is given on @xmath47 , so @xmath69 can be calculated on @xmath47 ( more precisely , since @xmath4 is defined in @xmath30 , it is the limit of @xmath4 as @xmath47 is approached ) .
then @xmath112 on @xmath47 is given by @xmath144 , according to eq .
( [ eq : fp_def ] ) . finally , we add back the source term subtracted before @xmath145 we now turn to the case when @xmath30 is not simply - connected .
we add source terms : one inside the domain , as discussed above ( eq . ( [ eq : source_in_d ] ) ) , and one inside every hole ( i.e. _ outside _
@xmath30 ) , such that the flux @xmath146 through every boundary connectivity element is zero , see fig .
[ fig : cauchy],b .. @xmath147 the @xmath111 can be the same points used in section [ sec : remove_harmonic_junc ] , or other points .
the results do not depend on the additional charges locations , as will be explained below .
we now introduce cuts so that the boundary contains a single connectivity element ( the cuts play a part in the derivation , but drop out of the final calculation ) , see fig .
[ fig : cauchy],b .. the cuts introduced also serve as the branch cuts of the logarithms .
denote this new boundary @xmath148 .
using @xmath149 , we proceed as when @xmath30 is simply connected : define @xmath116 as before , see fig .
[ fig : cauchy],b . , and subtract the source term inside @xmath30 , as in eq .
( [ eq : source_in_d ] ) . according to eq .
( [ eq : cauchy_2_parts ] ) : @xmath150 the second integral on the rhs denotes the integral over the cuts introduced to form @xmath149 .
note that each cut - path is traversed twice , back and forth .
since the flux @xmath146 through each and every hole boundary is zero , the value of @xmath151 , when integrated along @xmath149 , is continuous across the branch cuts , and the integrations over each cut traversed in both directions cancel each other , and drop from the total integration .
therefore @xmath152 as in eq . ( [ eq : cauchy_2_parts ] ) . as before , @xmath144 .
finally , the source term inside @xmath30 is added back , as in eq .
( [ eq : source_in_d_back ] ) .
the boundary value of @xmath141 obtained after this source term is added back may be multi - valued , with a branch cut discontinuity in the imaginary part that is a multiple of @xmath11 .
this is a valid input to the next step , as explained in section [ sec : calc_locations ] below ( following eq .
( [ eq : xsi_def ] ) ) .
the choice of the added singularities locations outside @xmath30 ( inside the holes ) does not affect the results : changing the location of a singularity with some charge from @xmath153 to @xmath154 is equivalent to adding a singularity with opposite charge at @xmath153 , and with the same charge at @xmath154 . since these two singularities lay outside @xmath116 , and
are of opposite charge , this amounts to adding an analytic function to @xmath81 , which is removed by the cauchy integral technique described above .
the result is also unaffected by the location @xmath128 of the source term subtracted inside @xmath30 , since its location only affects the @xmath141 and it is later added back . the choice of cuts in eq .
[ eq : cauchy_with_cuts ] , of course , does not affect the cauchy integral calculation , since the cuts do not enter the final calculation , eq .
( [ eq : cauchy_final_calc ] ) . at this point
, we are given the value of @xmath112 on the boundary @xmath47 of the domain @xmath30 .
( [ eq : f_def ] ) , we define @xmath155 = \exp\left [ { \displaystyle\sum\limits_{m=1}^{n_{c } } } k_{m}\ln\left ( z - z_{m}\right ) \right ] \text{.}\label{eq : xsi_def}\ ] ] note that while @xmath141 may be have a multi - valued imaginary part with a branch cut discontinuity : @xmath156 , for some integer @xmath14 , @xmath157 is single valued , since @xmath158 = 1 $ ] .
since the @xmath38-values , @xmath37 , are integers , we can assume without loss of generality that @xmath159 .
other charges may be formed by placing several sources at the same location .
we can therefore write eq .
( [ eq : f_def ] ) as : @xmath160 where @xmath161 are the number of sources with @xmath162 respectively . according to the divergence theorem
, the difference @xmath163 is equal to @xmath164 ( this can also be shown directly by calculating the integral of @xmath141 along @xmath47 ) .
since @xmath120 is known ( see also remark [ remark : flux_calc ] ) , @xmath163 is known .
therefore , to fix @xmath161 only one number remains to be chosen , e.g. @xmath165 .
the choice of the total number of sources affects the accuracy of the reconstruction of the input data , see sections [ sec : ipsummary ] and section [ sec : implementation].[remark : divergence ] substituting eq .
( [ eq : fp_only_plus_minus ] ) , eq .
( [ eq : xsi_def ] ) now reads @xmath166 in eq .
( [ eq : rational_roots ] ) , the numbers @xmath167 are the roots of polynomials , and we proceed by presenting the polynomials in a different form . later ,
once the polynomials are found , the @xmath168 are recovered by finding the roots of the polynomials .
we rewrite eq .
( [ eq : rational_roots ] ) as @xmath169 where @xmath170 are unknown coefficients .
note that the prefactors of @xmath171 are indeed equal to @xmath76 , as can be seen by expanding the polynomials in eq .
( [ eq : rational_roots ] ) . to find the unknown coefficients numerically , we discretize eq .
( [ eq : rational_standard_poly ] ) , and give @xmath157 at @xmath172 points @xmath173 , @xmath174 , and write @xmath175 : @xmath176 the rhs of eq .
( [ eq : rational_discrete ] ) is a _ rational function interpolation _ of the @xmath177 data ( see e.g. @xcite ) . in the implementation described below ,
the unknowns @xmath178 are evaluated as follows . rearranging , eq .
( [ eq : rational_discrete ] ) reads @xmath179 these are @xmath172 linear equations for the @xmath165 unknowns : @xmath178 .
since @xmath172 will typically be larger than @xmath165 , the solution will only be approximate , e.g. , a solution in the least - mean - square ( lms ) sense . once the unknown variables @xmath178 are found , the source terms locations @xmath180 are calculated by finding the roots of the two polynomials appearing in the rhs of eq .
( [ eq : rational_discrete ] ) . a solution in the lms sense , as described in eq .
( [ eq : linear_system ] ) , is not recommended if the resulting errors in the @xmath177 s are large , when compared with @xmath177 .
this is due to two reasons : first , the input data @xmath177 appears in both sides of eq .
( [ eq : linear_system ] ) , so the errors in the lms solution might not reflect the errors in @xmath177 , and moreover , the lms solution may be sensitive to the choice of the coordinate system origin .
secondly , an error in @xmath177 scales as a difference in the resolution ( since @xmath181 ) , rather than the more natural error definition given by the _ ratio _ of input and obtained resolution .
this means that at low resolutions ( large cell - sizes ) , errors in @xmath182 might represent large relative resolution deviations .
once the sources locations have been determined , the function @xmath2 can be calculated everywhere in the domain .
the solution will usually be approximate , i.e. , the resulting sum of sources will only approximate the required @xmath183 , and some deviation from both boundary alignment and resolution requirements may be found . the trade - off between the two requirements
can be partly controlled by choosing an appropriate harmonic @xmath33 in eq .
( [ eq : cond1 ] ) . in order to satisfy the cell - size requirements exactly
, one can solve the ( direct ) poisson equation , given by eq . ( [ eq : cond1 ] ) , but this time with a known charge distribution , and with dirichlet boundary conditions , i.e. by specifying @xmath2 on the boundary .
a different approach is to try and satisfy the boundary alignment conditions , by solving eq .
( [ eq : cond1 ] ) with newmann boundary conditions .
this gives a new @xmath2 on the boundary .
the difference between the input ( required ) @xmath184 , and the @xmath2 obtained on the boundary can be used to evaluate the quality of the source distribution calculated by the ip algorithm , and the total number @xmath165 of sources can be changed accordingly . note that when @xmath30 is not simply connected , solving @xmath33 with neumann boundary conditions ( condition 2 ) does _ not _ mean that exact boundary alignment is obtained , because condition 4 might not be exactly fulfiled .
the steps followed in calculating the source distribution ( locations and charges ) can be summarized as follows : 1 .
adjusting the input resolution near the junction points ( section [ sec : input_near_junc ] ) .
2 . subtracting the junction source terms ( sections [ sec : remove_harmonic_junc],[sec : remove_harmonic ] ) .
@xmath185 3 . calculating the cauchy integral to find @xmath69 on the boundary .
finding @xmath183 on the boundary .
( section [ sec : remove_harmonic ] ) .
adding back the source inside @xmath30 ( section [ sec : remove_harmonic ] ) : @xmath186 5 . calculating the locations of sources ( section [ sec : calc_locations ] ) , using some initial total number of sources @xmath165 .
6 . solving the ( direct ) poisson problem to find the new @xmath81 approximating the original @xmath81 .
according to the quality of reconstruction , step 5 can be repeated with a different total number of charges .
once the function @xmath187 is set throughout the domain , the final mesh can be constructed .
there are various possible approaches to this problem .
we present a simple method that was used to create the examples in section [ sec : examples ] . in this method
, the domain is cut along geodesics that follow the cross - field directions .
every source is at the end of a cut , and every inner boundary component is connected to the outer boundary by a cut , see fig .
[ fig : mapping_sketch],a .. the direction of the geodesics emanating from the sources and directed along the cross - field are calculated as explained in appendix a. the cut domain , which we denote by@xmath188 , does not contain any sources in its interior , and its boundary consists of a single connected component .
a conformal mapping @xmath189 of @xmath190 into the plane can be constructed , with a conformal factor ( local scaling ) of @xmath191 .
we consider @xmath190 and @xmath192 as lying in the complex plane .
first , note that since @xmath190 contains no sources in its interior and is simply - connected , @xmath193 is uniquely defined in @xmath190 .
denote by @xmath194 this single - valued @xmath193 . to calculate the mapped boundary @xmath192
, we calculate the function @xmath195 along the ( single ) boundary @xmath196 of @xmath190 : @xmath197 [ ptb ] if this mapping is used to create a mesh , by placing a cartesian grid on @xmath192 , and mapping it back to @xmath190 by taking the inverse of @xmath189 , the mesh obtained will be invalid : the grid will not be continuous across the cuts , will contain invalid cells at the sources , and cut cells at the boundaries . to correct this , and obtain a valid final mesh
, we seek a function @xmath198 on @xmath196 approximating @xmath199 .
a related problem , for surface meshes without boundary alignment , was addressed in @xcite .
first , note that because the cuts are added along geodesics aligned with the cross field , @xmath199 is a piecewise straight line , whose straight segments are either horizontal or vertical . in order to obtain a valid final mesh , the new function @xmath198 should have the following properties : 1 .
@xmath198 must map the closed path @xmath196 to a closed path .
2 . @xmath200 must be a piecewise straight line , composed of horizontal and vertical lines only .
to each segment in @xmath201 corresponds a segment in @xmath202 , at the same order along the paths .
each segment in @xmath200 must be directed in the same direction as the corresponding segment in @xmath203 , or be of zero length .
if @xmath204 is a geodesic cut of @xmath196 , it is followed in both directions : @xmath204 and @xmath205 .
then @xmath206 and @xmath207 must have the same length .
this ensures the same number of cells on both sides of a single cut .
we describe a simple algorithm for creating @xmath208 , meeting these requirements .
to create @xmath200 , the path @xmath199 is followed , and the segments of @xmath209 are modified one after the other , in the order they appear in @xmath210 , see fig .
[ fig : mapping_sketch ] .
each segment is shifted to a parallel line of the integer grid .
this is usually the nearest segment , except in the case the length of the segment directly before this segment dictates a different shift , according to properties ( 3,4 ) above , see fig . [
fig : mapping_sketch],c .. this algorithm requires that at least one junction exist in @xmath47 , which is mapped to a right angle in @xmath201 .
the algorithm starts from the segment following this junction , so that when the last segment is reached and shifted , a change in the length of the first segment will be allowed .
the resulting modified path is similar to @xmath211 , only with different segments lengths .
@xmath200 is then defined as the composition of @xmath199 followed by a linear mapping of each segment of @xmath203 to the corresponding segment in the modified path .
once @xmath212 is known , @xmath213 can be defined . in the implementation described below this
is done by placing a triangular mesh in @xmath190 , and solving the laplace equation for both @xmath214 and @xmath215 on @xmath190 .
@xmath198 obtained in this way is not in general analytic , since the cr conditions might not hold between @xmath214 and @xmath216 , so the final mesh will not exactly conformal , but only approximately , as @xmath198 approximates the analytical function @xmath217 .
the mesh edges are then extracted by tracing the integer - valued lines of @xmath218 and @xmath219 .
once the connectivity of the mesh has been established , a final smoothing procedure which involves all the interior vertices , including those on the cut path , is preformed . in the definition of @xmath198 ,
both @xmath214 and @xmath215 are harmonic functions .
a standard finite difference approximation to a harmonic function on a uniform grid , assigns to each vertex the average of its four neighboring vertices .
this is applied to all vertices inside the mesh , including those on the cuts of @xmath196 , which can be interpreted as a generalization of the defintion of @xmath198 ( to be more precise , in this smoothing procedure it is @xmath220 which is assumed to be harmonic ) .
this procedure is the well - known laplacian smoothing procedure , in which the location of every vertex is equal to the average location of its neighboring vertices , and in this application it is theoretically justified if @xmath221 is approximately @xmath217 , i.e. if @xmath212 is close to @xmath199 .
the algorithm described in the previous section was implemented in matlab @xcite , using standard matlab functions , such as a least - mean square solver for linear systems , an ode solver , etc .. to calculate the cauchy integral ( section [ sec : remove_harmonic ] ) , the boundary was approximated by a piecewise linear path between sampled points , with a constant value of @xmath2 on each segment .
this allows the integral over each segment of the path to be calculated analytically , and the path integral there is equal to the sum of the integral over the segments .
this is a low order approximation , but was sufficient for the present purposes .
the linear system , eq . ( [ eq : linear_system ] ) , was solved in the least - mean - square sense , using matlab s ` mldivide ` function .
note that the coefficients in this system contain powers of the @xmath173 s , the boundary data locations .
this leads to ill - conditioned linear systems , which are more sensitive to the numerical round - off errors , as the number of sources @xmath161 increases , or the range of @xmath222 increases .
tests indicate that when the solver issues a warning , indicating that the matrix is rank deficient to the working precision , the results of the ip algorithm become unreliable ( this was not the case in the examples in section [ sec : examples ] ) . for input domains where many sources or a high @xmath222 range are required ,
a different solver implementation may be needed , such as a varying precision arithmetic computation , working at higher floating - point accuracies .
after sources locations are recovered by taking the roots of the polynomials , the function @xmath33 of eq .
( [ eq : cond1 ] ) is calculated , given the newmann boundary conditions in eq .
( [ eq : cond2 ] ) .
this was done using the method of fundamental solutions ( mfs ) , see e.g. @xcite , where the solution to the laplace for @xmath33 is approximated by a sum of ( real ) source terms , i.e. functions of the form @xmath223 , with the locations @xmath224 lying _ outside _ the domain @xmath30 where @xmath33 is harmonic .
the charges @xmath225 are then calculated to best fit the boundary conditions . once the function @xmath2 is given everywhere in @xmath30 , the directions of the star - geodesics at each source are calculated , as described in appendix a. the star - geodesics , emanating from the source were then traced by solving the differential equation , eq .
( [ eq : kappa_dphi_dn ] ) , using a runge - kutta ode solver .
the cut - tree was constructed by taking one source at a time at some arbitrary order , and introducing a cut along a star - geodesic of the added source , that is closest to the cut - tree constructed so - far .
this construction of the cut tree is admitantly arbitrary , and may not be optimal in some cases .
lastly , the algorithm described in section [ sec : create_final ] is applied .
in the first example , a domain with two boundary components is meshed , see fig .
[ multi_diamond_var_res],a .
the side lengths of the outer and inner squares are @xmath226 and @xmath227 respectively , centered at the origin , and the inner square is rotated at angle @xmath228 radians . the ratio of input resolution between outer and inner boundaries is @xmath229 .
the inner and outer boundaries were sampled with @xmath230 and @xmath231 points , respectively , which were used both for calculating the cauchy integral , and for the sources locations calculation .
singularities of @xmath2 are not required at the junctions points , since all inner angles are multiples of @xmath11 radians . as there is no flux of @xmath232 through either boundary ( because the boundaries curvature is @xmath233 , see eq .
( [ eq : cond1 ] ) , and @xmath2 is regular at the junctions ) , the total charge inside is zero ( see remark [ remark : divergence ] ) .
hence an equal number of @xmath234 and @xmath235 charges were used .
[ multi_diamond_var_res],a . , shows the domain boundary , together with the calculated sources locations .
squares mark sources with @xmath234 , and circles sources with @xmath235 .
the distribution is composed of @xmath236 sources of each type , i.e. a total of @xmath237 sources .
the value of @xmath2 obtained by solving the poisson equation with this charge distribution is shown in fig .
[ multi_diamond_var_res],b . , and
compared to the input @xmath2 requested , denoted by @xmath238 .
the difference @xmath239 is also plotted . for visual clarity ,
the average of the @xmath238 was subtracted from both @xmath2 and @xmath238 ( this is just a scaling of the resolution by a constant factor ) ; this was also done in the resolution comparisons below , in fig .
( [ fig : multi_ex2]),c . , fig .
( [ fig : multi_ex3 ] ) , b . and fig .
( [ fig : multi_ex4]),c .. the number of sources was chosen to be the smallest for which @xmath240 at all boundary points .
this criterion is used for choosing the number of sources is also used in the following examples . the cut - tree used in creating the final mesh
is shown in fig .
[ multi_diamond_var_res],c . , and the final mesh in shown fig .
[ multi_diamond_var_res],d .
[ ptbh ] the domain in the second example is the union of three unit radius circles which pass through the origin , see fig .
[ fig : multi_ex2],a .
the boundary has three junctions , all with the same inner angle , @xmath241 .
boundary alignment requires @xmath2 to be singular at the junctions , corresponding to sources with charge @xmath242 , or @xmath38-value @xmath243 , see eq .
( [ eq : junc_charge]),([eq : junc_incidence ] ) .
as explained in section [ sec : input_near_junc ] , singular functions can be added the input resolution in order to avoid clustering of sources near the junction points . in this example , we compose the input @xmath2 from a sum of pairs of source terms , with @xmath38-values as marked in fig .
[ fig : multi_ex2 ] , a .. the pairs of nearby sources with opposite charge create the desired singularities at the junctions , and have only a small effect at distances large compared with the distance between the sources of each pair .
we can now calculate the total flux of @xmath2 through the boundary in this example .
the ( singular ) flux through each junction is @xmath244 ( positive value when @xmath232 is directed outward ) .
the flux through each of the three arcs is @xmath245 along the arc , and since for each arc @xmath246 , the total flux through the boundary is @xmath247 .
this is in agreement with the formula given in remark [ remark : flux_calc ] .
therefore , the sum of @xmath38-values of the sources inside the domain should be @xmath248 , so @xmath249 . hence , @xmath250 of eq .
( [ eq : source_in_d ] ) was @xmath248 ; @xmath251 was used .
the lowest number of sources for which @xmath240 , shown in fig . [
fig : multi_ex2],b .
, has @xmath252 , @xmath253 .
the boundary was sampled at @xmath254 points .
[ ptbh ] the boundary in the third example is composed of two boundary components , each with a single junction point .
the inner angles at the inner and outer boundary junctions are @xmath255 respectively , so a singularity of @xmath2 will form at the inner boundary junction . as in the previous example , we add a singular function to the input @xmath2 , by adding a pair of source terms , one at the junction point with @xmath256 , as obtained from eq .
( [ eq : junc_charge]),([eq : junc_incidence ] ) with @xmath257 .
no inner source term is required , and from the flux trough the hole boundary it follows that @xmath258 . after subtracting the junction source term with @xmath256 ( see section [ sec : junc_points ] ) a source with total charge of @xmath259
is added inside the hole , see fig .
[ fig : multi_ex3],a .. the inner and outer boundary components are sampled at @xmath260 and @xmath261 points , respectively .
the distribution with @xmath262 , the smallest number of sources for which @xmath240 , is plotted at fig .
[ fig : multi_ex3],a . , with squares and circles representing charges as in previous examples .
two pairs of @xmath263 charges are located inside the inner boundary , i.e. , outside the domain .
their constitute a correction to the homogeneous @xmath33 , which is very small , since in each pair the opposite charges almost overlap .
the obtained @xmath2 reconstruction is given in fig .
[ fig : multi_ex3],b .. the final mesh , fig .
[ fig : multi_ex3],c . , contains @xmath264 irregular vertices , @xmath265 of each type .
this is because two pairs of sources lay outside the domain , and another pair ( the pair of sources closest to the inner junction ) was eliminated in the creation of the final mesh ( section [ sec : create_final ] ) : the two opposite charges where united , giving a zero total charge . [ ptbh ] in the final example the boundary of the square @xmath266\times\lbrack-2,2]$ ] was assigned a varying input resolution proportional to : @xmath267 the resolution thus varies by a factor of @xmath268 within the domain . as always , eq .
( [ eq : phi_from_res ] ) , @xmath269 .
the total charge is zero , so @xmath270 .
the boundary was sampled at @xmath271 points .
a source distribution with @xmath240 was obtained with just @xmath226 charges , i.e. with @xmath272 .
[ fig : multi_ex4],a . shows the source distribution , fig .
[ fig : multi_ex4],b . shows the obtained @xmath2 vs. the input @xmath2 , and the final mesh is shown in fig .
[ fig : multi_ex4],c .. it is instructive to compare the star - geodesics and cut - tree with the final mesh .
the two are overlaid in fig .
[ fig : multi_ex4],d .
( this would create a crowded appearence when more sources are present ) .
the black solid line traces the cut - tree , the dashed lines are additional star - geodesics , and the mesh is drawn with the background in gray lines .
[ ptbh ] cell shape quality was measured using a variant of the quality measure @xmath273 @xcite , as defined in @xcite .
@xmath274 represents a square cell , while @xmath275 represents a cell with an inner angle of @xmath276 .
cells with a high aspect ratio are also given low @xmath273-values .
table 1 shows the minimum and average @xmath273 values , and the total number of cells , for the examples in fig .
[ multi_diamond_var_res]-[fig : multi_ex4 ] . for each example , the tabulated information is given for the mesh shown in the corresponding figure , and for a finer mesh of the same domain , where the resolution function was doubled ( prodcing a mesh with about four times the number of cells ) .
note that multiplying the resolution by a factor amounts to adding a constant to @xmath2 , which does not change the source distribution obtained from the ip algorithm , as this constant , which is part of @xmath69 , is readily removed when @xmath69 is subtracted .
@xmath277{|c|c|c|c|}\hline & num .
cells & min .
$ ] @xmath278@xmath279 @xmath280 table 1 : mesh statistics .
an unstructured quadrilateral mesh generation scheme in planar domains was presented .
the method rests on a theoretical foundation , linking the mesh generation problem with the inverse poisson ( ip ) problem .
an ip solution algorithm is presented , whose output is interpreted as the location and type ( degree ) of irregular vertices in the domain .
the continuum fields obtained , describing mesh resolution and directionality , are conformal everywhere except on the irregular vertices , and fit the required input properties at the boundary , or at other user - defined locations .
an algorithm for creating a valid final mesh is also presented .
example meshes feature irregular vertices where they are needed , in combination with highly regular regions where possible .
directions for future work include more sophisticated methods for solving the rational function interpolation equations , and for constructing the final mesh .
the relations between conformal unstructured mesh generation , the ip problem and rational function interpolation raise many research questions .
these may lead to a deeper understanding of the properties of high quality meshes , and to better algorithms for creating them .
* acknowledgements . *
the author would like to thank mirela ben - chen , shlomi hillel , dov levine , yair shokef and vincenzo vitelli for helpful discussions and critical reading of the manuscript .
a cross is defined at every point @xmath49 of the domain , that is not a singularity of the function @xmath2 .
geodesic curves that start at @xmath49 can be drawn in all four directions of the cross . by the definition of a cross - field
, such a geodesic will be aligned with the cross - field everywhere along the curve . if @xmath49 is a singular point , with @xmath38-value @xmath281 , a cross
is not defined at @xmath49 , but there are @xmath282 geodesics that are incident on @xmath49 and follow the cross - field directions elsewhere on the curve .
these will be called _ star - geodesics_. for example , the geodesics drawn in fig .
[ multi_diamond_var_res],c . and [ fig : multi_ex4],d .
are star - geodesics .
denote the angle from the @xmath10-axis around @xmath49 by @xmath283 , and the cross direction when @xmath49 is approached from direction @xmath283 by @xmath284 , see fig .
[ fg : star_directions ] . to calculate the directions in which star - geodesics emanate from @xmath49 ,
we first calculate the cross when @xmath49 is approached from some direction , e.g. the positive @xmath10-axis .
this can be done using eq .
( [ eq : theta_change_par_tranport ] ) along a curve from the boundary to @xmath49 which approaches @xmath49 from the direction @xmath285 .
once @xmath286 is known , @xmath287 for any @xmath283 can be calculated by using eq .
( [ eq : theta_change_par_tranport ] ) along a small circular arc around @xmath49 at radius @xmath53 , @xmath97 .
the singularity term at @xmath2 is @xmath288 , and according to eq .
( [ eq : theta_change_par_tranport ] ) @xmath289 note that contributions to @xmath2 that are regular at @xmath49 do not affect @xmath287 for @xmath290 .
the star - geodesic directions are those for which the cross is directed along the ray from @xmath49 , or @xmath291 with @xmath46 . substituting eq .
( [ eq : app_a_th_calc ] ) into eq ( [ eq : app_a_star_cond ] ) and rearranging , we find @xmath292 it is easy to show that there are exactly @xmath293 different @xmath283-values for which this equation is fulfilled .
they are equally distributed around @xmath49 .
m. hmlinen , r. hari , r. j. ilmoniemi , j. knuutila and o. v. lounasmaa , magnetoencephalography - theory , instrumentation , and applications to noninvasive studies of the working human brain , _ reviews of modern physics _ * 65 * issue 2 ( 1993 ) 413 - 497 . | a novel approach to unstructured quadrilateral mesh generation for planar domains is presented .
away from irregular vertices , the resulting meshes have the properties of nearly conformal grids .
the technique is based on a theoretical relation between the present problem , and the inverse poisson ( ip ) problem with point sources .
an ip algorithm is described , which constructs a point - source distribution , whose sources correspond to the irregular vertices of the mesh . both the background theory and the ip algorithm address the global nature of the mesh generation problem .
the ip algorithm is incorporated in a complete mesh generation scheme , which also includes an algorithm for creating the final mesh .
example results are presented and discussed . |
after bachelier s seminal paper @xcite and its re - discovery @xcite , random walk theory has been the most crucial cornerstone in economics and finance .
an assumption that price dynamics is governed by stochastic process has become popular and useful in asset valuation theories such as option pricing theory @xcite or interest rate models @xcite .
however , the assumption also claims that prices of financial instruments can not be predicted exactly because of the nature of brownian motion .
this unpredictable nature of financial markets helps economists to establish a belief that there are no tools to find arbitrage opportunities and to make money systematically in the financial markets .
it is also imposed that successful investors are considered nothing but luckier than others .
the idea is crystallized in the form of the efficient market hypothesis by eugene fama @xcite and paul samuelson @xcite . according to the efficient market hypothesis , financial markets are informationally efficient and
this efficiency can not make participants systematically achieve excessive returns over the market portfolio in the long run .
although there are three slightly different versions of the hypothesis to cover more general cases , what the hypothesis generally emphasizes has not been changed .
however , many market practitioners intrinsically have an idea that the market could be predictable regardless of their methods used for forecast and investment because it is partially or totally inefficient .
the idea is opposite to the belief of proponents for the efficient market hypothesis and it is empirically supported by the fact that there are actual market anomalies which are used as the sources of systematic arbitrage trading .
these anomalies and trading strategies include fundamental analysis , technical analysis , pair trading , price momentum , sector momentum , mutual fund arbitrage , volatility arbitrage , merger arbitrage , january effect , and weekend effect etc .
the anomalies let market participants create profits by utilizing the trading strategies based on the market inefficiencies .
even if the market is efficient in the long run , practitioners assure that they are able to find opportunities and timings that the market stays in the inefficient phase within very short time intervals .
the existence of a shortly inefficient market state is guaranteed by the success of high frequency trading based on quantitative analysis and algorithmic execution in a short time scale automated by computers . in these cases
, the arbitrage does not have the traditional definition that non - negative profit is gained almost surely .
it can create positive expected return with high probability but there are also downside risks which make the portfolio underperform .
this kind of arbitrage is called statistical arbitrage and the arbitrage in this paper means mostly statistical arbitrage .
not only the practitioners but some academic researchers also have different opinions to the efficient market hypothesis . they have taken two approaches to check the validity of the efficient market hypothesis . on the one hand , the market anomalies of which the practitioners believe the existence are empirically scrutinized .
some results on the famous market anomalies are reported in academic literatures and seem to be statistically significant while their origins are not clearly revealed yet . for more detailed discussions on the market anomalies , see singal @xcite or lo and mackinlay @xcite . on the other hand ,
psychological and behavioral aspects of investors begin to be paid attention in order to find the explanatory theories on the market anomalies @xcite .
the behavioral economists focus on cognitive biases such as over- and under - reaction to news / events and bounded rationality of investors @xcite .
they claim that those biases can create the inefficiencies in the market .
the cognitive biases lead the investors to group thinking and herding behavior that most of investors think and behave in the similar ways .
the good examples of herding are speculative bubbles , their collapses , market crashes , and panics during financial crises .
momentum effect on price or valuation in assets is one of the famous examples which have attracted the interest of industry and academia . as an implemented trading strategy , momentum strategy
is frequently used for statistical arbitrage .
additionally , an investor gains a maximum 1.31% of monthly return by the monthly momentum strategy that constructs the portfolio which buys past winners and short - sells losers in the u.s .
market @xcite .
since price dynamics has a tendency that price moves along the direction it has moved , price momentum becomes the systematic proxy for forecasting future prices .
if an investor buys past winners , short - sells past losers , and repeats execution of the strategy , then he / she is expected to gain positive return with high probability in the long run .
it is exactly a counterexample to the efficient market hypothesis . despite the success of the momentum strategy ,
the origin of price momentum is not well - understood and remains rather unclear .
some possible explanations on the existence of the momentum effect answer parts of the question and the behavioral direction is one of them for understanding the nature of price momentum .
physicists also have become interested in the characteristics of financial markets as complex systems .
mainly , econophysics and statistical mechanics communities have used their methodologies to analyze the financial markets and several research fields have attracted their interests . in the sense of correlation ,
the financial markets are interesting objects .
since there are many types of interactions between market building blocks such as markets - markets , instruments - instruments , and investors - investors , correlations and correlation lengths are important .
in other directions , speculation and its collapse are always hot topics because they are explained as collective behavior in physics .
the analysis on speculation gives some partial answers that speculations have patterns including the resilience effect .
additionally , market crash or collapse of a bubble can be understood by the log - periodic pattern . for more details , see @xcite and references therein .
in particular , sornette introduced the concept of spontaneous symmetry breaking ( ssb ) of stock price to explain speculation and to resolve the growth stock paradox @xcite .
he pointed out that economic speculation is understood as price dynamics caused by desirable / undesirable price symmetry .
if stocks of a certain company are desirable to hold , investors try to buy the equities at extremely high prices which are the spontaneous symmetry breaking mode .
however , when the equities are not desirable any more , the investors do not want to hold it and try to sell them as soon as possible to avoid damages from the downslide of price caused by the situation that nobody in the market prefers the equities . in his paper , the phase transition is induced by riskless interest rate above risk - adjusted dividend growth rate which also expresses herding in the sense that large growth rate gets more attention from investors and it leads to herding .
positive dividend payment breaking the symmetry makes the price positive and this is why the positive price is observed .
these are the origins of speculation in economic valuation .
the result is also related to the well - known financial valuation theory called the gordon - shaprio formula .
his work is important in speculation modeling not only because symmetry breaking concept is applied to finance but also because speculation , its collapse , and market crash are indispensable parts of the market dynamics . in this paper ,
the concept of spontaneous symmetry breaking is applied to arbitrage modeling . unlike sornette s work
@xcite which uses spontaneous symmetry breaking to explain speculation in the asset valuation theory , the phase transition is emergent directly from arbitrage dynamics .
wyarta and bouchaud also consider symmetry breaking @xcite but their concern is self - referential behavior explained by spontaneous symmetry breaking of correlation in macroeconomic markets such as indexes not of arbitrage return generated by the trading strategy . from the viewpoint of symmetry breaking
, this paper pays attention to portfolio / risk management rather than explanations on macroeconomic regime change on which both of the previous works focus .
based on the dynamics which gives a spontaneous arbitrage phase and a no - arbitrage phase , the arbitrage strategy can be executed upon the phases of arbitrage .
the phases are decided by a control parameter which has the same meaning to speed of adjustment in finance .
the execution of the strategy aided by spontaneous symmetry breaking provides better performance than the naive strategy and also diminishes risk of the strategy . in section [ ssb ] , a brief introduction to arbitrage modeling is given and then the spontaneous arbitrage modes are emergent from the return dynamics .
the momentum strategy aided by spontaneous symmetry breaking is simulated on real data and the results in various markets are posted in section [ result ] . in section [ conclu ] ,
we conclude the paper with some discussions and future directions .
introducing the existence of arbitrage opportunity , the value of portfolio @xmath0 is governed by the following differential equation , @xmath1 where @xmath2 is risk free rate , @xmath3 is excessive return of the portfolio @xmath0 , @xmath4 is volatility of portfolio return , and @xmath5 is a stochastic noise term .
if the no - arbitrage theorem is imposed , the excessive return becomes zero guaranteed by the girsanov theorem that the risk - neutral measure @xmath6 and the brownian motion @xmath7 always exist under no - arbitrage situation @xcite .
if the existence of arbitrage is assumed , there is no risk - neutral measure @xmath6 nor related brownian motion @xmath7 . in this case
, it is more important to know how its return series has evolved .
the reason why the dynamics is important has two facets .
first of all , for theorists , the dynamics encodes large amount of information on market macro- and microstructure .
secondly , it is helpful for practitioners to exploit the arbitrage opportunity by implementing trading strategies based on the dynamics . the excessive return @xmath3 is modeled by @xmath8 where @xmath9 is a white noise term .
the structure of @xmath10 is decided by properties of arbitrage .
one of the simplest forms for @xmath11 is a polynomial function of @xmath12 .
two properties of arbitrage dynamics help to guess the structure of the function @xcite .
when the excessive return of the strategy is large enough , the arbitrage opportunity usually disappears very quickly because many market participants are easily able to perceive the existence of the arbitrage and can use the opportunity profitably even with trading costs .
this property imposes a constraint that coefficients of @xmath11 have negative values . additionally , eq . ( [ arb_gen_dyn ] ) should be invariant under parity transformation @xmath13 because negative arbitrage return is also governed by the same dynamics .
this property makes even order terms in the function vanish . considering these properties of arbitrage
, the form of @xmath11 is given by @xmath14 where @xmath15 for odd positive integer @xmath16 .
in traditional finance , these @xmath17s are also able to be considered as the proxies incorporating the information on changes of discount rates which are covered in @xcite .
the dynamics describes reversal of return that the return becomes decreased when being large and it is increased when under the trend line . in other words ,
the reversal makes the return stay near the equilibrium around the trend line . by dimensional analysis , @xmath18 is a speed of adjustment and is broadly studied in finance @xcite . larger @xmath19 means the arbitrage opportunity dies out much faster
. meanwhile , smaller @xmath19 corresponds to the situation that chances for arbitrage can survive longer .
as @xmath19 goes to infinity , the arbitrage return goes to zero extremely quickly and this limit corresponds to the no - arbitrage theorem .
when only the linear term is considered for the simplest case , the dynamics is an ornstein - uhlenbeck process in mathematical finance , @xmath20 where the trend line @xmath21 is zero .
this stochastic differential equation is invariant under parity transformation of @xmath22 because @xmath23 is an ito process with standard normal distribution which has symmetric distribution around mean zero .
although there are higher order terms in eq .
( [ arb_dyn_poly ] ) , the dynamics is still considered as a generalized ornstein - uhlenbeck process because it is the mean - reverting process around the trend line . we begin to introduce a cubic term to the ornstein - uhlenbeck process to extend it to more general cases .
the introduction of higher order terms is already used in the market crash model @xcite .
then the dynamics is changed to @xmath24 where @xmath25 , @xmath26 , and @xmath9 is a white noise term .
after the cubic term is introduced , adjustment on arbitrate return occurs quicker because the coefficients are all negative .
the negative coefficient condition needs to be modified in order to describe not only reversal but also trend - following arbitrage return which is explained by positive coefficients . in real situations ,
the trend - following arbitrage strategies are also possible to make profits by exploiting market anomalies because arbitrage opportunities fortified by transaction feedback do not disappear as quickly as expect and there could be more chances for investors .
speculation , as one of the examples , can create more opportunities for the trend - following arbitrage and increases expected return . under speculation ,
the investors buy the instrument even though the price is high .
this transaction induces to generate the trend line and is able to give feedback to the investors trading patterns . during market crash or bubble collapse , they want to sell everything at very low prices although the intrinsic values of instruments are much higher than the price at which they want to sell . not in extreme cases but under the normal market condition , people tend to buy financial instruments which have shown better performance than others because they expect that the instruments will provide higher returns in the future .
the prices of the instruments become higher because the investors actually buy with the expectation @xcite .
it seems to be very irrational but happens frequently in the markets . to integrate these kinds of situations
, we can introduce the cutoff value which can decide whether the arbitrage is originated from reversal or trend - following dynamics rather than the negative speed of adjustment . with the cutoff value ,
let us change @xmath18 and @xmath27 into the forms of @xmath28 where @xmath17 , @xmath29 , and @xmath30 are positive .
although the number of parameters seems to be increased , this is not true because @xmath29 is an external parameter . under these changes ,
the arbitrage dynamics is given by @xmath31 after relaxation time @xmath32 , eq .
( [ arb_dyn_ssb ] ) becomes zero up to the noise term because other transient effects die out .
in other words , the deterministic part of arbitrage dynamics arrives at the equilibrium state . by setting the deterministic part of the r.h.s . in eq .
( [ arb_dyn_ssb ] ) to zero , stationary solutions are found .
the interesting point is that the number of stationary solutions is dependent with @xmath17 and @xmath29 . in the spontaneous symmetry breaking argument
, @xmath17 is a control parameter and @xmath12 is an order parameter . when @xmath33 , there is only one asymptotic solution @xmath34 which shows the property of usual arbitrage opportunities .
the meaning of this solution is that the arbitrage return finally becomes zero up to noise .
it is obvious that the arbitrage opportunity vanishes after the relaxation time because it is taken by market participants who know the existence and use the chance . for @xmath35 ,
there are three asymptotic solutions with @xmath34 and @xmath36 the solution @xmath37 has the same meaning to the solution for @xmath33 .
it means that the arbitrage opportunity finally dies out .
the latter solutions , @xmath38 , are more interesting because there exist long - living arbitrage modes in return . after the relaxation time ,
the arbitrage chance still exists and lifetime of the spontaneous market anomaly is longer than that of the usual short - living arbitrage . it is noteworthy that these solutions unlike @xmath37 are symmetry breaking solutions although the dynamics is conserved under parity .
the spontaneous mode also has the coherent meaning in the sense of speed of adjustment @xmath17 .
if @xmath17 is smaller than the critical value @xmath29 , it is slower adjustment and the arbitrage opportunity can have longer lifetime .
these solutions are also well - matched to the no - arbitrage theorem that the arbitrage chance does not exist because it disappears very quickly .
the no - arbitrage theorem which corresponds to @xmath39 does not make the arbitrage possible after the relaxation time because @xmath17 is always greater than @xmath29 .
when a weak field term is introduced to eq .
( [ arb_dyn_ssb ] ) , the observation becomes more interesting . introducing the constant term @xmath40 ,
the equation is given by @xmath41 where @xmath40 can be considered the velocity of @xmath12 . if @xmath35 , the asymptotic solution is also changed from @xmath42 to @xmath43 as positive @xmath40 is changed to negative .
the asymptotic behaviors described in the previous subsection can be cross - checked with exact solutions . in the long run ,
the noise term is ignored because its average is zero . under this property , the exact solutions of eq .
( [ arb_dyn_ssb ] ) are given by @xmath44 where @xmath45 is the initial time .
when @xmath33 , exponential functions in the nominator and the denominator go to zero in the large @xmath46 region and it makes @xmath3 zero .
this corresponds to the symmetry preserving solution which is the usual arbitrage . if @xmath35 , the exponential functions become dominant as @xmath46 goes to infinity . at that time , @xmath3 approaches @xmath47 which are the symmetry breaking solutions .
these solutions are already seen in the asymptotic solutions . with the long - living arbitrage solutions in eq .
( @xmath48 ) , properties of the solutions are checked graphically in fig .
[ ssb_return_lambda ] and [ ssb_return_time ] . in fig .
[ ssb_return_lambda ] , the left graph shows time evolution of the solutions as @xmath49 . in the small t region , there exist non - zero arbitrage returns regardless of the value of @xmath50 .
however , as @xmath49 , the return approaches to non - zero if @xmath51 and it vanishes if @xmath52 . in the asymptotic region ,
the difference becomes clear and phase transition happens where @xmath17 is at the critical value @xmath29 .
it is easily seen in the graph on the right .
the region @xmath51 is called the long - living arbitrage phase , spontaneous return phase , or arbitrage phase .
another region where @xmath53 is considered the short - living arbitrage phase or no - arbitrage phase . in the model , market anomalies survive if they are in the long - living modes . in fig .
[ ssb_return_time ] , the spontaneous arbitrage returns approach to @xmath42 whatever initial return values are .
however , the no - arbitrage phase finally goes to zero .
this property does not depend on the size of the initial return values .
even if the initial value is smaller than the asymptotic value , it grows up to the asymptotic value .
for example , if investors realize the arbitrage opportunity and if they begin to invest into the chance , their trading behavior affects price dynamics and the trend - following investors pay attention to the instruments .
the interest leads to trading which gives feedback to their trading patterns and can increase the profitability .
in other words , money flows into the instrument , boosts its price , and gives feedback to investors behaviors .
if transaction cost is smaller than the asymptotic value , arbitrage opportunities created by spontaneous symmetry breaking can be utilized by the investors .
when the long - living arbitrage mode is possible , @xmath3 can be re - parametrized by @xmath54 where @xmath55 is a dynamic field for expansion around @xmath47 .
plugging this re - parametrization into ( [ arb_dyn_ssb ] ) , the differential equation for @xmath56 is solved and its solutions are given by @xmath57 since the latter solution goes to @xmath58 in the asymptotic region , we can check the transition between @xmath42 and @xmath43 .
if @xmath59 , the initial modes stay in themselves , i.e. @xmath47 go to @xmath47 . however ,
if @xmath56 is the latter solution , they evolve to @xmath60 in large @xmath46 limit even though we start at @xmath47 initially .
in order to test the validity of spontaneous symmetry breaking of arbitrage , we apply the following scheme depicted in fig .
[ flow_chart ] to trading strategies over real historical data . in backtest , the control parameter @xmath17 for the strategy should be forecasted based on historical data . at certain specific time @xmath45
, it is assumed that data only from @xmath61 are available and the control parameter for next period is forecasted from them . if the forecasted @xmath17 is smaller than the forecasted @xmath29 , the strategy which we want to test is expected to be in spontaneous arbitrage mode in the next period and the strategy will be executed .
when the forecast tells that the strategy would not be in spontaneous arbitrage mode , it will not be exploited and the investor waits until the next execution signals .
the weak field is also able to decide the method of portfolio construction .
if the constant term is positive , the portfolio which the strategy suggests to build will be constructed .
however , if the constant term becomes negative , weights of portfolio will become opposite to those of the portfolio originated from the positive constant term .
simply speaking , the portfolio is not constructed if the speed of adjustment is larger than the critical speed .
when it is smaller than the critical value , the weight of the portfolio is @xmath62 if the weak field is positive and the portfolio is @xmath63 if the weak field is negative .
this kind of multi - state models is popular in the names of hidden markov model @xcite or threshold autoregressive model @xcite in econometrics and finance .
the scheme is repeated in every period over the whole data set .
to apply the model to real data , the model considered in the continuous time frame needs to be modified to discrete version because all financial data are in the forms of discrete time series . in the discrete form , eq .
( [ arb_dyn_ssb ] ) is changed into @xmath64 and an additional @xmath65 related to the coefficient @xmath66 in the first term on the r.h.s comes from the time derivative in eq .
( [ arb_dyn_ssb ] ) .
the next step is estimation of parameters in eq .
( [ arb_dyn_ssb_disc ] ) with real data .
regression theory gives help on estimation but it is not easy to estimate the parameters with real data because the model is nonlinear and many methods in the regression theory are for linear models . in statistics , these parameters can be estimated by nonlinear regression theory but it is not discussed in this paper . instead of using nonlinear regression theory directly
, we can get some hints from linear regression theory .
with consideration on financial meanings and physical dimensions of the parameters , linear regression theory enables us to estimate the model parameters .
there are some issues on the estimation of parameters .
the first issue is related to stability of the parameters .
when the parameters are fit to real data , if values of the parameters severely fluctuate over time , those abruptly - varying parameters hardly give a steady forecast .
one of the best ways to avoid this is taking a moving average ( ma ) over a certain period .
moving average over the period can make the parameters smoothly - changing parameters .
for longer ma windows , the parameter is stable but it would be rather out of date to tell more on the recent situation of the market .
if it is short , they can encode more recent information but they tend to vary too abruptly to forecast the future values . to check ma window size dependency , a range of mas needs to be tested and the results from different mas should be compared . another issue is the method to estimate parameters in the model . since two or three internal parameters and one external parameter are given in the model , the same number of equations should be prepared . for the simpler case ,
the coefficient for each term can be considered as one parameter . in this case , two equations need to be set up .
however , the values of two parameters found from two equations sometimes diverge when real data are plugged . since @xmath17 and @xmath29 are the speeds of adjustment and have same physical and financial meanings
, they need to be derived from the same origin .
the only difference is that @xmath29 is external .
in addition to that , the symmetry breaking needs comparison between two different speeds , @xmath17 and @xmath29 .
one of the possible solutions is that @xmath17 is derived from the return series of the strategy and @xmath29 comes from the benchmark return as the definition of an external parameter .
this interpretation can give two parameters the same physical dimensions and financial meanings .
the specification on @xmath17s is also reasonable in the sense of the efficient market hypothesis . since the hypothesis tells that it is impossible to systematically outperform the benchmark , it is obvious that we compare the performance of the strategy with that of the benchmark in order to test the hypothesis . in the case of @xmath67 ,
the volatility of the strategy or benchmark return can be a good candidate because @xmath67 also has the same meaning and dimension to variance .
for the constant term @xmath40 , the average value of strategy return or benchmark return would be considered .
dividend payment rate is also a good candidate .
however , since the most important parameters in the model are @xmath17 and @xmath29 , we focus on the estimation of these two parameters . the intuitive way to get @xmath17 and @xmath29 is using a hint from the autoregressive model of order 1 called the ar(1 ) model . ignoring the cubic term
is also justified by the fact that the returns are much smaller than 1 . starting with the simpler model which does not have the cubic term , multiplying @xmath65 to both sides and taking ma over @xmath68 periods make the last term zero on the r.h.s . of eq .
( [ arb_dyn_ssb_disc ] ) and give the form of @xmath17 .
the one - step ahead forecasted @xmath17 is @xmath69 where @xmath70 . in longer ma windows ,
we can change @xmath71 in the denominator to @xmath72 because @xmath71 is close to @xmath72 . in shorter ma windows ,
the change is meaningful because it is capable of considering more recent informations will be tested in next subsections . ] . based on this argument ,
the final form of forecasted speed of adjustment is given by @xmath73 this @xmath17 has the same form to the parameter in ar(1 ) model which is found in @xmath74}{e[x_{i}x_{i } ] } \nonumber \end{aligned}\ ] ] where @xmath75 $ ] is the expectation value .
the estimator ( [ lambda_estimator ] ) is intuitively estimated but the hand - weaving argument is available .
since the benchmark return tends to be weakly autocorrelated and the return series by the arbitrage strategy is expected to be strongly positive - autocorrelated , the estimator for the arbitrage strategy is usually smaller than that for the benchmark . in this case , the strategy is in the long - living arbitrage mode . when the estimator for the strategy is larger than that for the benchmark , it is highly probable that return series for the strategy becomes much more weakly autocorrelated than the benchmark return
this tells that the strategy has recently suffered from large downslide and it can be used as the stop signal to strategy execution .
additionally , since the estimator is related to the correlation function which is in the range of -1 and 1 , the value of the estimator fluctuates between 0 and 2 and it is well - matched to the positiveness condition on @xmath17 .
the momentum strategy is one of the famous trading strategies which use market anomalies .
it is well - known that the strategy that buys past winners , short - sells past losers in returns , and then holds the portfolio for some periods in the u.s .
market enables to provide positive expected returns in intermediate time frames such as 112 months @xcite .
the basic assumption of the strategy is that since price has momentum in its dynamics , it tends to move along the direction it has moved .
based on the assumption , the financial instruments which have shown good performance in the past are highly probable to gain profits in the future .
opposite to winners , it is likely that losers in the past would underperform the benchmark in the monthly time frame . over other trading strategies such as pair trading or merger arbitrage strategies ,
it is advantageous that the momentum strategy is able to be exploited at any time and in any markets .
pair trading is utilized only when the correlation of two instruments weakens and when investors can find it .
merger arbitrage is able to make benefits if m&a rumors or news begin to be spread in the market and if there is a price gap between actual and buy prices .
when using these strategies , the investors become relatively passive to market conditions and events . however , in the case of the momentum strategy , if they look back at the price history , market participants make use of momentum strategy and the trading frequency is up to their time frames from high frequency trading to long - term investment .
in addition to that , unlike merger arbitrage which is possible only in equity markets , momentum strategy can be applied to various asset markets including local equity , global index / equity @xcite , currency @xcite , commodity @xcite , future @xcite , and bond markets @xcite . to exploit momentum strategy ,
first of all , returns of all equities in the market universe during a certain period called the look - back or ranking period are calculated from closing prices on the first and the last trading dates of the lookback period .
then equities are sorted by their returns in ascending order .
grouped into ten groups , the first group contains the worst performers and the tenth group is for the best performers .
each equity in winner or loser baskets has the same weight in the basket and each basket is also equal in absolute wealth but opposite in position to make the whole portfolio zero cost .
the portfolio constructed in zero cost is held during the holding period .
the construction of the portfolio occurs on the first day of the holding period and it is liquidated on the last day of the holding period .
the transactions happen at the closing prices of each day .
the strategy with @xmath76 period look - back and @xmath77 period holding period is shortly called @xmath78 strategy .
the expected return by momentum strategy is dependent with lengths of ranking and holding periods . as explained , the strategy with intermediate lengths such as monthly lookback and holding periods generates positive expected return . for longer period strategies ,
the momentum strategy suffers from reversal such that the winner group loses its price momentum and shows poor performance . meanwhile , the loser basket shows the opposite behavior such that the basket outperforms and can provide positive return @xcite . according to lo and mackinlay @xcite , the short - term momentum strategy in weekly scale
also has negative expected return . in both longer and shorter term strategies ,
it is impossible to make a profit on the momentum portfolio but it does not mean that there is no statistical arbitrage chance nor that the market is efficient . if the portfolio is constructed by reversal momentum i.e. contrarian strategy which buys losers and sells winners , there still exist the chances of profits .
the position by contrarian strategy is exactly opposite to that of the momentum strategy .
the reason why the momentum strategy generates positive expected return has attracted the interest of researchers but it is not clearly revealed yet .
the sector momentum is considered one of possible explanations @xcite . a behavioral approach to momentum
also can give more explanations such as under - reaction @xcite or over - reaction @xcite of market participants to news or psychology @xcite .
it is ambiguous whether the momentum effect comes from either which of them or from a combination of these possible explanations .
however , this paper focuses on how to use the strategy based on symmetry breaking rather than what makes markets inefficient .
two different market universes are used for analysis to avoid sample selection bias .
the first universe is the s&p 500 index that is the value / float - weighted average of the top 500 companies in market capitalization in nyse and nasdaq .
it is one of the main indexes such as the dow jones index and russell 3000 in the u.s .
standard & poor s owns and maintains the index with regular rebalancing and component change .
another universe is kospi 200 in the south korean market operated by korea exchange ( krx ) .
it is the value - weighted average of 200 companies which represent the main industrial sectors . unlike the s&p 500 index , kospi 200
contains small - sized companies in market capitalization and considers sector diversification .
its components and weights are maintained regularly and are also irregularly replaced and rebalanced in the case of bankruptcy or upon sector representation issues such as change of core business field or increase / descrease of relative weight in the sector .
the significance of each index in the market is much higher than those of other main indexes such as dow jones 30 index or russell 3000 in the u.s . and the kospi index , the value - weighted average of all enlisted equities in the south korean market , because futures and options on the indexes have enough liquidities to make strong interactions between equity and derivative markets . in the case of the korean market ,
the kospi 200 index among main indexes is the only index which has the futures and options related to the main indexes . additionally , many mutual funds target the indexes as their benchmarks and various index - related exchange - traded funds are highly popular in both markets .
the whole time spans considered for two markets are slightly different but have large overlaps .
s&p 500 is covered in the term of jan .
1st , 1995 and jun .
30th , 2010 , 15.5 years which includes various macro - scale market events such as the russian / asian crisis ( 19971998 ) , dot - com bubble ( 19952000 ) , its collapse ( 20002002 ) , bull market phase ( 2003 - 2007 ) , sub - prime mortgage crisis ( 20072009 ) , and the recovery boosted by quantitative easing i ( 20092010 ) . in the case of kospi 200 ,
the market in the period of jan .
2000 to dec . 31st .
2010 , had experienced not only economic coupling to the u.s .
market but also local economic turmoils such as the credit bubble and crunch ( 20022003 ) . given the market and time span ,
the price history of each stock and whether it was bankrupt or not are stored on a database in order to remove survivor bias and all records for component change are also tracked to keep the number of index components the same .
the s&p 500 data are downloaded from bloomberg .
the whole data of kospi 200 components and their change records are able to be downloaded from krx .
the total number of equities in the database during the covered period is 968 and 411 for s&p 500 , kospi 200 , respectively . in both markets , @xmath79
weekly strategies are considered and the contrarian portfolios are constructed .
the reason for choosing @xmath79 weekly strategies is that they show the best performance in each market among 144 strategies derived from maximum 12-week lookbacks and holdings .
excessive weekly returns of the portfolios are calculated from risk - free rates and proxies for the risk - free rate are from the u.s .
treasury bill with 91 days duration for s&p 500 , cd with 91 days duration for kospi 200 . since the weekly momentum portfolio is constructed at the closing price of the first day in the week and is liquidated at the closing price of the last day , the benchmark return is also calculated from the closing prices of the first and the last days in the week .
the results for these markets are given in fig .
[ grp_return ] there are similarities and differences in two markets .
first of all , it is easily seen that @xmath79 strategy shows a reversal that if the winner basket is bought , it is impossible to get a significant positive return but we can achieve positive return from the loser basket . in particular , the contrarian portfolio beats the benchmark over whole periods and this comes from the fact that the loser basket outperforms and the winner basket underperforms the benchmark . additionally , the contrarian strategy looks more profitable in the korean market and it can be explained that developed markets have weaker anomalies than emerging markets because the investors in the developed market have utilized the anomalies during longer periods . in the south korea market ,
the winner and the loser have more clear directions and magnitudes of the returns are much greater than those of the u.s .
it is easily seen in table [ tbl_stat ] .
.statistics for contrarian strategy and benchmark in s&p 500 and kospi 200 . [ cols="^,^,^,^,^,^,^,^",options="header " , ] [ tbl_stat ] in table [ tbl_stat ] , the numbers from the kospi 200 confirm much stronger and clearer contrarian patterns as shown in fig .
[ grp_return ] .
the contrarian return in the korean market is weekly 1.325% which is much greater than 0.225% from s&p 500 contrarian strategy and the t - value of the kospi 200 contrarian strategy is 9.491 which is 0.1% statistically significant but the u.s .
strategy has only 2.065 , 5% statistically significant .
the null hypothesis is that the expected excessive return is zero .
similar to the contrarian returns , the winner basket and the loser basket have larger absolute returns and t - statistics in kospi 200 .
both of them are 0.1% statistically significant but the s&p 500 loser return only has a 5% statistically significant t - value and a less significant t - value for another . in both markets ,
benchmarks have much smaller weekly expected returns than those by the contrarian strategies and t - values are not significant .
standard deviation gives another reason why the portfolio by the momentum / contrarian strategy needs to be constructed .
after construction of the contrarian portfolio , the volatility of the portfolio is smaller than the volatility of the winner group and the loser group . in particular , in the south korea market , the contrarian portfolio has a smaller volatility than the benchmark and has a greater sharpe ratio than each of the winner and the loser basket has .
a larger sharpe ratio imposes that the strategy is good at minimizing the risk and maximizing the return .
winners , losers , and contrarian portfolios have large kurtosis by fat - tailed distribution .
the results by symmetry breaking with 99 different ma windows are given in fig .
[ sp500_ssb ] and [ kospi200_ssb ] .
the strategies aided by spontaneous symmetry breaking show better performance than the naive momentum strategy in both markets and the results are not particularly dependent on the market where the strategy is used . in the case of return
, the strategies with shorter ma windows have improved returns than longer ma windows or naive momentum strategy .
as the length of the ma window becomes longer , the return plunges sharply and this plummet is observed in both markets .
the sharpe ratio is also increased with the ssb - aided strategy and it is obvious that the modified strategy is under better risk management .
the winning percentage also increases and it is larger for shorter ma windows .
the application of spontaneous symmetry breaking also has the minor market dependencies . in s&p 500 ,
average returns and sharpe ratios increase after a drop around the 20-ma window but the kospi 200 momentum does not recover its average return level and remains stagnated around returns by the naive strategies . in the case of volatility , it is helpful to reduce volatility with ssb in kospi 200 but is not useful in s&p 500 . the constant term in spontaneous symmetry breaking is also considered . as described before , average return over the ma window or
return in previous term of the raw strategy are used as the forecasted constant term .
if the constant is positive , the contrarian portfolio is constructed and if the constant is negative , the momentum strategy is used . however , the strategy including the constant term does not provide better results than the strategy without the constant .
the same approach is applied to mean return or return in previous terms of the benchmark but it is not possible to find the better strategy . with these facts , it is guessed that the constant term is zero or the constant term is always positive if it exists and if these returns are the only possible candidates for the constant .
the positiveness of the constant can be guaranteed by the fact that the arbitrage portfolio is constructed to get a positive expected return . with other estimators for speed of adjustment
, it is found that the ssb - guided strategies provide similar results although the results are not given in the paper . in both markets ,
the patterns of results are similar to the results depicted in fig .
[ sp500_ssb ] and [ kospi200_ssb ] . specifically speaking
, the estimator with @xmath71 in the denominator gives similar patterns in kospi 200 but the performance is slightly poorer than the result in fig .
[ kospi200_ssb ] . in the u.s .
market , similar patterns in longer ma windows are found but the results with shorter ma windows are worse than the result given in the paper . this is well - matched to the assumption that @xmath71 is almost identical to @xmath72 in longer ma windows .
when the estimator uses the covariance of @xmath65 and @xmath80 , similar results are found but the performance becomes much poorer , especially in shorter window length . although the whole time period is same for each of the ma windows , longer ma strategies have fewer data points when the performance is calculated in backtest .
this difference in number of data points comes from the assumption that even though we work with historical data already known , we pretend to be unaware of the future after the moment at which the forecast is made in backtest . in the simulation with each ma ,
the first few data whose length is the same as the size of the ma window are used for forecast and are ignored in the calculation of performance .
however , the difference does not make any serious difference in the patterns of performance .
when the tests to calculate the performance are repeated over the same sample period for all ma windows , notable differences are not observed and the results are similar to fig . [ sp500_ssb ] and [ kospi200_ssb ] .
the cubic order term and parity symmetry on return introduce the concept of spontaneous symmetry breaking to arbitrage dynamics . in the asymptotic time region , the dynamics has symmetry breaking modes triggered by the control parameter
. it can provide the long - living arbitrage modes including the short - living mode in the dynamics .
spontaneous symmetry breaking generated by the control parameter @xmath17 imposes phase transition between the arbitrage phase and no - arbitrage phase .
contrasting to the short - living mode which is expected in the frame of the efficient market hypothesis , the long - living modes are totally new and exotic .
the existence of a spontaneous arbitrage mode explains why the arbitrage return survives longer than expected and why the trading strategies based on market anomalies can make long term profits . with the existence of the weak field , it is possible to consider the transition between two long - living arbitrage modes , @xmath47 in the asymptotic region . based on spontaneous symmetry breaking of arbitrage , the control parameter enables to decide execution of the trading strategy .
if @xmath17 for the strategy is smaller than @xmath29 for the benchmark , the strategy will be executed in next period .
if the speed of adjustment for the strategy is greater than that of the benchmark , nothing will be invested .
since it is difficult to estimate the parameter in the nonlinear model , the ar(1 ) model gives an insight for estimation .
the estimated @xmath17 based on the ar(1 ) model has the theoretical ground that the speed of estimation is derived from the autocorrelation function .
it is also reasonable in the sense of testing the efficient market hypothesis because it is capable of comparing the strategy with the benchmark .
the simplest but most meaningful estimator for the control parameter is applied to momentum strategy in the u.s and south korean stock markets .
the ssb - aided momentum strategy outperforms and has lower risk than the naive momentum strategy has .
since the strategy applied to two different markets shows similar patterns , the results are not achieved by data snooping .
it is also not by estimator bias because three different estimators for speed of adjustment are tested and provide similar results with some minor differences .
the future study will be stretched into a few directions .
first of all , parameter estimation needs to be more precise and statistically meaningful . in this paper ,
the estimator for the control parameter @xmath17 is from the ar(1 ) model and @xmath17 for benchmark serves as the critical value @xmath29 although the cubic term exists .
although it provides better performance and lower risk , estimation of the parameters is from the reasonable intuition not from regression theory . for the more precise model ,
they need to be estimated from nonlinear regression theory .
in particular , a statistical test on estimation should be done . in the case of @xmath29
, it can be estimated with the help of other researches on market phase such as wyarta and bouchaud s work @xcite .
other parameters , @xmath30 or @xmath40 , also help to find the better performance strategy if they are statistically well - estimated .
the second direction is considering the stochastic term in arbitrage dynamics . in the paper , only the deterministic part
is considered and the stochastic term is out of interest in finding the exact solutions . if the spontaneous symmetry breaking modes are found not as the asymptotic solutions but as the exact solutions of the stochastic differential equation , they would extend our understanding on arbitrage dynamics .
in addition to that , specification of relaxation time can be found from the correlation function of the stochastic solutions .
finally , it would be interesting if validity of the arbitrage modeling with spontaneous symmetry breaking is tested over other arbitrage strategies . since only the momentum / contrarian strategy is the main concern in the paper , tests on other trading strategies including high frequency trading look very interesting .
additionally , a cross - check with momentum strategies for different markets and frequencies would be helpful to check the effectiveness and usefulness of spontaneous symmetry breaking concepts in arbitrage modeling .
it is our pleasure to thank matthew atwood , jonghyoun eun , robert j. frey , wonseok kang , andrew mullhaupt , and svetlozar rachev for useful discussions .
we are especially indebted to sungsoo choi for helpful discussions from the early stage of this work .
we are grateful to didier sornette for providing valuable advice in the revision of the first draft .
we express thanks to sungsoo choi , jonghyoun eun , and wonseok kang for cooperation on the construction of the financial database on the korean stock market .
we are thankful to xiaoping zhou for collecting parts of price histories on s&p 500 components . 99 bachelier , l. , theorie de la spculation , annales scientifiques de lcole normale suprieure 3 ( 1900 ) 21 - 86 cootner , p. , the random character of stock market prices , mit press , 1964 black , f. and scholes m. the pricing of options and corporate liabilities , journal of political economy 8 ( 1973 ) 637 - 654 merton , r. c. , theory of rational option pricing , bell journal of economics and management science ( the rand corporation ) 4 ( 1973 ) 141 - 183 .
vasicek , o. , an equilibrium characterization of the term structure , journal of financial economics 5 ( 1977 ) 177 - 188 cox , j.c . , ingersoll , j.e . and ross , s. a. , a theory of the term structure of interest rates , econometrica 53 ( 1985 ) 385 - 407 hull , j. and white , a. , pricing interest - rate derivative securities , the review of financial studies 3 ( 1990 ) 573 - 592 fama , e. , the behavior of stock market prices , journal of business 38 ( 1965 ) 34 - 105 samuelson , p. , proof that properly anticipated prices fluctuate randomly , industrial management review 6 ( 1965 ) 41 - 49 singal , v. , beyond the random walk : a guide to stock market anomalies and low risk investing , oxford university press , 2003 lo , a. and mackinlay , a. , a non - random walk down wall street , princeton university press , 2001 shleifer , a. , inefficient markets : an introduction to behavioral finance , oxford university press , 2000 kahneman , d. and tversky , a. , choices , values and frames , cambridge university press , 2000 kahneman , d. , slovic , p. , and tversky , a. judgment under uncertainty : heuristics and biases , cambridge university press , 1982 conlisk , j. , bounded rationality and market fluctuation , journal of economic behavior and organization 29 ( 1996 ) 233 - 250 jegadeesh , n. and titman , s. , returns to buying winners and selling losers : implications for stock market efficiency , journal of finance 48 ( 1993 ) 65 - 91 mangtegna , r. n. and stanley , h. e. , introduction to econophysics : correlations and complexity in finance , cambridge university press , 2007 roehner , b. m. , patterns of speculation : a study in observational econophysics , cambridge university press , 2005 sornette , d. , why stock markets crash : critical events in complex financial systems , princeton university press , 2004 sornette , d. and malevergne , y. , from rational bubbles to crashes , physica a 299 ( 2001 ) 40 - 59 ( 2001 ) malevergne , y. and sornette , d. , multi - dimensional rational bubbles and fat tails , quantitative finance 1 ( 2001 ) 533 - 541 sornette , d. , stock market speculation : spontaneous symmetry breaking of economic valuation , physica a 284 ( 2000 ) 355 - 375 wyarta , m. and bouchaud , j. p. , self - referential behaviour , overreaction and conventions in financial markets , journal of economic behavior and organization 63 ( 2007 ) 1 - 24 girsanov , i. v. , on transforming a certain class of stochastic processes by absolutely continuous substitution of measures , theory prob .
. 5 ( 1960 ) 285 - 301 ilinski , k. , physics of finance : gauge modelling in non - equilibrium pricing , wiley , 2001 cochrane , john h. , presidential address : discount rates , journal of finance 66 ( 2011 ) 1047 - 1108 amihud , y. and mendelson , h. , trading mechanisms and stock returns : an empirical investigation , journal of finance 42 ( 1987 ) 533 - 553 damodaran , a. , a simple measure of price adjustment coefficients , journal of finance 48 ( 1993 ) 387 - 400 theobald , m. , and yallup , p. , measuring cash - futures temporal effects in the uk using partial adjustment factors , journal of banking and finance 22 ( 1998 ) 221 - 243 theobald , m. , and yallup , p. , determining security speed of adjustment coefficient , journal of financial markets 7 ( 2004 ) 75 - 96 bouchaud , j. p. and cont r. , a langevin approach to stock market fluctuations and crashes , european physical journal b 6 ( 1998 ) 543 - 550 hong , h. and stein , j. c. , a unified theory of underreaction , momentum trading and overreaction in asset markets , journal of finance 54 ( 1999 ) 2143 - 2184 terence , h. , hong , h. , lim , t. , and stein , j. , bad news travels slowly : size , analyst coverage , and the profitability of momentum strategies , journal of finance 55 ( 2000 ) pp .
265 - 295 baum , l. e. and petrie , t. statistical inference for probabilistic functions of finite state markov chains .
the annals of mathematical statistics 37 ( 1966 ) : 1554 - 1563 tong , h. , threshold models in nonlinear time series analysis , springer - verlag , new york , 1983 rouwenhorst , k. g. , international momentum strategies , journal of finance 53 ( 1998 ) 267 - 284 rouwenhorst , k. g. , local return factors and turnover in emerging stock markets , journal of finance 54 ( 1999 ) 1439 - 1464 okunev , john , and derek white , do momentum - based strategies still work in foreign currency markets ? ,
journal of financial and quantitative analysis 38 ( 2003 ) 425 - 447 erb , claude b. , and campbell r. harvey , the strategic and tactical value of commodity futures , financial analysts journal 62 ( 2006 ) 69 - 97 asness , c. s. , moskowitz , t. , and pedersen , l. h. , value and momentum everywhere , university of chicago working paper , 2008 moskowitz , t. j. , ooi , y. h. , and pedersen , l. h. , time series momentum , university of chicago working paper , 2010 de bondt , w. f. m. and thaler , r. , does the stock market overreact ?
, journal of finance 40 ( 1985 ) 793 - 805 lo , a. and mackinlay , a. , when are contrarian profits due to stock market overreaction ? , review of financial studies 3 ( 1990 ) 175 - 205 moskowitz , t. j. and grinblatt m.
, do industries explain momentum ?
, journal of finance 54 ( 1999 ) 1249 - 1290 daniel , k. , hirshleifer , d. , subrahmanyam , a. , investor psychology and security market under- and over - reactions
. journal of finance 53 ( 1998 ) 1839 - 1886 barberis , n. , shleifer , a. , vishny , r. a model of investor sentiment .
journal of financial economics 49 ( 1998 ) 307 - 343 | we introduce the concept of spontaneous symmetry breaking to arbitrage modeling . in the model , the arbitrage strategy is considered as being in the symmetry breaking phase and the phase transition between arbitrage mode and no - arbitrage mode is triggered by a control parameter .
we estimate the control parameter for momentum strategy with real historical data . the momentum strategy aided by symmetry breaking shows stronger performance and has a better risk measure than the naive momentum strategy in u.s . and south korean markets . spontaneous symmetry breaking , arbitrage modeling , momentum strategy |
ngc 5128 , the nearest giant elliptical galaxy @xcite , was associated with the strong radio source centaurus a by @xcite . in the optical
, the galaxy has an elliptical light distribution crossed in the center by a strong warped dust lane .
the diffuse optical extensions were first discovered by @xcite and subsequently the rich system of shells around the elliptical galaxy was detected by @xcite .
@xcite and @xcite discovered the optical jet or filaments in ngc 5128 . the origin and the nature of the emission - line filaments have been hotly debated @xcite .
the origin of diffuse shells @xcite can be explained with phase wrapping of stars from an accreted disk galaxy @xcite .
@xcite discussed the stellar populations in the prominent diffuse north - eastern shell and in the halo of ngc 5128 .
the ` outer filament ' situated at @xmath5 from the centre extends for @xmath6 in the direction of position angle @xmath7 @xcite overlaping partially with the diffuse stellar shell . while the stellar populations in the shell are similar to the rest of the halo of the giant elliptical , the stars close to the emission - line filament are predominantly young ( see also * ? ? ?
* ; * ? ? ?
the ` inner filament ' is located at the distance of @xmath8 from the nucleus and extends outward over @xmath9 at a position angle @xmath10 @xcite .
it consists of three bright knots , identified as a , b and c by @xcite and two more diffuse structures , e and f @xcite . here
we report the detection of a young stellar population associated with the optical emission - line inner ( @xmath11 ) and outer ( @xmath12 ) filaments in the north - eastern halo of ngc 5128 . throughout the paper
we adopt distance modulus to ngc 5128 of 27.8 @xcite that corresponds to a distance @xmath13 mpc . at this distance
, @xmath14 on the sky translates to 1.047 kpc in the galaxy .
we used fors1 at vlt antu to image the outer field because it is less crowded and the filaments cover a larger area than in the inner field . on the contrary in the inner field
magellan i was necessary because despite of having a smaller area it gives higher resolution for the crowded field .
the field with the inner filaments ( @xmath15 , @xmath16 ) was imaged with the magellan i ( walter baade ) telescope , one of the twin 6.5 m magellan telescopes at cerro manqui at the las campanas observatory .
the observations were taken on 22 february 2001 , during the first chilean run on the new telescope and one of the first 10 nights of scientific operations .
the instrument was the direct ccd camera tektronix @xmath17 with gain of 2 @xmath18 . at the f/11 focus of magellan
i the pixel scale is @xmath19/pixel with a total field size of @xmath20 .
the small pixel scale and the extraordinary good and stable seeing of @xmath21 during the whole night enabled us to obtain well sampled point spread function ( psf ) images with almost no crowding problems deep in the halo of ngc 5128 .
the observations consist of @xmath22 , @xmath23 and @xmath24 exposures in harris @xmath25- , @xmath0- and @xmath26-bands .
] , respectively . the color composite image of the field can be found at the following url @xmath27 .
the standard stars from the @xcite catalogue in two different fields were observed at four different airmasses on 22 and 23 february 2001 with the identical instrumental setup .
additional photometric standards in these fields were defined by @xcite in the @xmath0 and @xmath26 bands , making the total number of useful standard stars observed 15 , 50 and 48 in @xmath25 , @xmath0 and @xmath26 , respectively .
the following calibration equations reproduce the calibrated magnitudes with one - sigma scatter around the mean of 0.08 , 0.024 and 0.031 for @xmath25 , @xmath0 and the @xmath26 band , respectively .
@xmath28 @xmath29 @xmath30 the observations of the field that contains outer filaments ( @xmath31 , @xmath32 ) were carried out in the service mode with fors1 instrument on the 8.2 m eso very large telescope ( vlt ) antu at paranal , on july 11@xmath33 and 12@xmath33 1999 .
they consist of pairs of 15-min optical ( bessel @xmath25- and @xmath0-band ) exposures taken in photometric conditions with seeing ranging from @xmath34 to @xmath35 .
the fors1 detector has pixel scale of @xmath36/pixel and the field of view @xmath37 .
a detailed description of these data and their reduction is given in @xcite .
the photometric measurements of the outer filament images were described in ( * ? ? ?
* field 1 ) .
the photometry in the inner filament field followed a similar procedure .
the stars were identified on the median combined image with the find algorithm that is part of the daophot ii programme @xcite .
the psf magnitudes were measured on all individual frames simultaneously with allframe @xcite .
restricting the psf fitting parameters sharpness and chi we rejected bad pixels and extended objects like galaxies and globular clusters @xcite from the final photometry list . only the stars with good photometry ( @xmath38 ) in at least two images
were kept .
the final photometric catalogue of of the inner filament field contains 1612 , 8003 and 7683 stars with good photometry in the @xmath25 , @xmath0 and @xmath26 bands , respectively .
in fig . [ uvcmds ] we compare the uv color - magnitude diagrams of the inner and outer filament fields . the much larger field of vlt than magellan i and the more extended star formation region in the outer field accounts for higher number of blue stars with colors @xmath39 . in both fields
there is a young blue main sequence extending to magnitudes of @xmath40 .
we `` cleaned '' the cmds from foreground contamination using the besanon group model of stellar populations in the galaxy .
some of the stars in the red part ( @xmath41 ) of the cmd are most probably still the remaining galactic foreground stars , but in that area of the cmd supergiants on their blue red excursions are also located ( see also @xcite ) .
the width of the blue to red supergiant loops in the red part constrains the abundance , while the brightness of the main sequence stars constrains the age of the population . in order to determine an upper limit for the temperature and ionizing fluxes of the stars we assumed the age and abundance from the isochrone fitting .
the best fit from the available set of padova isochrones @xcite is obtained for z=0.004 and the log(age@xmath42 yr . the brightest stars in both fields are lying along the @xmath4 yr old isochrone . assuming then that the brightest stars have z=0.004 and log(age)=7.0 - 7.1 yr , the temperature of the most luminous stars is between @xmath43 and @xmath44 k. the comparison with another halo field in the southern region of the galaxy ( field 2 in @xcite ) indicates that the young population is confined to the regions aligned with the optical filaments .
the ( u - v ) vs. ( v - i ) color - color diagram for all the stars in the inner filament field with photometric errors @xmath45 mag and magnitudes brighter than 50% completeness limits in all bands is presented in fig .
[ ccduvi ] .
the foreground stars have not been subtracted .
they form the sequence that crosses the diagram from upper right towards lower left corner .
the objects in the lower left part of the color - color diagram are young supergiants in ngc 5128 .
the spread is due to possible differential reddening and spread in ages of the young stars .
most galaxies have @xmath46 and @xmath47 @xcite , so only few of them at higher redshift may contribute to the spread in the bluest part of the color - color diagram .
the isolated point to the left of the stellar sequence at @xmath48 and @xmath49 is most probably a galaxy .
absence of other points with similar colors indicate a good selection criteria and low contamination by background galaxies in our stellar photometry .
the reddening value inferred from the isochrone fitting to the stellar sequence is @xmath50 in good agreement with the values from the literature @xcite .
the width of the young blue sequence is somewhat larger in the inner than in the outer filament field .
this may indicate the presence of the small differential reddening amounting at most to @xmath51 .
however , a possible spread in ages of young stars and the photometric errors at fainter magnitudes complicate the interpretation .
the photometry of knot a @xcite , a possible compact star forming region in the jet , required particular care .
it consists of a mixture of stellar like objects , extended nebular emission and an unresolved background of faint stars . in order to estimate the relative contribution of these components to our photometry
, we deconvolved a small portion of the magellan i image of @xmath52 arcsec around the brightest hot spot using the mcs deconvolution algorithm @xcite . in this way we obtained a new image with improved spatial resolution ( fig .
[ knota ] , top right ) and a background map ( or the `` extended source channel '' ) containing the extended sources .
the final pixel scale on the deconvolved image is @xmath53 and the resolution of point sources is @xmath19 . in the mcs algorithm
, one has to give as an input , initial estimated positions and intensities for all suspected point sources .
the program then iteratively changes them while trying to deconvolve the data .
the difference between the data and the deconvolved image ( reconvolved back with the psf ) in units of the noise , is used to check the quality of the process , a posteriori . this difference image or residual map
should be flat and equal to 1 everywhere in the field in case of perfect deconvolution . since introducing a higher number of point sources in the image will always produce a better result
, we deliberately choose the minimum number of point sources required for the deconvolution process to yield acceptable residual maps .
some faint stars in the deconvolved image are therefore present in the `` extended source channel '' of the data and we `` see '' fewer stars with the deconvolution process than with allframe . however , the stars we see are decontaminated from the extended underlying nebular emission and we can infer their magnitudes and colors with no contribution from the emission of the surrounding gas .
the comparison between these new measurements and the photometry obtained with allframe ( fig .
[ knota ] ) indicates that the magnitudes and colors of blue stars measured with allframe are not excessively contaminated by line emission from the surrounding gas .
in particular we show that knot a is composed of at least two point sources embedded in a more extended envelope , possible nebular emission or a very compact and still unresolved cluster .
other objects in the field do not show such extended emission apart from the compact object @xmath54 arcsec sw of knot a. the residual flux of knot a after star subtraction is @xmath55 erg/@xmath56/sec / .
many high - redshift radio galaxies show optical images elongated in the same direction as their double radio sources , the so called alignment effect @xcite .
the alignment has often been explained with models of star formation induced by shocks associated with the passage of radio components @xcite , but in some cases scattering of nuclear light by dust or electrons and line - emission excited by the continuum from the obscured active nucleus are dominating optical and uv light @xcite . in ngc 5128
we can observe in more detail physical processes that are acting in these more distant galaxies .
the direct observations of young blue supergiants associated with the outer filaments were recently reported by several authors @xcite .
now for the first time we have resolved the young stars associated with the ionized filaments and knots in the inner part of the galaxy , and mapped these young stars over a large field along the radio jet .
the remarkable alignment of the young stars and ionized gas over @xmath57 kpc is shown in fig .
[ blue_red_inner ] and [ blue_red_outer ] .
the red stars ( @xmath58 ) are evenly distributed in both fields as expected for foreground galactic stars , while almost all recently formed stars lie along the knots and filaments of ionized gas .
the north - south ( vertical ) alignment in the outer field is along the eastern edge of the h i cloud that is associated with the stellar shells @xcite .
we checked the locations of all of the blue stars and in particular of the most luminous ones . in the outer field
they are evenly distributed along the star forming region , while the four brightest blue stars in the inner field lie within a very small region , known as knot a @xcite .
@xcite found that the spectrum of the knot a shows characteristics of a normal h ii region and conclude that there must be hot stars embedded in that gas .
we resolved here the brightest stars in that h ii region .
the northern extension of knot a is distinguished by a higher excitation and electron temperature similar to what is observed in other knots and filaments in the inner and outer halo @xcite .
the high excitation lines observed to be prominent in the filaments in the halo of ngc 5128 ( e.g. he ii @xmath594686 and high ratio of [ oiii ] @xmath595007/[oii ] @xmath593727@xmath60 ) show that the stars can not be the only source of ionizing photons .
the complex velocity field of the filaments creates a puzzle .
the spectroscopy of the ionized gas revealed a very wide range of velocities covering more than 550 km s@xmath61 @xcite and varying significantly on scales of @xmath62 pc .
the mean velocity of the gas , around @xmath63 km s@xmath61 , is a few hundred km s@xmath61 lower than the systemic velocity of the central region @xcite , but is close to the measured velocity of the h i cloud that is located to the nw of the outer filaments @xcite .
if the stars shared the turbulent velocities of the gas , in @xmath4 yr they would travel @xmath64 kpc , which correspond to approx .
@xmath65 at the distance of ngc 5128 ( note that the field of view of magellan is less than @xmath66 and that of vlt @xmath67 ) .
if they had these velocities , the gas and the stars would have been dispersed in the halo of the galaxy .
the large amount of the h i and co gas @xcite and its association with the diffuse stellar shells in the halo can be interpreted with the dynamical scenario of phase - wrapping , following the merger of a spiral galaxy with ngc 5128 approximately @xmath68 yr ago @xcite . the similar mean velocity of the h i cloud and the ionized gas in the filaments indicates the common origin of the gas .
we detect here for the first time the young stars associated with the inner filaments .
their strong alignment with the line emitting filaments and with the radio and x - ray jet coming from the agn nucleus of ngc 5128 suggests a connection between them and poses interesting questions like `` did the jet trigger the star formation ? '' or `` are the jets seen because of the star formation ? '' and `` would we see the jet if there were no blue stars there ? '' .
the answer to the first question is more complex and thus we address the others first .
if we assume that the brightest stars have ages of @xmath69 myr and metallicities of z=0.004 , then they have temperatures of @xmath70 k. the high excitation lines observed in the spectra of the ionized gas require ionization temperatures of the order of @xmath71 k. it is clear that an additional ionization mechanism is exciting the gas . the alignment of the ionized gas with the direction towards the nucleus of the galaxy and the fact that even though it is present , the gas in the other parts of the halo of ngc 5128 is not ionized , leaves as the most simple possibility either an interaction with the radio jet that leads to the shock excitation @xcite or photoionization by the radiation field of a nuclear continuum source @xcite . in any case , the `` jet '' of optical emission line features is not seen due to the star formation and it would most probably be visible even if the stars were not there . the young stars formed after the collapse of the gas probably have the same origin as the h i and molecular clouds @xcite .
the presence of this cool gas , with @xmath72 k , is the necessary ingredient in the jet induced star formation models . in the model described by @xcite
the prime effect of the radio outburst in a galaxy would be to heat the surrounding medium and evacuate the radio lobe of thermal gas .
however , it would not disturb the cool - phase clouds , which would then be triggered to collapse in the overpressured medium .
so , the jet could have induced the collapse and the star formation without transferring to the stars the high turbulent velocities that are observed today in the gas .
the star formation process driven by the external pressure may not be dissimilar from the one observed in the classical stellar associations , such as orion .
this means that one needs to find an alternative mechanism in order to explain the velocities of the gas that have been observed .
one possibility is that supernova explosions induced shocks into the surrounding medium .
however , the problem with this scenario is that the first generation of supernovae would explode some @xmath73 yr after the onset of the star formation .
so , the gas would acquire the high velocities still @xmath74 million years ago and would have had plenty of time to disperse .
it would have too low density and would probably not be visible any more .
thus either the first supernovae were not efficient enough to blow away all the gas or some new gas came from elsewhere . also , the jet that triggered the star formation would have induced high velocities in the hot gas , which would then have plenty of time to disperse in the halo of the galaxy .
the obvious source of the new gas is the close - by h i cloud that will continue to supply fresh material that can be ionized as long as it stays in the direction of the jet .
the inner and the outer filaments are not mutually aligned .
similarly the inner and the outer radio jets are misaligned .
if we assume that the jet is precessing as suggested also by @xcite , it should have first hit the outer filament and precess in the eastward direction hitting later the inner filament field .
the recent observations of the radio jet with vlbi @xcite and the x - ray jet with chandra @xcite show close alignment with the inner filaments at the position angle of @xmath10 .
stars in the inner and the outer filament have the same ages to within our precision , of the order of @xmath75 yr .
the distance between the inner and the outer filaments is @xmath76 , implying the projected precession rate of the the order of @xmath77 degrees per year .
the presence of the large quantity of neutral gas towards the west of the outer filament implies that either the jet is precessing in a cone whose right edge does not pass the outer filaments on the west or that it is intermittent and started shining only @xmath78 yr ago .
given the facts that the blue stars are aligned along the direction of the jet , that they have formed recently , that there is a lot of gas all over the halo of the galaxy @xcite , but new stars are formed only where the jet passes across or close to that gas , we can conclude that the young stars observed in the filaments in the ne halo of ngc 5128 are result of the jet induced star formation . the optical emission from the gaseous filaments associated with those stars
is mainly powered by the radiation field of the hidden nuclear continuum source @xcite or by the shock excitation due to the interaction of the radio jet with a dense cloud of material at the location of the filaments @xcite .
if this cause - effect scenario is correct , we can put a lower limit to the age of the jet itself : it must have been acting for at least @xmath4 yr in order to form the stars observed .
we have used the direct camera at the magellan i and fors1 at antu vlt to measure the stellar photometry in two fields in the halo of the giant elliptical galaxy ngc 5128 .
the fields were located at the position of the inner and outer optical filaments @xcite .
we detected the chain of blue stars aligned with the line - emitting filaments and with the radio jet .
the main conclusions from this work are : * we see young blue stars aligned with the jet over @xmath57 kpc . *
the young stars have ages of @xmath4 yr and their uv color - magnitude diagram is best fit with padova isochrones @xcite for abundance of z=0.004 . *
the jet triggered star formation in the halo of ngc 5128 , presumably when hitting dense clouds of neutral gas left in the halo by the previous accretion event @xcite .
* the main excitation mechanism that ionizes the gas associated with the young stars is either beamed photoionization from the obscured agn @xcite or shock ionization by the radio jet @xcite .
the young stars with temperatures of the order of @xmath70 k contribute only a smaller part of ionizing photons .
* our results support the alignment effect observed in high redshift radio galaxies , and suggest that radio galaxies should have higher than normal star formation rates .
* there is evidence for the jet to be processing from the outer to inner filament field eastward in the sky .
* we infer a lower limit for the age of the jet as @xmath4 yr .
we are grateful to r. fosbury for stimulating discussion and valuable comments that improved this paper and to p. stetson for the careful reading of the manuscript .
we thank the staff of las campanas observatory for their help during the observations as well as the eso astronomers who took the vlt observations for us in service mode .
mr acknowledges eso studentship programme .
this work was performed in part under the auspices of the chilean fondecyt no .
01990440 .
bertelli , g. , bressan , a. , chiosi , c. , fagotto , f. & nasi , e. , 1994 , a&as , 106 , 275 best , p.n . , longair , m.s . & rttgering , 1996 , h.j.a . ,
mnras , 280 , l9 blanco , v.m . ,
graham , j.a . , lasker , b.m .
& osmer , p.s . , 1975 ,
apj , 198 , l63 bolton , j.g . ,
stanley , g.j .
& slee , o.b . , 1949 ,
nature , 164 , 101 chambers , k.c . ,
miley , g.k . &
van breugel , w. , 1987 , nature , 329 , 604 charmandaris , v. , combes , f. & van der hulst , j.m . , 2000 ,
a&a , 356 , l1 daly , r.a . , 1990 , apj , 355 , 416 de breuck , c. , van breugel , w. , minniti , d. et al . , 1999 ,
a&a , 352 , l51 di serego alighieri , s. , fosbury , r.a.e . ,
tadhunter , c.n . & quinn , p.j . , 1989 , nature , 341 , 307 dickson , r. , tadhunter , c. , shaw , m. , clark , n. , & morganti , r. , 1995 , mnras , 273 , l29 dufour , r.j .
& van den bergh , s. , 1978 , apj , 226 , l73 fassett , c.i . & graham , j.a . , 2000 , apj , 538 , 594 graham , j.a . , 1998 , apj , 502 , 245 graham , j.a . , 1983
, apj , 269 , 440 graham , j.a . , 1979 , apj , 232 , 60 graham , j.a . & price , r.m .
, 1981 , apj , 247 , 813 harris , g.l.h . , harris , w.e . & poole , g.b . , 1999 , aj , 117 , 855 haynes , r.f . ,
cannon , r.d . & ekers , r.d .
, 1983 , pasau , 5 , 241 israel , f.p .
, 1998 , a&ar 8 , 237 johnson , h.m . , 1963 , aj , 68 , 76 kraft , r.p . ,
forman , w. , jones c. et al . , 2000 , apj , 531 , l9 landolt , a.u .
, 1992 , aj 104 , 340 magain , p. , courbin , f. & sohy , s. , 1998 apj 494 , 472 malin , d.f .
, quinn , p.j .
& graham , j.a . , 1983 ,
apj 272 , l5 mccarthy , p.j . , van breugel , w. , spinrad , h. & djorgovski , s. , 1987 , apj , 321 , l29 morganti , r. , killeen , n.e.b . , ekers , r.d . & oosterloo , t.a . , 1999 , mnras , 307 , 750 morganti , r. , fosbury , r.a.e . ,
hook , r.n . ,
robinson , a. & tsvetanov , z. , 1992 , mnras , 256 , 1p morganti , r. , robinson , a. , fosbury , r.a.e .
, di serego alighieri , s. , tadhunter , c.n . & malin , d.f .
, 1991 , mnras , 249 , 91 mould , j.r . ,
ridgewell , a. , gallagher , j.s .
iii et al . ,
2000 , apj , 536 , 266 peterson , b.a , dickens , r.j .
& cannon , r.d .
, 1975 , proc .
aust . , 2 , p. 366
quinn , p.j .
, 1984 , apj , 279 , 596 rees , m.j . , 1989 , mnras , 239 , 1p rejkuba , m. , 2001 , a&a , 369 , 812 rejkuba , m. , minniti , d. , silva , d.r . & bedding ,
t. , 2001 , a&a , submitted robin , a. & creze , m. , 1986 , a&a , 157 , 71 robin , a.c . ,
haywood , m. , creze , m. , ojha , d.k .
& bienayme , o. , 1996 , a&a , 305 , 125 rodighiero , g. , granato , g.l . ,
franceschini , a. , fasano , g. & silva , l. , 2000 , a&a , 364 , 517 schlegel , d.j . ,
finkbeiner , d.p . &
davis , m. , 1998 , apj , 500 , 525 schiminovich , d. , van gorkom , j.h . , van der hulst , j.m . & kasow , s. , 1994 , apj 423 , l101 soria , r. , mould , j.r . , watson , a.m. , et al . , 1996 ,
apj 465 , 79 sutherland , r.s . , bicknell , g.v . &
dopita m.a . , 1993 ,
apj , 414 , 510 stetson , p.b . , 2000 , pasp , 112 , 925 stetson , p.b . , 1994 , pasp , 106 , 250 stetson , p.b . , 1987 , pasp , 99 , 191 tingay , s.j . , jauncey , d.l . , reynolds , j.e .
et al . , 1998 ,
aj , 115 , 960 vernet , j. , fosbury , r.a.e .
, villar - martn , m. et al . , 2001 , a&a , 366 , 7 | we used the direct ccd camera at the magellan i telescope at las campanas observatory and fors1 at antu vlt at eso paranal observatory to image fields centered on the inner and outer optical filaments in the halo of ngc 5128 . in the @xmath0 vs. @xmath1 color - magnitude diagrams
we have identified young blue supergiants associated with these line - emitting filaments located between the inner radio lobe and the northern middle lobe . around the outer filament stars as young as 10 myr were detected .
they are principally aligned with the direction of the radio jet , but a vertical north - east alignment along the edge of the h i cloud is also present .
young stars in the inner filament field are found inside the bright knots of photoionized gas and are strongly aligned in the direction towards the center of the galaxy at the same position angle as the inner radio jet . fitting the padova isochrones on uv color - magnitude diagrams
we find that the blue stars around the inner filaments have ages similar to the ones around the outer filaments @xmath2 myr and the same abundance of z=0.004 .
the presence of young blue supergiants clearly shows that the bright blue knots in the north - eastern halo of ngc 5128 are associations of young stars with photoionized gas .
the temperature of the brightest stars is t@xmath3 k , insufficient to account alone for the high excitation lines observed in the surrounding ionized gas .
thus the optical emission jet is principally seen due to its alignment with the radio structure of the agn .
the highly collimated star formation is present only in the north eastern halo of the galaxy , suggesting the interaction of the jet with the gas clouds deposited during the last accretion event as the preferred triggering mechanism . from these observations
we infer a lower limit for the age of the ngc 5128 jet of @xmath4 yr .
the triggering of the star formation in the dense clouds in the halo of the galaxy by the jet supports the alignment effect observed in high redshift radio galaxies .
it also suggests that radio galaxies should have higher than normal star formation rates . |
anomalous x - ray pulsars ( axps ) , about half a dozen of which are known , have properties significantly different from those of binary x - ray pulsars ( see mereghetti 1999 for a recent review ) .
axps are sources of pulsed x - ray emission with relatively low persistent x - ray luminosities , @xmath1 erg / sec , and soft spectra which are well - fitted by a combination of blackbody and power - law contributions , with effective temperatures and photon indices in the range @xmath2 kev and @xmath3 , respectively .
they have relatively long spin periods of about @xmath4 seconds , which increase steadily with time .
their characteristic ages are about @xmath5 years .
no binary companions have been detected for these objects , and observations have placed strong constraints on companion masses ( e.g. mereghetti , israel & stella 1998 ; wilson et al .
at least three axps have been associated with young supernova remnants which limit their ages to @xmath6 kyr .
there are at present two theories seeking to explain the properties of axps . in one ,
they are modeled as isolated , ultramagnetized neutron stars or `` magnetars '' spinning down by the emission of magnetic dipole radiation , which is powered either by residual thermal energy ( heyl & hernquist 1997a , b ) or by magnetic field decay ( thompson & duncan 1996 ) . the required surface magnetic field strength is @xmath7 g , which is similar to the values inferred from timing data for soft gamma repeaters or sgrs ( e.g. kouveliotou et al . 1998 , 1999 ; see , however , marsden , rothschild & lingenfelter 1999 ) .
the other class of theories proposes that the x - ray emission from axps is powered by accretion from binary companions of very low mass ( mereghetti & stella 1995 ) , from the interstellar medium ( see wang 1997 in reference to the axp candidate rx j0720.4 - 3125 ; however , see heyl & hernquist 1998 for a magnetar model of this object ) , or from the debris of a disrupted high - mass x - ray binary system , after a stage of common - envelope evolution ( van paradijs et al .
1995 ; ghosh , angelini & white 1997 ) . the inferred magnetic field strengths of the neutron stars then are similar to those of ordinary radio pulsars and luminous x - ray pulsars ( @xmath8 g ) .
more recently , chatterjee , hernquist and narayan ( 1999 ; hereafter chn ) proposed another accretion model in which axps are neutron stars with standard magnetic fields accreting from a disk formed after fallback of material from a supernova explosion .
the star is rapidly spun down close to the observed axp periods on a timescale of @xmath9 years by the action of a propeller effect ( for other accretion scenarios for axps and sgrs based on the propeller effect , see alpar [ 1999 ] and marsden et al . [
1999 ] ) .
however , it has been observed that the spin - down of certain axps is very stable over time ( kaspi , chakrabarty & steinberger 1999 ) .
this could pose a problem for all accretion models , in which a higher level of timing noise might be expected .
moreover , it is not clear whether optical and infra - red emission from such accretion disks would be low enough to evade the limits set by observations ( perna , hernquist & narayan , 1999 ; hulleman et al .
it remains to be seen whether accretion models can circumvent these difficulties .
a question that has hitherto largely been unaddressed is the reason for the puzzlingly strong clustering of spin periods of the observed axps .
colpi , geppert & page ( 2000 ) contend that this phenomenon implies magnetic field decay in the magnetar model .
marsden et al .
( 1999 ) seek to explain this by limiting the magnetic fields of the neutron stars and the strength of the propeller wind emission by which they spin down to narrow ranges of values . in this paper , we show that the observed narrow period range of axps is consistent with the accretion model described in chn if the neutron stars are drawn from an underlying population characterized by broad and reasonable distributions of magnetic field strength , initial spin and accretion disk mass .
it has been suggested that axps and sgrs are drawn from the same underlying population of objects .
this is motivated by the observation that sgrs ( see hurley 1999 for a review ) have properties that are extremely similar to those of axps , except for the fact that they occasionally undergo energetic outbursts .
we show in this paper that the narrow period range of axps and sgrs taken together is also consistent with the above model .
the theoretical model used in this paper is that which is set out in detail in chn , and to which the reader is referred for details . here
we only summarize the different possible histories of a neutron star accreting from a debris disk .
such a star will have various spin - down histories depending on the initial values of its free parameters : its surface magnetic field strength @xmath10 , its initial spin period @xmath11 , and the mass of its accretion disk @xmath12 .
for some values of initial parameters , the star goes into an extended radio pulsar phase in which the magnetospheric radius is greater than the light cylinder radius ; accretion action is then assumed to stop .
for other initial values , the star is spun down in the propeller phase , during which accreted matter is flung out by centrifugal forces prior to reaching the surface of the star ; the star then reaches the quasi - equilibrium `` tracking phase '' , in which the spin of the star roughly matches the equilibrium period but never quite equals it since the mass accretion rate declines with time .
if the star enters the tracking phase at a time @xmath13 , then we identify the period between @xmath13 and the `` accretion death time '' @xmath14 to be the axp phase ; here , @xmath15 is the time at which the mass accretion rate falls to about 1% of the eddington value and the infall becomes an advection - dominated accretion flow or adaf . for yet other initial values of the star s parameters , it remains in the propeller phase until after the time @xmath14 ; in other words , @xmath16 . such a system was called a `` propeller system '' in chn .
the x - ray luminosity of the system is determined by the mass accretion rate on to the surface of the star , @xmath17 , which could be different from the mass accretion rate through the disk , @xmath18 , if some of the material is driven away from the system before reaching the surface of the neutron star .
thus , when accretion action ceases and the star is in the radio pulsar phase , the x - ray luminosity falls to zero . during the propeller phase , most of the matter will be ejected by centrifugal forces , and the system will be x - ray faint , since @xmath19 . during the tracking phase , we assume that @xmath20 if @xmath21 , where @xmath22 is the mass accretion rate corresponding to the eddington luminosity , but that @xmath23 if @xmath24 ; in this phase , the system will be x - ray bright . note that the assumption here is that even though @xmath18 could be highly super - eddington , especially at early times , at no point would the x - ray luminosity exceed the eddington limit , owing to ejection of matter by radiation pressure prior to reaching the surface of the star . it is not known how rapidly the luminosity will fall during the adaf phase of the axp ; presumably it will do so faster than linearly in @xmath18 ; in this paper , for purposes of numerical calculation , we assume the dependence of luminosity between between times @xmath15 and @xmath14 to be @xmath25 ( of course , for @xmath26 , @xmath27 ; and for @xmath28 , @xmath29 ) .
there are 6 axps known today : 1e 2259 + 586 , rxs j170849 - 4009 ,
4u 0142 + 615 , 1e 1048 - 5937 , 1e 1841 - 045 and axj 1845 - 0300 ( see mereghetti 1999 ) .
their spin periods , which are very tightly clustered together , are , in seconds , 6.98 , 11.00 , 8.69 , 6.45 , 11.76 and 6.97 , respectively .
associated young supernova remnants ( snrs ) of estimated age @xmath30 years have so far been detected for 3 of these 6 objects . if it is assumed , as in the model described in the previous section , that all the 6 observed axps are the result of supernova explosions @xmath30 years ago , and that the birthrate of radio pulsars in the galaxy is approximately 1 every 100 - 200 years ( see , e.g. , lyne et al .
1998 ) , then the relative birthrates of axps and radio pulsars turns out to be @xmath31 .
how is the distribution of spin periods of the axps to be calculated ?
the evolution in time of the properties of an axp depends on the three free parameters of the model , namely , the surface magnetic field strength of the neutron star @xmath32 , the mass of the initial circumstellar accretion disk @xmath12 , and the initial spin period of the star @xmath33 . the way we calculate the distribution of axp properties is as follows : we specify intrinsic prior probability distributions for each of the parameters @xmath34 and @xmath35 , and endow each neutron star with initial values of @xmath36 and @xmath35 sampled randomly and independently from their respective distributions , as well as a time of birth chosen randomly from a distribution uniform in time , since we expect the birthrate of neutron stars to be constant over time .
having done this for a large number of model neutron stars , we evolve each until a later fiducial time , when we take account of how many stars at that fiducial time are bright axps , dead axps ( i.e. , an axp which is past its @xmath14 time ) , propeller systems or radio pulsars
. the number of sample neutron stars should be sufficiently large , and the width of their time - of - birth distribution sufficiently long , that the results of this calculation are stable across repetitions of the process .
as an example , we present in detail the results of one such calculation . we take the surface magnetic fields of the neutron stars to be the distribution reported in narayan & ostriker ( 1990 ) for radio pulsars : namely , that log @xmath10 is distributed normally with a mean of 12.5 and a standard deviation of 0.4 ( note that in the rest of this paper , all indicated logarithms are to be taken to base 10 ) .
the distributions of @xmath35 and @xmath12 are poorly constrained by observations .
we choose distributions for them that are simple and cover fairly broad ranges of values .
@xmath35 is chosen to be distributed uniformly between 2ms and 50ms , and log@xmath37 is assumed to be distributed uniformly between -6 and -1.8 .
in one particular realization of this calculation , the initial number of neutron stars was selected to be 5,000,000 ; at the chosen fiducial time , the number of bright axps was 139,306 ; bar graphs showing their distribution of period , luminosity and age are shown in figure 1 .
the fractions of all initial neutron stars becoming axps , propeller systems and radio pulsars were 0.07 , 0.10 , and 0.83 , respectively .
thus , the relative birthrates of axps and radio pulsars was 0.09 .
it will be observed that the axp spin period distribution peaks at roughly 5 - 6 seconds , that the ages of the axps are of the order of a few times @xmath38 years , and that the x - ray luminosity distribution peaks at about @xmath39 ergs / s ; all of these properties correspond reasonably well with the properties of observed axps .
notice also that the spin period distribution has a long tail towards high periods .
however , given that the total number of observed axps ( whose periods are much more closely clustered ) is only 6 , they could still constitute a representative sample from the above distribution . to quantify this statement
, we carry out the nonparametric kolmogorov - smirnov ( k - s ) test for two discrete samples ( see , e.g. , press et al . 1988 ) to test whether the hypothetical and observed axp spin period distributions are compatible .
we define the test statistic to be the quantity @xmath40 , where @xmath41 and @xmath42 are the sizes of the observed and hypothetical spin period distributions , respectively ( here , @xmath43 and @xmath44 ) , and @xmath45 is the maximum absolute difference between the cumulative distribution functions of the above distributions . in the above realization of the calculation
, @xmath46 has the value 0.76 , and it is just significant at the level of significance 0.61 .
the cumulative distributions of the observed and hypothetical spin periods are shown in figure 2 .
we thus conclude that the observed axp period distribution is consistent with the hypothetical distribution predicted by our model .
the strongest factor determining the axp properties in our model is the @xmath10 distribution ; the spin period distribution in figure 1 can be reasonably well - fitted by a log - normal distribution , reflecting the log - normal distribution of @xmath10 .
the weakest factor is the initial spin period @xmath35 ; changing its distribution does not significantly alter the final axp properties .
the above results are for the case in which log@xmath47 is distributed uniformly between -6 and -1.8 .
the axps in this model come from the relatively high end of the @xmath12 distribution ( in the above realization , the peak value for the axps of log@xmath47 is -1.8 , and the minimum is at approximately -3.4 ) ; neutron stars with smaller disk masses end up as propeller systems or radio pulsars .
one would thus expect that lowering the lower limit of the log@xmath47 distribution would not change the axp properties described above , except that it would lower the ratio of birthrates of axps and radio pulsars , since a greater fraction of neutron stars would then have accretion disks too little massive to become axps .
this is indeed the case : for lower limits of -5 , -6 and -7 , the corresponding fractions of neutron stars which become axps are 0.09 , 0.07 and 0.06 ; the fractions of neutron stars becoming x - ray dim propeller systems are 0.14 , 0.10 and 0.08 , respectively ; the remaining neutron stars become radio pulsars , the fractions of which are 0.77 , 0.82 and 0.86 , respectively ; hence , the corresponding ratios of birthrates of axps and radio pulsars are 0.12 , 0.09 and 0.07 . on the other hand , changing the upper limit of the log@xmath47 distribution changes not only the axp - radio pulsar birthrate ratio , but also the axp spin distribution : raising it shifts the peak of the period distribution towards lower periods , and lowering it shifts the peak towards higher periods .
this affects substantially the level of significance of the k - s test : for upper limits of -1.7 , -1.8 and -1.9 , the values of @xmath46 are approximately 0.90 , 0.76 and 0.89 respectively , with corresponding levels of significance of 0.39 , 0.61 and 0.41 .
the relative birthrates of axps , propeller systems and radio pulsars also change : the respective fractions of neutron stars becoming axps are 0.09 , 0.07 and 0.06 ; the fractions of neutron stars which become propeller systems are 0.11 , 0.10 and 0.10 , repsectively ; the corresponding fractions of neutron stars which become radio pulsars are 0.81 , 0.83 and 0.84 ; hence , the ratios of birthrates of axps and radio pulsars for the above three cases are 0.11 , 0.09 and 0.07 , respectively .
it is interesting to apply our model to try to explain the narrow period distribution of the axps and sgrs taken together .
there are four known sgrs : sgr 1627 - 41 , sgr 0525 - 66 , sgr 1806 - 20 and sgr 1900 + 14 ; their spin periods are , in seconds , 6.41 , 8.0 , 7.48 and 5.16 respectively .
if we take the 10 axps and sgrs together and apply the k - s test to check whether this distribution is consistent with our usual model with log@xmath47 distributed uniformly between -6 and -1.8 , we find ( with @xmath41 = 10 , in this case ) that @xmath48 , and it is just significant at the level of significance 0.29 .
thus , the combined axp and sgr spin distribution is also consistent with our hypothetical distribution .
we have shown in this paper that according to our model for fallback accretion disks around solitary neutron stars , not all isolated neutron stars which are the product of supernova explosions become radio pulsars ; a substantial fraction of them ( @xmath49 , for the models considered here ) end up as x - ray bright axps , a possibly larger fraction ( @xmath50 ) as x - ray faint propeller systems , while the remainder end up as radio pulsars .
if the 6 observed axps were produced by supernova explosions @xmath30 years ago , then the ratio of birthrates of axps and radio pulsars is expected to be of the order of 10% . whether a neutron star becomes a radio pulsar , an axp or a propeller system depends on the values of its surface magnetic field @xmath10 , initial spin period @xmath35 , and initial circumstellar disk mass @xmath12 . using the observationally constrained distribution function of @xmath10 for radio pulsars , and assuming simple distribution functions of initial spin period and disk mass covering a broad range in @xmath35 and log@xmath51 respectively , it has been shown that our accretion model produces a population of axps with a spin period distribution peaking at @xmath52 seconds , x - ray luminosity @xmath53 ergs / s , age @xmath30 years , and with approximately the above axp - radio pulsar birthrate ratio .
systems with relatively high values of @xmath10 and @xmath12 end up as axps , while those with low values end up as radio pulsars ; systems with intermediate values of these parameters become propeller systems .
it is difficult to make definitive statements about possible relations between the properties of the underlying neutron star , such as its magnetic field strength or its initial accretion disk mass , and its observable properties such as spin period , age or luminosity , since the wide ranges of values the initial parameters can take give rise to a lot of scatter in the final axp properties that can be observed .
nevertheless , some rough general trends may be discerned .
for example , if a neutron star has a low disk mass ( `` low '' and `` high '' here are to be understood relatively in the context of the corresponding distributions considered in the previous section and in figure 1 ) , then it would need a high surface magnetic field strength in order to become a visible axp ; this would in turn tend to produce a probable observed spin period which lies slightly rightwards of the peak of the spin period distribution in figure 1 ; the probable observed age is low , and the observed luminosity could take a range of values . on the other hand , a star with a high disk mass could reach the axp phase with a range of surface magnetic field strengths , and the probable observed spin period , luminosity and age could also span a range of values . in this paper , we wished also to address ourselves to the narrow spin period range observed in axps .
one might suspect that this would in turn constrain the parameters of any model of axps to unphysically slender ranges of values .
however , we have shown in this paper that even under the assumption of quite reasonable and broad distributions of the parameters , our model leads to a spin period distribution that is statistically consistent with the narrowly clustered distribution observed in the case of the 6 known axps , and also in the case of the 6 axps and 4 sgrs taken together ( if indeed they can be grouped together as fundamentally the same kind of objects ) .
we note in conclusion that the above model would not work as a viable explanation for axps if the underlying neutron stars had magnetar - like field strengths ( @xmath54 g ) ; in this case , spin - down would be extremely rapid and would produce spin periods of hundreds of seconds , rather than close to 10 seconds .
thus , it is unnecessary to invoke neutron stars with non - standard magnetic fields in the context of this model for axps . | we consider the accretion model for anomalous x - ray pulsars proposed recently by chatterjee , hernquist and narayan , in which the emission is powered by accretion from a fossil disk formed by the fallback of material from a supernova explosion .
we demonstrate that this model is able to account for the spin period , luminosity and age distributions of the observed population of axps for reasonable and broad distributions of the free parameters of the model , namely , the surface magnetic field of the neutron star , the mass of its accretion disk and its initial spin period . in particular , this model is able statistically to account for the puzzlingly narrow observed spin distribution of the axps .
we show also that if the establishment of fallback accretion disks around isolated neutron stars is a universal phenomenon , then a fairly large minority ( @xmath0 ) of these objects become x - ray bright axps or x - ray faint systems spinning down by propeller action , while the rest become radio pulsars . |
all graphs are assumed to be simple , finite and undirected , and any graph - theoretic notations and definitions not explicitly defined can be found in @xcite or @xcite . a _ vertex - deleted subgraph _
@xmath15 is the unlabelled graph obtained by deleting , from the graph @xmath0 , a vertex @xmath16 and all edges incident to @xmath16 .
the _ deck _ of @xmath0 , denoted @xmath17 , is the multiset of vertex - deleted subgraphs of @xmath0 and each member of @xmath17 is referred to as a _
card_. our main focus in this paper will be on the analogously defined _ edge - cards _ of @xmath0 which are edge - deleted subgraphs @xmath18 of @xmath0 .
the collection of the edge - cards is called the _ edge - deck _ of @xmath0 , denoted by @xmath19 . the reconstruction conjecture proposed in 1942 by
kelly @xcite and ulam @xcite states that a graph with at least three vertices , is uniquely determined , up to isomorphism , from its collection of vertex - deleted subgraphs .
the most natural variation of the reconstruction conjecture ( for a recent survey see @xcite ) is an analogue for deletion of edges .
this is the edge - reconstruction conjecture [ harary , 1964 ] which states that all graphs on at least four edges are edge reconstructible @xcite .
a _ reconstruction _ ( _ edge - reconstruction _ ) of @xmath0 is a graph @xmath20 with @xmath21 ( @xmath22 ) .
a graph @xmath0 is _ reconstructible _ ( _ edge - reconstructible _ ) if every reconstruction of @xmath0 is isomorphic to @xmath0 .
this means that @xmath0 is reconstructible ( edge - reconstructible ) if it can be obtained uniquely , up to isomorphism , from its deck ( edge - deck ) . for a reconstructible graph @xmath0 , harary and
plantholt @xcite introduced the notion of the _ reconstruction number _ of @xmath0 , denoted by @xmath23 , which is defined as the least number of vertex - deleted subgraphs of @xmath0 required in order to identify g uniquely ; that is , @xmath24 is the size of the smallest subcollection of the deck of @xmath0 which is not contained in any other deck of another graph @xmath20 where @xmath25 .
it can be considered as a measure of the level of difficulty in reconstructing @xmath0 uniquely .
the simplest observation we can make is that @xmath26 .
reconstruction numbers are now known for various classes of graphs such as disconnected and regular graphs and trees @xcite .
a variation of reconstruction numbers is the _ class - reconstruction numbers_. let @xmath27 be a class of graphs and let @xmath28 .
then the class - reconstruction number @xmath29 is defined as the least number of vertex - deleted subgraphs required to determine @xmath0 from any other graph in @xmath27 .
non - trivial class reconstruction numbers which have been studied include maximal planar graphs , trees and unicyclic graphs @xcite .
reconstructing the graph from the deck seems to be more difficult than reconstructing the graph from its edge - deck since more of the graph is left in an edge - deleted subgraph than in a vertex - deleted subgraph .
however , it sometimes happens that more edge - deleted subgraphs are required for unique reconstruction than vertex - deleted subgraphs .
motivated by the falsity of the reconstruction conjecture for directed graphs , ramachandran @xcite weakened the reconstruction conjecture by considering the degree of the deleted vertex along with each vertex - deleted subgraph in a _ degree - associated card _
, referred to as _ da - card_. the _ degree _ of a vertex @xmath16 is the number of edges of @xmath0 incident to @xmath16 .
a vertex of degree 0 is called an _ isolated vertex _ and a vertex of degree 1 is called an _ endvertex_. the _ minimum degree _ of a graph @xmath0 , denoted by @xmath30 , is the smallest number of edges incident to any vertex @xmath16 in @xmath0 . a da - card denoted by @xmath31 consists of a card @xmath15 in the deck of @xmath0 and the degree @xmath32 in @xmath0 of the deleted vertex @xmath16 .
the _ da - deck _ is the multiset of da - cards .
ramachandran defined the _ degree - associated reconstruction number _ of a graph @xmath0 , denoted by @xmath33 , to be the minimum number of da - cards necessary to determine @xmath0 uniquely .
clearly @xmath33 is equivalent to the class reconstruction number of @xmath0 given that @xmath0 is in @xmath34 , the class of graphs on @xmath35 edges . the _ edge - reconstruction number _ ,
the _ class edge - reconstruction number _ and the _ degree - associated edge - reconstruction number _ are analogously defined .
degree - associated edge - card _ or _ da - ecard _ is a pair @xmath36 consisting of an edge - card @xmath37 in the edge deck of the graph @xmath0 and the degree of the edge @xmath38 , denoted by @xmath39 , which is the number of edges adjacent to @xmath38 , that is , @xmath40 where @xmath41 .
the multiset of all da - ecards is the _ da - edeck_. monikandan and raj @xcite initiated the study of the degree - associated edge - reconstruction number , denoted by @xmath42 which is the minimum @xmath43 such that some multiset of @xmath43 da - ecards determines @xmath0 .
clearly @xmath44 .
they determined @xmath42 where @xmath0 is a regular graph , a complete bipartite graph , a path , a wheel or a double star .
they also proved that @xmath45 where @xmath0 is a complete 3-partite graph whose part - sizes differ by at most 1 .
they showed that if @xmath0 is a graph obtained from @xmath46 by subdividing each edge at most once , then @xmath45 . in her study ,
myrvold @xcite proposed the _ adversary reconstruction number _ of @xmath0 , denoted by @xmath47 , which is the smallest value of @xmath43 such that no subdeck of @xmath0 containing @xmath43 cards is in the deck of any other graph which is not isomorphic to @xmath0 .
therefore @xmath47 equals 1 plus the largest number of cards which @xmath0 has in common with any graph not isomorphic to it . the _ adversary edge - reconstruction number _ of @xmath0 , denoted by @xmath48 ,
is analogously defined .
recently monikandan et al .
@xcite also introduced a similar parameter called the _ adversary degree - associated edge - reconstruction number _ of a graph @xmath0 , denoted by @xmath49 , which is the least number @xmath43 such that every collection of @xmath43 da - ecards of @xmath0 is not contained in the da - edeck of any other graph @xmath20 such that @xmath50 . from the definitions
, it follows that @xmath51 and @xmath52 .
moreover , if all the da - ecards of a graph @xmath0 are isomorphic , then @xmath53 .
ma , shi , spinoza and west @xcite recently showed also that for all complete multipartite graphs and their complements @xmath54 is usually 2 except for some exceptions .
they also pointed out that a significant difference between vertex and edge degree - associated reconstruction number is that while trivially , a graph and its complement have the same @xmath55 @xcite , they need not have the same value of @xmath54 . in this paper we first state some known results on the edge - reconstruction number of disconnected graphs and trees and then present some new results on their respective degree - associated edge - reconstruction numbers .
there is a large gap between the value of @xmath56 for disconnected graph @xmath0 with at least two non - isomorphic components and the value @xmath57 for disconnected graph @xmath0 with all components isomorphic on @xmath58 edges .
we therefore study whether this gap can be narrowed by considering the corresponding degree - associated edge - reconstruction numbers
. then we shall shift focus onto the degree - associated edge - reconstruction number of caterpillars and some other special classes of trees , with the main aim of obtaining some results towards determining the degree - associated edge - reconstruction number of a tree .
in @xcite , molina started to tackle the edge - reconstruction number of disconnected graphs .
he showed that the edge - reconstruction results are similar to the vertex reconstruction results stated by myrvold @xcite , but a significant difference is that whereas the vertex reconstruction number of a graph is always three or more , the edge - reconstruction number of a disconnected graph is often two . in summary , these are molina s main results : let @xmath0 be a disconnected graph with at least four edges and at least two non - trivial components ( that is , components that have more than one vertex ) .
then * if not all components are isomorphic , then @xmath59 ; * if all components are isomorphic , then @xmath60 where @xmath58 is the number of edges in a component ; * if there exists a pair of non - isomorphic components in which one component has a cycle and @xmath0 does not have any components isomorphic to either @xmath61 or @xmath62 , then @xmath63 .
he also observed that the value of @xmath64 is attained , giving as an example the graph consisting of @xmath12 copies of @xmath65 . in @xcite , asciak and lauri used line graphs in order to prove and extend molina s results .
in fact they proved the following results : let @xmath0 be a disconnected graph with at least four edges and the property that all components are isomorphic to a graph @xmath20 .
then * if @xmath20 is isomorphic to @xmath61 , then @xmath66 ; * if @xmath20 is isomorphic to @xmath62 , then @xmath67 ; * if @xmath20 is not isomorphic to @xmath61 or @xmath62 , then @xmath68 , where @xmath58 is the number of edges in @xmath20 . moreover , if @xmath69 then @xmath70 . let @xmath0 be a disconnected graph consisting of exactly two types of non - trivial components , namely those isomorphic to @xmath61 and those isomorphic to @xmath62
. then @xmath71 .
they also tried to investigate conditions which force or do not allow @xmath1 to be equal to 2 and also showed that in general , there is no straightforward relationship between the edge - reconstruction number of @xmath0 and that of its components .
these results and data from rivshin s computer search @xcite ( which showed that out of more than a billion graphs on at most eleven vertices , only fifty - six disconnected graphs have edge - reconstruction number greater than 3 and that out of these disconnected graphs , only four graphs do not have isolated vertices as components , namely @xmath72 and @xmath73 ) led asciak and lauri to make the following conjecture .
suppose that @xmath74 for a disconnected graph all of whose components are isomorphic to @xmath20 .
then @xmath20 is isomorphic to the star @xmath75 where @xmath76 is the number of edges .
we shall first describe some special types of trees and also give some basic definitions on general trees .
a _ caterpillar _ is a tree whose non - leaf vertices ( a _ leaf _ or an _ endvertex _ is a vertex of degree 1 ) induce a path called the _ spine _ of the caterpillar .
a caterpillar will be represented as the sequence @xmath77 which denotes a caterpillar with spine @xmath78 such that @xmath79 leaf vertices are incident to @xmath80 for each @xmath81 . clearly
this representation of a given caterpillar is unique up to left to right orientation .
note that for @xmath82 , each @xmath83 , but @xmath84 .
such a sequence is called a _
caterpillar sequence_. a special type of tree denoted by @xmath85 is a tree similar to a star ( a star is the tree on @xmath7 vertices , @xmath86 of which are endvertices ) which consists of three paths on @xmath87 , @xmath88 and @xmath89 edges , respectively , emerging from a common vertex .
some examples are shown in figure 1 . ;
( b ) @xmath90 ; ( c ) @xmath91 a tree is called a _ quasipath _ if it is either a path @xmath6 on @xmath7 vertices or one of the two special trees @xmath92 and @xmath90 .
we define the _ weight _ of a vertex @xmath16 of a tree @xmath13 , denoted by @xmath93 , to be the number of vertices in a largest component of @xmath94 .
the _ centroid _ of a tree @xmath13 is the set of all vertices with minimum weight ; this weight denoted by @xmath95 .
centroidal vertex _ is a vertex in the centroid .
it is well - known that the centroid of a tree consists of either one vertex or two adjacent vertices .
a tree with one centroidal vertex is called _ unicentroidal _ while a tree with two centroidal vertices is called _ bicentroidal_. in the latter case , the edge joining the centroidal vertices is called the _ centroidal edge_. when t is bicentroidal with centroidal edge @xmath38 , then the two components of @xmath96 are said to be _
centroidal components_. rivshin @xcite , using his computer program , obtained the following result which helps complete the theoretical results given in @xcite .
+ * if @xmath97 is either one of the two trees @xmath98 and @xmath99 shown in figure 2 , then @xmath100 . *
if @xmath13 is a bicentroidal tree in which vertices @xmath87 and @xmath88 are the centroidal vertices so that the two centroidal components of @xmath101 are @xmath90 , then @xmath102 .
but if the two components are @xmath92 then @xmath103 unless @xmath13 is the tree @xmath104 shown in figure 2 where @xmath105 .
( b ) @xmath99 ( c ) @xmath104 ] molina @xcite gave the following result .
if @xmath13 is a unicentroidal tree on at least four edges then @xmath107 then asciak , lauri , myrvold and pannone @xcite proved the following . every bicentroidal tree except the caterpillars with sequences @xmath108 and
the graph @xmath99 shown in figure 2 has edge - reconstruction number equal to 2 . for unicentroidal trees , the results of the computer searches in @xcite and those of rivshin , led asciak , lauri , myrvold and pannone to the infinite family of trees @xmath109 @xmath110 , where @xmath43 is the degree of the central vertex ( depicted in figure 3 ) having @xmath111 .
note that when @xmath112 , @xmath109 is the caterpillar @xmath113 . with @xmath111 ]
they also found the graph @xmath115 on fifteen vertices , shown in figure 4 , which does not fall within any known infinite class but which also has @xmath111 . these computer searches and results presented in @xcite led them to make the following conjecture for unicentroidal trees .
@xcite the only infinite classes of trees which have @xmath111 are the paths on an odd number of vertices , the caterpillars @xmath116 of even diameter , and the family of trees @xmath109 described above . on fifteen vertices with @xmath111 ]
continuing on the work presented by molina and more recently by asciak and lauri as described in section 2 , we shall shift our study to the degree - associated edge - reconstruction number of disconnected graphs wherein the degree of the deleted edge @xmath39 is given together with the edge - card .
but first we need the following two lemmas , the first of which is due to ma et al . in @xcite .
if @xmath0 has an edge @xmath38 such that @xmath117 or no two non - adjacent vertices in @xmath118 other than the endpoints of @xmath38 have degree - sum @xmath39 , then da - ecard @xmath119 determines @xmath0 .
the condition in lemma 4.1 is sufficient but not necessary for @xmath119 to determine @xmath0 .
in fact if @xmath0 is a graph in which an edge joins two disjoint complete graphs , then the condition fails , but @xmath120 .
let @xmath38 be an edge in a graph @xmath0 and @xmath119 be a given da - ecard . if @xmath118 is without isolated vertices , then * if @xmath121 , then @xmath38 is incident to two endvertices from @xmath118 ; * if @xmath122 , then @xmath38 is incident to at most one endvertex from @xmath118 .
however , if the edge - card @xmath118 has isolated vertices and @xmath123 , then the deleted edge can either join an isolated vertex to a non - endvertex or is incident to two endvertices whenever @xmath121 or is incident to two non - endvertices whenever @xmath122 .
* let @xmath124 be an edge in a graph @xmath0 .
suppose first that the edge - card @xmath118 has no isolated vertices .
we show that ( 1 ) and ( 2 ) hold . if @xmath121 , then there must be two edges adjacent to @xmath38 , so the only way to place the edge @xmath38 in order to obtain @xmath119 as a da - ecard is to join two endvertices .
if however @xmath125 , then the missing edge @xmath38 can join either an endvertex to a non - endvertex or two non - endvertices .
now suppose that the edge - card @xmath126 has at least one isolated vertex . then if @xmath123 , there is the possibility ( apart from the previous situation ) that the missing edge joins an isolated vertex to a non - endvertex .
+ we are now in a position to find the degree - associated edge - reconstruction number of disconnected graphs for a number of cases which include those mentioned in section 2 . but first we need the following definition .
suppose that a graph @xmath127 has in its edge - deck the edge - cards @xmath128 , we then say that @xmath20 is a _ blocker _ for these edge - cards or that @xmath20 _ blocks _ these edge - cards .
let @xmath0 be a disconnected graph with at least two non - trivial components all of which have at least three edges . then * if all components are isomorphic to @xmath61 , then @xmath120 ; * if all components are isomorphic to @xmath62 , then @xmath129 ; * if all components are isomorphic to @xmath65 , where @xmath58 is greater than 3 , then @xmath120 ; * if all components are of exactly two types , namely those isomorphic to @xmath61 and those isomorphic to @xmath62 , then @xmath130 ; * if @xmath0 has either * * only @xmath61 components and at least one isolated vertex , or * * only @xmath62 components and at least one isolated vertex , + then in both cases @xmath129 .
* proof * * ( 1)*. as the graph @xmath0 is made up of only @xmath12 copies of the @xmath61 component , then its da - edeck consists of @xmath131 copies of da - ecard @xmath132 where @xmath133 has @xmath134 copies of @xmath61 and a path @xmath135 . by lemma 4.1
, every da - ecard ( @xmath133,2 ) determines @xmath0 . *
( 2)*. since all components in graph @xmath0 are isomorphic to @xmath62 then all da - ecards of @xmath0 are copies of @xmath136 where @xmath137 consists of @xmath134 copies of @xmath62 , the path @xmath135 and an isolated vertex . by lemma 4.2 ,
the possible graphs having @xmath136 as a da - ecard is either the original graph @xmath0 or graph @xmath20 which is obtained from @xmath137 by joining the two endvertices of @xmath135 . since there exist only three da - ecards in the da - edeck of @xmath0 that are also in the da - edeck of @xmath20 , therefore @xmath129 . *
( 3)*. since @xmath0 consists of @xmath12 copies of @xmath65 where @xmath58 is the number of edges , then all da - ecards are of the form @xmath138 where @xmath139 has @xmath140 components , a @xmath141 component and an isolated vertex . by lemma 4.1 ,
every da - ecard @xmath142 determines @xmath0 . *
( 4)*. let @xmath0 consist of @xmath12 copies of @xmath62 and @xmath143 copies of @xmath61 , so the da - edeck of @xmath0 consists of two different da - ecards @xmath144 and @xmath145 .
the da - ecard @xmath144 is the graph @xmath146 where @xmath147 is an edge in @xmath61 , so @xmath148 consists of @xmath12 copies of @xmath62 , @xmath149 copies of @xmath61 and a component @xmath135 .
the da - ecard @xmath145 is the graph @xmath150 where @xmath151 is an edge in @xmath62 and therefore @xmath152 consists of @xmath143 copies of @xmath61 , @xmath134 copies of @xmath62 , a component @xmath135 and an isolated vertex .
suppose that graph @xmath153 can be obtained from @xmath150 by adding an edge @xmath38 whose @xmath121 . by lemma 4.2
, @xmath154 must be one of the following graphs : * @xmath98 is a disconnected graph whose components are a caterpillar @xmath155 , @xmath156 copies of @xmath62 , @xmath143 copies of @xmath61 and an isolated vertex , * @xmath99 is a disconnected graph whose components are a caterpillar @xmath157 , @xmath158 copies of @xmath62 , @xmath143 copies of @xmath61 , a component @xmath135 and an isolated vertex , * @xmath104 is a disconnected graph whose components are a component @xmath159 , formed by adding an edge to two endvertices of a @xmath62 component , @xmath156 copies of @xmath62 , @xmath143 copies of @xmath61 , a component @xmath135 and an isolated vertex .
* @xmath160 is a disconnected graph whose components are component @xmath159(a similar component to the above ) , which is obtained by joining an isolated vertex to a @xmath61 component , @xmath134 copies of @xmath62 , @xmath149 copies of @xmath61 and a component @xmath135 .
in all cases there is at most one da - ecard in the da - edeck of @xmath0 that is in the da - edeck of @xmath154 , @xmath161 .
thus @xmath130 since graphs @xmath98 ,
@xmath99 , @xmath104 can only contain one da - ecard in common with @xmath0 , while the specific choice @xmath144 , @xmath145 of da - ecards rules out graph @xmath160 . * ( 5)*. case ( i ) : let @xmath0 consist of at least two @xmath61 components and @xmath76 isolated vertices , @xmath162
. then all da - ecards of @xmath0 are copies of @xmath163 where @xmath164 is a disconnected graph consisting of a @xmath135 component , @xmath134 copies of @xmath61 and @xmath76 isolated vertices .
a disconnected graph @xmath165 made up of @xmath134 copies of @xmath61 , a @xmath62 component and @xmath166 isolated vertices has three copies of @xmath163 in its da - edeck , so @xmath167 .
using lemma 4.2 , the missing edge @xmath168 can join the two endvertices in @xmath164 in which case the resultant graph @xmath169 ; otherwise , an isolated vertex is joined either to a @xmath61 component or @xmath135 component . in the latter cases , each resulting graph can share at most three da - ecards with @xmath0 , so @xmath170 . hence @xmath171 .
case ( ii ) : similar arguments presented in case ( i ) hold for this case . in the cases considered so far , we did not investigate the case when all the components of @xmath0 are all isomorphic and not equal to either @xmath61 or @xmath172 .
the next results will address this situation . in this section
we consider only @xmath173 , that is , the disjoint union of copies of the connected graph @xmath20 , since otherwise as pointed out in section 2 , @xmath1 is at most 3 @xcite .
it is also known that if @xmath174 , @xmath1 can be as large as @xmath64 and this can happen only when @xmath175 @xcite . however , in this case , @xmath120 except for the case @xmath176 considered above .
therefore , can the degree - associated edge - reconstruction number help to reduce the gap between 3 and @xmath64 ?
but first we need to define some definitions and state some important results which will be required throughout this sub - section .
bipartite graph _ is one whose vertex set can be partitioned into two sets such that no edge joins two vertices in the same set .
these two sets are called the _
colour classes _ of the bipartite graph , and they are uniquely defined if the graph is connected .
a graph @xmath0 is _ edge - transitive _ if , given any two edges @xmath177 and @xmath178 , there is an _ automorphism _ @xmath179 ( that is , a one - one mapping of the vertex set of graph @xmath0 onto itself which preserves adjacency ) such that @xmath180 .
it is easy to show that if all edge - cards @xmath181 are isomorphic then @xmath20 must be edge - transitive .
the complete bipartite graph @xmath182 is an example of an edge - transitive graph .
we shall also need the following easy result .
let @xmath20 be an @xmath76-regular graph .
+ then @xmath183 .
recall that the minimum degree of a graph @xmath0 denoted by @xmath30 is the smallest number of edges incident to any vertex @xmath16 in @xmath0 .
we are now able to start considering the case when all the edge - cards of a disconnected graph are isomorphic .
let @xmath0 be a disconnected graph , all of whose components are isomorphic to @xmath20 , and suppose that all edge - cards of @xmath20 are isomorphic .
suppose also that @xmath184 .
then @xmath45 . *
* since all edge - cards @xmath181 are isomorphic then @xmath20 is edge - transitive .
also all edges have degree @xmath185 , where @xmath186 are the degrees of the two endvertices of the edge ; @xmath76 and @xmath187 are the same for all edges , that is , @xmath20 is either regular ( when @xmath188 ) or bi - degreed ( that is , all vertex degrees are either @xmath76 or @xmath189 ) although we do not know , from the da - ecards , the values of @xmath76 and @xmath187 .
due to proposition 4.1 we may assume that @xmath20 is not regular .
in this case , @xmath20 is bipartite , with the vertices of different degree forming the two colour - classes of @xmath20 .
we shall assume that @xmath20 has @xmath87 vertices of degree @xmath76 and @xmath88 vertices of degree @xmath187 , where @xmath190 . also , since we are assuming that @xmath184 , then @xmath191 .
all da - ecards are of the type @xmath192 .
suppose we are given two such da - ecards @xmath193 .
then we need to consider the following two cases .
+ + _ case 1 _ : @xmath193 are blocked by a graph @xmath194 , where @xmath195 with @xmath196 in @xmath197 .
but since @xmath198 must be @xmath12 in @xmath197 and its endvertices can only have degrees @xmath76 and @xmath187 in @xmath20 , the only possibility is that @xmath199 joins two vertices of degree @xmath187 in @xmath20 , and therefore @xmath200 @xmath201
so , we have that @xmath202 has @xmath87 vertices of degree @xmath76 and @xmath203 vertices of degree @xmath187 . also , @xmath197 has @xmath87 vertices of degree @xmath76 , @xmath204 vertices of degree @xmath187 , and two vertices of degree @xmath205 .
moreover , @xmath197 must have an edge @xmath206 such that @xmath207 in order for @xmath133 to be a blocker of the da - ecards @xmath208 .
but @xmath209 has a vertex of degree @xmath210 which @xmath20 does not .
this contradiction shows that @xmath208 can not have a blocker of this type .
+ + _ case 2 _ : @xmath193 are blocked by a graph @xmath211 , where @xmath212 but @xmath213 has two da - ecards in common with @xmath20 .
therefore we can take @xmath213 to be equal to @xmath214 with @xmath41 , @xmath215 , @xmath216 in @xmath20 and @xmath217 , but such that @xmath218 in @xmath213 is equal to @xmath219 in @xmath20 . but @xmath218 in @xmath213 is equal to @xmath220 where @xmath221 ( these being the degrees of the vertices in @xmath181 ) but @xmath222 ( since @xmath223 ) .
so we have these two possibilities : * @xmath168 is adjacent to two vertices of degree @xmath187 in @xmath181 .
therefore @xmath224 * @xmath168 is incident to two vertices of degree @xmath187 and @xmath225 in @xmath181 .
therefore @xmath226 we shall now consider each of these two possibilities : + _ case 2(a ) _ : let @xmath227
. therefore @xmath213 has one vertex @xmath228 of degree @xmath225 , the vertex @xmath16 of degree @xmath229 and the vertices @xmath230 of degree @xmath205 . since @xmath231 then @xmath232 , and @xmath233 are the only vertices of @xmath213 with this degree .
also , since @xmath234 , each of the two vertices @xmath230 are adjacent , in @xmath20 , to at least three vertices of degree @xmath76 . therefore , in @xmath213 ,
each of @xmath230 is adjacent to at least two vertices of degree @xmath76 .
but we know that for @xmath133 to block both @xmath148 and @xmath152 , @xmath213 must have an edge @xmath235 such that @xmath236 is isomorphic to an edge - card of @xmath20 . comparing the degrees in an edge - card of @xmath20 with the degrees in @xmath213
we see that @xmath199 must be adjacent , in @xmath213 , to two vertices of degree @xmath205 .
but since @xmath237 , the endvertices of @xmath199 must be @xmath16 and one of @xmath238 or @xmath239 .
this means that in @xmath236 there still is a vertex ( @xmath238 or @xmath239 ) of degree @xmath205 adjacent to a vertex of degree @xmath76 . but @xmath181 contains no such pair of adjacent vertices .
therefore , this case can not happen .
+ + _ case 2(b ) _ : we proceed similarly to the previous case .
let @xmath227 and suppose that @xmath240 and @xmath241 in @xmath181
. therefore @xmath238 must be the vertex @xmath228 and @xmath239 can not be the vertex @xmath16 , since @xmath242 .
therefore , in @xmath213 , @xmath239 is a vertex of degree @xmath243 adjacent to one vertex , @xmath238 , of degree @xmath187 , possibly another vertex , @xmath16 , also of degree @xmath187 , and the remaining @xmath244 neighbours all of degree @xmath76 .
since @xmath245 and @xmath246 , this means that @xmath239 is adjacent to at least two non - adjacent vertices of degree @xmath76 . as before , for @xmath133 to be a blocker of @xmath148 and @xmath152 , @xmath213 must have an edge @xmath235 such that @xmath247 is the same as a da - ecard of @xmath20 .
therefore @xmath199 must be incident to a vertex of degree @xmath187 and a vertex of degree @xmath76 .
now , since @xmath235 , these two vertices can not be @xmath238 and @xmath239 .
also , @xmath199 can not be @xmath248 ( supposing @xmath16 were adjacent to @xmath239 in @xmath20 ) , because in this case , @xmath16 would be a vertex of degree @xmath249 adjacent to vertices of degree @xmath187 in @xmath236 , which no edge - card of @xmath20 has .
but then , since @xmath239 is adjacent to at least two non - adjacent vertices of degree @xmath76 in @xmath213 , then it must be adjacent to at least one vertex of degree @xmath76 in @xmath236 , that is , @xmath236 contains an edge joining two vertices of degree @xmath76 , which no edge - card of @xmath20 has .
therefore even here we have shown that this case can not hold
. + + * remark . *
if we allow @xmath250 and @xmath251 , then @xmath252 has @xmath54 equal to 4 and @xmath253 ( see later ) and if we allow @xmath254 then @xmath255 has @xmath54 equal to 3 .
in fact we believe that these are the only three cases when @xmath256 , and we conjecture : let @xmath0 be as in theorem 4.2 and suppose that the condition @xmath184 is replaced by @xmath257 , @xmath258 and @xmath259 . then @xmath45 .
we now consider the case when not all the edge - cards of @xmath20 are isomorphic .
it seems that this case is more difficult to settle .
so we shall give only a very partial result .
but first we need to define the minimum multiplicity of a graph .
the _ multiplicity _ of an edge - card of @xmath20 is the number of times it appears in the edge - deck of @xmath20 , and the _ minimum multiplicity _ of @xmath20 , denoted by @xmath260 is the minimum amongst all multiplicities of edge - cards appearing in the edge - deck of @xmath20 .
we give this simple result which does not use the concept of @xmath261 .
we believe that the result is far from being the best possible .
let @xmath173 and suppose not all edge - cards of @xmath0 are isomorphic . then @xmath262 * proof .
* let @xmath263 be two non - isomorphic edge - cards of @xmath20 .
then , any collection of edge - cards of @xmath0 which contains @xmath264 and @xmath265 can only be blocked by a graph @xmath266 . in this case , if the collection of edge - cards is @xmath267 then @xmath213 must block the edge - cards @xmath268 . if these edge - cards are all obtained from the same component of @xmath0 , and @xmath269 then no @xmath213 can block them .
if we let @xmath270 be isomorphic copies of a single edge - card of @xmath20 and @xmath271 then again no @xmath213 can block @xmath268 . in the following section we shall investigate @xmath2 , where @xmath0 is a special case of disconnected graphs whose components are all isomorphic but not all of its edge - cards are isomorphic .
we can easily verify that some special trees such as paths @xmath6 and stars @xmath11 have degree - associated edge - reconstruction number of 1 .
in theorem 4.1 we have shown that even @xmath43 copies of stars @xmath11 for @xmath272 have degree - associated edge - reconstruction number equal to 1 .
but we shall show that while @xmath273 for @xmath4 , @xmath274 .
@xmath275 for @xmath4 and @xmath276 let @xmath0 be a graph consisting of @xmath43 copies of @xmath6 .
then there are two types of da - ecards which are either @xmath277 , where @xmath278 is a disconnected graph having @xmath279 components isomorphic to @xmath6 , a @xmath280 component and an isolated vertex , or @xmath281 , where @xmath282 is a disconnected graph having @xmath279 components isomorphic to @xmath6 and two other path components @xmath283 and @xmath284 where @xmath285 .
now graphs other than @xmath0 having either da - ecard @xmath148 or @xmath152 are @xmath286 and @xmath287 where @xmath285 , respectively .
this means that @xmath288 , so @xmath148 or @xmath152 alone do not give @xmath289 uniquely where @xmath290 .
we now show that @xmath148 , @xmath152 determine @xmath0 .
let @xmath133 be a reconstruction from \{@xmath291 } and suppose that @xmath133 is obtained by adding a new edge @xmath168 to the edge - card @xmath278 .
now since the degree of @xmath168 is 1 , the isolated vertex can join an endvertex of a component @xmath6 to obtain the graph @xmath292 .
but the resultant graph @xmath292 does not have the edge - card @xmath282 .
so the only possibility is that @xmath133 is obtained by joining the isolated vertex to an endvertex of path @xmath293 , so @xmath294 .
@xmath274 let @xmath0 be a graph consisting of @xmath43 copies of @xmath135 .
then @xmath0 has only one distinct da - ecard which is of the form @xmath295 where @xmath278 is a disconnected graph having @xmath279 copies of component @xmath135 , and components @xmath296 and @xmath297 . with just one such da - ecard
one can conclude that any reconstruction has paths for all its components .
now the graph @xmath298 has two da - ecards in common with @xmath0 and so two da - ecards are not sufficient to reconstruct @xmath0 uniquely . but three da - ecards suffice in order to force a reconstruction to have all components isomorphic to @xmath135 because the reconstruction can only be obtained from @xmath278 by joining the isolated vertex to the smallest component @xmath299 .
the empirical evidence provided by david rivshin showed that , after investigating graphs on at most eleven vertices , only seventeen trees have edge - reconstruction number equal to 3 .
three out of these seventeen trees are the bicentroidal trees @xmath98 , @xmath99 and @xmath104 shown in figure @xmath300 .
we can easily check directly that @xmath301 while @xmath302 .
the remaining trees are unicentroidal and we can also show by hand that , the caterpillars @xmath113 , @xmath303 , @xmath304 , the paths of odd order @xmath305 , @xmath306,@xmath307 and @xmath308 and the tree @xmath292 shown in figure 3 have @xmath309 while the remaining six unicentroidal trees , namely , @xmath310 , @xmath311 , @xmath312 , @xmath313 , @xmath314 and @xmath315 can also be directly checked by hand to show that their @xmath316 .
previously we have stated that from computer search , the edge - reconstruction number of graph @xmath115 of figure 4 is 3 . but also
, in this case , we can show directly that @xmath317 .
barrus and west @xcite have shown that , except for the case of the 6-vertex caterpillar @xmath318 shown in figure 2 , @xmath55 of a caterpillar is 2 , or 1 , for stars .
we shall now restrict our study in this section to the degree - associated edge - reconstruction number of caterpillars .
but we first quote a very useful result by molina @xcite which allows us to identify a graph as a tree from two given edge - cards .
let @xmath0 be a graph with edges @xmath147 and @xmath151 .
suppose that edge - card @xmath146 has two components which are trees of orders @xmath319 and @xmath320 while edge - card @xmath150 has another two components which are trees of orders @xmath321 and @xmath322 .
if @xmath323 , then @xmath0 is a tree .
moreover , we shall also make use of the next observation in order to identify when a tree is a caterpillar .
let @xmath13 be a tree and let @xmath38 be an edge whose degree @xmath39 is greater than 1 and that @xmath96 is a caterpillar with an isolated vertex .
then @xmath13 is a caterpillar .
therefore the fact that @xmath13 is a caterpillar , is recognisable from any two da - ecards as long as at least one of them is not obtained by deleting an edge which changes the spine of the caterpillar . from now on
, we shall assume that @xmath13 is a caterpillar and since a caterpillar sequence @xmath324 denotes a path which has @xmath325 we shall henceforth assume that @xmath13 is not a path .
we shall only consider da - ecards corresponding to @xmath96 and @xmath326 where @xmath327 and @xmath199 are end - edges which do not change the spine of @xmath13 .
this is equivalent to reconstructing a caterpillar sequence @xmath328 from two sequences @xmath329 and @xmath330 and we shall consider the problem this way . note that the two sequences are given only up to left to right orientation ; we call two such orientations the reverse of each other
. we shall also call @xmath329 a _ reduction _ of the sequence @xmath328 .
a reduction is therefore a description of the corresponding da - ecard of @xmath13 without the isolated vertex .
if a number of reductions of a sequence determines the sequence , we can say that the reductions reconstruct the sequence .
we shall need the definition of conjugate entries in a caterpillar sequence . given the caterpillar sequence @xmath331 ,
then the entries @xmath79 , @xmath332 are said to be _ conjugate_. + * notation .
* the term @xmath333 will stand for @xmath334 and @xmath335 for @xmath336 .
if the caterpillar sequence @xmath337 is not reconstructible from the reductions @xmath338 and @xmath339 then @xmath340 and all other conjugate entries are equal , except for a pair which differ exactly by one
. * proof . * for non - reconstructibility we must be able to get another sequence from the alignment @xmath341 and @xmath342 suppose , for contradiction , that @xmath343 , therefore @xmath344 .
therefore for the above alignment to lead to a reconstruction of @xmath278 , all other conjugate entries must be equal and @xmath345 must differ by exactly one , say , without loss of generality , @xmath346 .
then the given alignment reconstructs as @xmath347 but , since all conjugate entries apart from @xmath348 are equal , this sequence is simply the reverse of the original sequence @xmath278 , therefore reconstruction is unique .
( the above reasoning can be noted by following example @xmath349 ) .
hence , for non - unique reconstruction , @xmath350 and @xmath351 .
therefore , for the above alignment to lead to a reconstruction of @xmath278 , we must have for some @xmath352 , that the conjugate pair @xmath353 differ by exactly one , say , @xmath354 , and all other conjugate entries are equal , as required .
thus we get the following reconstruction @xmath355 ( example @xmath356 illustrates this situation ) .
+ a caterpillar sequence can be reconstructed from two reductions .
therefore , if @xmath13 is a caterpillar with at least two end - edges whose removal does not change its spine , then @xmath357 .
* suppose the caterpillar sequence @xmath358 is not reconstructed from @xmath359 and @xmath360 .
then by lemma 5.2 , @xmath361 and , for some @xmath352 , the conjugates @xmath362 differ by one . but then , since @xmath362 are not equal , the sequence is reconstructed from @xmath363 and @xmath364 . if @xmath13 is a caterpillar which is not a path , then @xmath357 . * proof .
* the only remaining case to consider is when the caterpillar @xmath13 has only one end - edge whose removal does not change its spine .
but it is easy to check that for such caterpillars @xmath357 .
a caterpillar is expressed by the sequence @xmath365 and the following two reductions representing two edge - deleted caterpillars are @xmath366 @xmath367 where the second reduction is reversed with respect to the first one . comparing the two reductions will give the sequence @xmath368 , which is simply the reverse of the original sequence of the caterpillar .
let a caterpillar be expressed by the sequence @xmath369 and the deletion of two of its end - edges gives the following reductions @xmath370 @xmath371 in which the second reduction is reversed . by comparing the two reductions ,
the sequence @xmath372 is obtained , which is an alternative sequence to the original one .
so the two sequences given by the two corresponding edge - cards do not reconstruct the given caterpillar . in @xcite ,
monikandan et al . defined a _
star - like tree _ to be a tree obtained from @xmath373 by subdividing each edge at most once .
they proved that any such tree has @xmath54 at most 2 .
as already pointed out in section 4.1 , stars @xmath11 have degree - associated edge - reconstruction number of 1 .
we shall show that if @xmath10 where @xmath374 is a tree obtained from the star @xmath11 by subdividing each edge @xmath12 times , as shown in figure 5 , then @xmath9 .
( b ) @xmath375 let @xmath10 be a tree defined as above
. then @xmath376 .
by considering all possible da - ecards of @xmath10 , it is easy to show that it is not possible to reconstruct @xmath10 uniquely from only one da - ecard , therefore @xmath377 .
let two da - ecards of the form @xmath378 each be obtained by deleting an end - edge from @xmath10 , so @xmath379 is made up of a unicentroidal tree in which all paths that emerge from the centroidal vertex are of the same order except one path whose order is one less than the others , and an isolated vertex .
we claim that these two da - ecards reconstruct @xmath10 uniquely .
since in both cases the degree of the deleted edge is 1 , the missing edge can only join the isolated vertex to any of the endvertices of @xmath379 .
but if the isolated vertex is joined to any one of the equal paths that emerge from the centroidal vertex then the resultant tree can not have the two above mentioned da - ecards as part of its degree - associated edge - deck .
hence the only possible alternative is to join the isolated vertex to the smallest path emerging from the centroidal vertex in @xmath379 . but
this is isomorphic to @xmath10 .
hence @xmath376 . based on these findings , the results in section 3 , and the fact that @xmath380
, we offer the following conjecture . if @xmath13 is a tree , then @xmath14 .
i am grateful to josef lauri for his many helpful suggestions which largely improved the style of this paper and sections 4.1 and 5 . 32 k.asciak , m.a.francalanza , j.lauri and w.myrvold . a survey of some open questions in reconstruction numbers . _ ars
_ , 97:443456 , 2010 .
k.asciak , j.lauri . on the edge - reconstruction number of disconnected graphs .
_ bulletin of the inst.of combin . and its appl._,63:87100 , 2011 .
k.asciak , j.lauri , w.myrvold and v. pannone . on the edge reconstruction number of a tree .
_ australasian j. of combin . _ vol 60(2 ) , 154175 , 2014 .
m.d.barrus and d.b.west .
degree - associated reconstruction number of graphs .
_ discrete math .
_ , 310:2600 - 2612 , 2010 .
j.a.bondy and u.s.r.murty .
_ graph theory ( graduate texts in mathematics ) .
_ springer , 2008 .
f.harary , on the reconstruction of a graph from a collection of subgraphs , in _ theory of graphs and its applications _ ( proceedings of the symposium held in prague , 1964 ) , edited by m.fiedler , czechoslovak academy of sciences , prague , 1964 , 47 - 52 .
f.harary and j.lauri . on the class - reconstruction number of trees .
_ oxford , 39:47 - 60 , 1988 .
f.harary and j.lauri .
the class - reconstruction number of maximal planar graphs . _ graphs and combinatorics _ vol 3 , 1:45 - 53 , 1987 .
f.harary and m.plantholt .
class reconstruction numbers of unicyclic graphs , _ appl.math .
( warsaw ) _ 20 ( 1988 ) 117123 .
f.harary and m.plantholt .
the graph reconstruction number .
_ j.graph theory _ 9 : 451 - 454 , 1985 .
p.j.kelly . on isometric transformations .
phd thesis , university of winconsin , 1942 .
the reconstuction problem , in j.l.gross and j.yellen editors , _ handbook of graph theory _ 2nd edition , pages 77 - 100 , crc press , 2014 .
j.lauri and r.scapellato .
topics in graph automorphisms and reconstruction . cambridge university press , 2003 .
g.li and f.ruskey .
the advantages of forward thinking in generating rooted and free trees . in _ 10th annual acm - siam symposium on discrete algorithms ( soda ) _ , 1999 .
m.ma , h.shi , h.spinoza and d.b.west .
degree - associated reconstruction parameters of complete multipartite graphs and their complements , _ taiwanese j. math_. 19 ( 2014 ) , 12711284 .
the edge reconstruction number of a disconnected graph._j .
graph theory _ , 19(3):375 - 384 , 1995 .
the edge reconstruction number of a tree .
_ vishwa international journal of graph theory _ , 2(2 ) : 117 - 130 , 1993 .
w.myrvold . ally and adversary reconstruction problems .
phd thesis , university of waterloo , ontario , canada , 1988 .
the ally - reconstruction of a disconnected graph . _ ars .
_ 28:123 - 127 , 1989 .
s.monikandan , p.a.devi and s.sundar raj .
degree - associated edge reconstruction number of graphs .
_ journal of disc .
_ 23:35 - 41 , 2013 .
s.monikandan and s.sundar raj .
degree - associated edge - reconstruction number , in w.f .
symth , s. arumugam ( eds),combinatorial algorithms , 23rd international workshop , iwoca 2012 , krishnankoil , india , lecture notes in computer science 7643 , springer - verlag , berlin 2012 , 100109 .
s.monikandan and s.sundar raj , c.jayasekaran , and a.p .
santhakumaran . a note on the adversary degree associated reconstruction number of graphs . _ j.discrete math _
vol 2013 ( 2013 ) , 5pp .
s.ramachandran . degree - associated reconstruction number of graphs and digraphs .
int . j. math .
sci . _ 41 - 53 , 2000 .
reconstruction number for ulam s conjecture .
_ , 78:289 - 296 , 2006 .
d.rivshin . personal communication .
s.m.ulam . a collection of mathematical problems .
wiley , new york , 1960 . | an _ edge - card _ of a graph @xmath0 is a subgraph formed by deleting an edge .
the _ edge - reconstruction number _ of a graph @xmath0 , @xmath1 , is the minimum number of edge - cards required to determine @xmath0 up to isomorphism .
a _ da - ecard _ is an edge - card which also specifies the degree of the deleted edge , that is , the number of edges adjacent to it . the _ degree - associated edge - reconstruction number _ , @xmath2 is the minimum number of da - ecards that suffice to determine the graph @xmath0 . in this paper we state some known results on the edge - reconstruction number of disconnected graphs and trees .
then we investigate how the degree - associated edge - reconstruction number of disconnected graphs and trees vary from their respective edge - reconstruction number .
we show how we can select two da - ecards to identify caterpillars uniquely .
we also show that while @xmath3 for @xmath4 , @xmath5 where @xmath6 is the path on @xmath7 vertices , and that , although @xmath8 , @xmath9 where @xmath10 is a tree obtained from the star @xmath11 by subdividing each edge @xmath12 times . finally we conjecture that for any tree @xmath13 , @xmath14 .
keywords : degree - associated edge - reconstruction number , trees , disconnected graphs .
ams subject classification : 05c60 |
the question concerning the form of the most general , consistent scalar tensor theory of gravity is frequently being reconsidered due to its foremost physical importance for building models of inflation and dark energy @xcite . the problem is deeper than expected , and its theoretical investigation leads to intriguing surprises .
subtleties arise when considering derivative couplings of gravity to the scalar field , due to the non - linearity of the equations of motion , and the non trivial structure of the constraint equations . a systematic study of scalar - tensor theories as a modification of general relativity ( gr ) started with the work of brans and dicke @xcite , and subsequent developments mainly considered scalar - tensor lagrangians with non - minimal couplings of the scalar to curvature , but containing at most first derivatives of the scalar field ( see @xcite for a comprehensive review ) . the modern approach to scalar tensor theories started with the introduction of galileons , the most general scalar field theory in flat spacetime with second order equations of motion @xcite .
galileon actions contain second order derivatives acting on the scalar field .
they have very attractive phenomenological and theoretical features .
in particular , they admit self - accelerating solutions that can explain the phenomenon of dark energy , and at the same time exhibit screening mechanisms , which ensure consistency with the tight bounds on small scale deviations from gr
. once galileons are covariantised by naively replacing partial derivatives by covariant derivatives , the theory leads to third order derivatives in the equations of motion .
a way to avoid them is to introduce counter terms , in the form of non - minimal couplings to gravity @xcite .
it was then shown that these theories can be further extended by including free functions of the scalar field and its kinetic term @xcite .
remarkably , it was found that this theory was already proposed in 1974 by horndeski @xcite and this theory is now called the horndeski theory @xcite . the basic strategy to construct the horndeski theory is to ensure that the equations of motion are of at most second order in derivatives .
if the equations of motion contain higher order time derivatives , this generally leads to a ghost instability due to the ostrogradsky theorem .
however , it was suggested that despite the appearance of higher order derivatives it is still possible to avoid the ostrogradsky ghost if the theory contains a hidden constraint and the equations of motion can be recast into a second order form , for example by field redefinitions .
this possibility was studied in detail by ref .
@xcite . in ref .
@xcite , the authors constructed theories in the so - called unitary gauge where the scalar field depends only on time , thus allowing it to be used as a clock .
it was then found that there are two more terms with two arbitrary functions of the scalar field and its kinetic term that can be added to the horndeski theory .
remarkably , these two terms can be obtained by naively covariantising the flat space galileon in the form written with the anti - symmetric levi - civita tensor .
although the equations of motion contain third order time derivatives , it was argued that they do not lead to the ostrogradsky instability , since constraints remove this additional degree of freedom .
these terms are dubbed ` beyond horndeski ' .
there still remains the question of whether beyond horndeski can be obtained from horndeski by a suitable transformation of the metric .
it is indeed possible to perform generalised conformal and disformal transformations @xcite , that depend on the scalar field and its kinetic term , on horndeski and obtain a theory which contains higher order derivatives @xcite .
as long as the transformation is invertible , it does not change the physics . thus , despite the appearance of higher order derivatives , the resultant theory propagates the same number of degrees of freedom as horndeski .
however , in this case , the theory can be mapped into horndeski , so unless we introduce a coupling to matter , the new theory is nothing but horndeski .
indeed , it has been shown that a beyond horndeski term can be generated by a generalised disformal transformation from horndeski @xcite .
this brings us to the question of whether beyond horndeski is really beyond horndeski or not . in this paper ,
we re - examine the problem of determining the number of degrees of freedom in the theory of beyond horndeski @xcite , without selecting any gauge .
we reformulate the theory using geometrical quantities defined with respect to a constant scalar field hypersurface .
this representation gives a simple geometrical understanding of why the higher derivatives in beyond horndeski do not lead to any additional degrees of freedom .
we show this fact by demonstrating the existence of a primary constraint in the hamiltonian formulation .
we further discuss cases in which , when the theory is combined with horndeski , this constraint is spoiled .
we finally study the conditions for mapping beyond horndeski to horndeski by a generalised disformal transformation .
this paper is organised as follows . in section [ sec - rev ] ,
we introduce horndeski and beyond horndeski theories .
we formulate beyond horndeski using geometrical quantities defined with respect to the constant scalar field hypersurface . in section [ sec - kin ] , we derive kinetic terms for the metric and a scalar field focusing on its higher derivative terms . in section [ sec - prim ] , we examine the existence of a primary constraint , which is necessary for the non - propagation of the additional mode due to the higher derivatives . in section [ sec - disf ]
, we study the possibility of mapping beyond horndeski to horndeski by a generalised disformal transformation .
section [ sec - disc ] is devoted to conclusions .
in this section we review the theories of horndeski and beyond horndeski , and reformulate the latter using geometrical quantities . the scalar tensor theory of horndeski is given by the action @xmath0 where the lagrangian densities are @xmath1 \ , , \\ { } ^{h}{\cal l}_5&= & g_5(\phi , x)g^{\mu\nu}\phi_{\mu \nu } + \frac{1}{3}g_{5x}(\phi , x)\left[(\box\phi)^3 -3\ , \box \phi\ , \phi_{\mu \nu } \phi^{\mu \nu } + 2\ , \phi_{\mu \nu}\phi^{\nu \rho } \phi^{\mu}_{\rho}\right]\,.\end{aligned}\ ] ] @xmath2 and @xmath3 , @xmath4 are arbitrary functions of @xmath5 and @xmath6 , defined as x_^. this is the most general scalar tensor theory of gravity , involving a single scalar field , that leads to covariant second order equations of motion .
the cubic horndeski theory is described by the lagrangian @xmath7 . in this case gravity
becomes dynamical only through a mixing with the scalar field , a phenomenon dubbed kinetic gravity braiding , see @xcite . in the present work ,
we focus on the quartic @xmath8 , and the quintic @xmath9 horndeski theories , where the tensor spin-2 degrees of freedom have their own kinetic terms .
general relativity is recovered by setting @xmath10 , and @xmath11 . in order to study the dynamics of theories with second order derivatives in the action , it is convenient to adopt the field redefinition introduced in ref .
this is the natural covariant extension of the one used by ostrogradsky in his pioneering work ( see @xcite ) .
we define therefore a four vector @xmath12 using @xmath13 , the quartic and quintic horndeski lagrangians are rewritten as @xmath14 } \ ,
\nabla_{\alpha } a^{\mu } \nabla_{\beta } a^{\nu } \ , , \\ { } ^{h}{\cal l}_5&= & g_5(\phi , x)g^{{\alpha}}_{\mu } \ , \nabla_{{\alpha } } a^{\mu } + \frac{1}{3}g_{5x}(\phi , x)\ , \delta^{\alpha}_{[\mu } \delta^{\beta}_{\nu } \delta^{\gamma}_{\rho]}\ , \nabla_{\alpha } a^{\mu } \nabla_{\beta } a^{\nu}\nabla_{\gamma } a^{\rho } \ , , \ ] ] where @xmath15 $ ] denotes anti - symmetrisation in front of ( anti-)symmetrised tensor . ] and x = a _
a^ = a^2 .
as stated above , the horndeski system is the most general scalar tensor theory leading to covariant second order equations of motion for the scalar and tensor field , hence it certainly propagates at most three degrees of freedom .
intriguingly , it is conceivable that more general scalar tensor theories exist , whose equations of motion are higher order , but at the same time are characterised by constraints that remove additional , undesired degrees of freedom .
a proposal in this direction is the theory of `` beyond horndeski '' , described by the following action @xmath16 where @xmath17 , \\ { } ^{bh}{\cal l}_{5 } & = & f_5(\phi , \,x ) \big [ x \left ( ( \box\phi)^3- 3\ , \box \phi\ , \phi_{\mu \nu } \phi^{\mu \nu } + 2 \phi_{\mu \nu}\phi^{\nu \rho } \phi^{\mu}_{\rho } \right ) \nonumber\\ & - & 3 \left ( \left(\box \phi\right)^2 \phi_{\mu } \phi^{\mu \nu } \phi_{\nu } - 2\ , \box \phi\ , \phi_{\mu } \phi^{\mu \nu } \phi_{\nu \rho } \phi^{\rho } - \phi_{\mu \nu } \phi^{\mu \nu } \phi_{\rho } \phi^{\rho { \sigma } } \phi_{{\sigma } } + 2 \phi_{\mu } \phi^{\mu \nu } \phi_{\nu \rho } \phi^{\rho { \sigma } } \phi_{{\sigma } } \right ) \big ] , \end{aligned}\ ] ] @xmath18 and @xmath19 , @xmath20 are arbitrary functions of @xmath21 .
these actions reduce to galileonic actions in an appropriate decoupling limit , when gravity is turned off ( this has been shown in @xcite using the results of @xcite ) . on the other hand , when gravity is fully dynamical , beyond horndeski is characterised by equations of motion of order higher than two . using diffeomorphism ( diff . )
invariance and selecting the unitary gauge , one can show that constraint equations exist that are able to avoid the propagation of additional degrees of freedom @xcite .
hence , one would be tempted to freely add the aforementioned beyond horndeski lagrangians to the horndeski action of the previous subsection with arbitrary coefficients , so to form a more general scalar tensor theory .
however , as anticipated in @xcite , and as we shall further develop later on , issues arise that suggest that an additional propagating mode could appear in some cases .
such problems become manifest when one does _ not _ select the unitary gauge in which the scalar field depends only on time @xcite . in the following sections , treating the general system without selecting any gauge , we will make some progress in understanding the genuine dynamics induced by beyond horndeski theories . before starting our analysis
we introduce a suitable geometrical formulation for these lagrangians that greatly simplifies the calculation .
the basic point of this interpretation can be already found in @xcite , but again the choice of the unitary gauge hides the proper construction . the approach is to introduce quantities according to the constant scalar field hypersurface , @xmath22 const .
this hypersurface is characterised by the following geometrical quantities @xmath23 where @xmath24 is the unitary normal vector , @xmath25 the projection tensor on the hypersurface , and @xmath26 is the extrinsic curvature of the constant @xmath5 hypersurface multiplied by @xmath27 .
notice that these objects have nothing to do with the usual ones associated with a space - time foliation .
the beyond horndeski lagrangians can be expressed in terms of the extrinsic curvature @xmath28 in the following way : @xmath29 } \ , \phi^{\mu}_{\alpha } \phi^{\nu}_{\beta}\ , , \\ { } ^{bh}{\cal l}_5 & = & - x f_5(\phi,\,x)\ , \delta^{\alpha}_{[\mu } \delta^{\beta}_{\nu } \delta^{\gamma}_{\rho]}\ , \phi^{\mu}_{\alpha } \phi^{\nu}_{\beta } \phi^{\rho}_{\gamma}\,.\end{aligned}\ ] ] using the expression for the extrinsic curvature ( [ phiextrinsic ] ) ,
this can be written as @xmath30 where @xmath31}\ , , \qquad { \cal m}^{\alpha \beta \gamma}_{\mu \nu \rho } = p^{\alpha}_{[\mu } p^{\beta}_{\nu}p^{\gamma}_{\rho]}\,.\ ] ] the matrix @xmath32 has a crucial property for the arguments that we are going to develop : @xmath33 this property stems from the fact that @xmath32 is built in terms of the projection tensor ( [ proj ] ) . as we will show in section [ sec - prim ] , this leads to the existence of a primary constraint necessary to avoid the propagation of an extra mode in beyond horndeski , despite the fact that the equations of motion have higher derivatives .
we stress that the unitary gauge used in the literature @xcite is a very special choice of gauge where the constant time hypersurface and the constant scalar field hypersurface coincide .
the analysis in this special gauge could lead to misleading results , as was explicitly pointed out first in @xcite and then in @xcite .
for this reason , we will refrain from making this choice in what comes next .
in order to identify the kinetic terms and carry out the analysis of constraints we need to separate space and time , performing a 3 + 1 decomposition .
we introduce the time vector flow @xmath34 decomposed as t^=nn^+n^ , where @xmath35 is the unit normal vector to the @xmath36 const .
hypersurface , @xmath37 the lapse function and @xmath38 the shift vector orthogonal to the normal vector .
the constant time hypersurface is then characterised by the following three quantities : @xmath39 where @xmath40 is the projection tensor on the hypersurface and @xmath41 the associated extrinsic curvature [ ext - def ] k_= ( _ -d _ ( n _ ) ) . with `` dot '' we mean the lie derivative respect to @xmath42
, @xmath43 is the 3d covariant derivative on the constant time hypersurface and the parenthesis @xmath44 on the indices denote symmetrisation .
following @xcite , we decompose @xmath13 into the normal and transverse components with respect to the aforementioned hypersurface : @xmath45 the expression for the covariant derivative of @xmath13 can be decomposed into various pieces depending on the derivatives of its components and of the metric : [ dmuanuc ] _ a _ = d _ a_-a _ * k_+n _ ( ( k _ ) ^-d_)a _ * ) + n_n_(v_*- _ a^ ) , where @xmath46 is the acceleration vector . in equation ( [ dmuanuc ] ) , as well as for the whole ( beyond )
horndeski lagrangian , time derivatives appear only for the three dimensional metric @xmath47 ( inside the extrinsic curvature ) and for the component @xmath48 ( inside what we called @xmath49 ) .
@xmath49 plays for @xmath48 the same role that @xmath50 plays for @xmath47 , i.e. v _ * n^ _ a _ * = ( _ * -n^d_a _ * ) .
it is convenient therefore to work directly with the extrinsic curvature and @xmath49 , instead of the real velocities @xmath51 and @xmath52 , identifying those terms as the kinetic contributions to the action .
this allows us to treat the fields in a decomposed space - time while still remaining in a covariant form .
remember however that @xmath49 contains the second time derivative of the scalar field , hence it represents a potentially dangerous contribution that could lead to the propagation of the ostrogradsky mode .
a further advantage of this procedure is that the lagrangian densities do not depend explicitly on the lapse and shift functions .
this is because such quantities are implicitly included in @xmath53 and @xmath54 .
this is a huge simplification that considerably reduces the number of fields involved in the calculation . performing a standard
adm canonical analysis would be very complicated , as was already shown in @xcite for the case of quartic beyond horndeski only .
on the other hand , ( beyond ) horndeski lagrangians are diffeomorphism invariant , so intuitively we do not expect any modification to the algebra of constraints associated with the lapse and shift , as they are the generators of such a symmetry . however , this is not straightforward to show and a general proof is still missing .
steps forward in this direction have been made very recently in @xcite , where for a simple quartic lagrangian , this was indeed shown to be the case .
for the purposes of this paper , it is sufficient to retain in the lagrangian only the highest order terms in the extrinsic curvature , so we obtain the following expressions ( we adopt the notation used in @xcite ) : @xmath55 in the following , for simplicity , we also assume that the functions @xmath56 and @xmath20 depend only on @xmath6 , and not on @xmath5 . for horndeski
, we obtain for the quantities @xmath57 and @xmath58 the following expressions @xmath59 } + 2 g_{4 x } \left ( a^2 h^{\alpha}_{[\mu } h^{\beta}_{\nu ] } - \hat{a}^2 \hat{p}^{\alpha}_{[\mu } \hat{p}^{\beta}_{\nu ] } \right ) \label{k - h4}\ , , \\ { } ^h { \cal k}_{\mu \nu
\rho}^{\alpha \beta \gamma } & = & \frac13 g_{5 x } a _ * \left ( a^2 h^{\alpha}_{[\mu } h^{\beta}_{\nu } h^{{\gamma}}_{\rho ] } - \hat{a}^2 \hat{p}^{\alpha}_{[\mu } \hat{p}^{\beta}_{\nu } \hat{p}^{{\gamma}}_{\rho ] } \right ) \label{k - h5}\ , , \ ] ] where @xmath60 is the three - dimensional projection tensor defined in terms of @xmath61 . note that there are no terms containing @xmath49 , since the @xmath62 quantities of equation vanish .
this ensures that the equations of motion are second order and that the horndeski lagrangian only propagates at most three degrees of freedom .
this is achieved by the special horndeski tuning between the ( non - minimal ) coupling of gravity to the derivatives of the scalar field . while the expressions associated with the quartic horndeski have already been derived in @xcite ,
this is the first time these expressions have been found for quintic horndeski .
the relevant kinetic terms for beyond horndeski are @xmath63 } \
, , \label{b - bh4 - 3d } \\ { } ^{bh } { \cal k}_{\mu \nu}^{\alpha \beta } & = & - f_4 \left [ a^4 h^{\alpha}_{[\mu } h^{\beta}_{\nu ] } - a^2 \hat{a}^2 \hat{p}^{\alpha}_{[\mu } \hat{p}^{\beta}_{\nu ] } + 2\hat{a}^4 \hat{p}^{\alpha}_{\mu } \hat{p}^{\beta}_{\nu } - \hat{a}^4 \left ( \hat{p}^{\alpha}_{\mu } h^\beta_\nu + \hat{p}^{{\beta}}_{\nu } h^{\alpha}_\mu \right ) \right ] \ , , \label{k - bh4 - 3d } \\ { } ^{bh } { \cal k}_{\mu \nu \rho}^{\alpha \beta \gamma } & = & f_5 a _ * \left[a^4 h^{\alpha}_{[\mu } h^{\beta}_{\nu } h^{{\gamma}}_{\rho ] } - a^2\hat{a}^2 \hat{p}^{\alpha}_{[\mu } \hat{p}^{\beta}_{\nu } \hat{p}^{{\gamma}}_{\rho ] } + \hat{a}^4 \left ( \hat{p}^{\alpha}_{[\mu } \hat{p}^{\beta}_{\nu ] } \left ( \hat{p}^{{\gamma}}_{\rho } - h^{\gamma}_\rho \right ) + \text{sym . } \right)\right ] \ , , \label{k - bh5 - 3d } \end{aligned}\ ] ] where `` @xmath64 '' in ( [ k - bh5 - 3d ] ) stands for the symmetric permutation of doublets of vertical indices ( e.g. the last term in eq .
( [ k - bh4 - 3d ] ) ) . since the mixing term only appears in beyond horndeski theory , we will mostly omit the superscript `` bh '' from @xmath65 and @xmath66 . for beyond horndeski the quantities in eq .
do not vanish , therefore there are potentially dangerous mixings between @xmath49 ( containing the second derivative of the scalar field ) and @xmath67 ( containing the first derivative of the spatial metric ) .
such contributions to the action lead to higher order equations of motion .
this suggests , but not necessarily implies , the presence of additional propagating degrees of freedom .
we now study the existence of a primary constraint that could prevent the propagation of an additional ( ghost ) mode .
the natural tool for counting the number of degrees of freedom is the ( dirac ) canonical analysis of constraints . a complete analysis
, however , is beyond the scope of this work and here we concentrate on studying the existence of primary constraints .
in particular we focus on the constraint that is able to remove the ostrogradsky mode , and assume that the four first class constraints associated with the diff .
invariance still remain .
of course , the existence of a primary constraint is not enough to remove a physical dof , nevertheless this would be the first necessary condition for it .
moreover , to our knowledge , there are no known lorentz invariant theories that propagate half degrees of freedom : this would be the case if there are an odd number of second class constraints .
a complete analysis for the quartic lagrangian has been recently performed in @xcite , confirming that a secondary constraint does indeed exist .
primary constraints exist when , passing to the hamiltonian formalism , all the velocities _ can not _ be expressed in terms of the fields and their conjugate momenta .
this translates to relations ( constraints ) between the fields and momenta that need to be added to the canonical hamiltonian through lagrangian multipliers . as explained in section [ sec - kin ] , instead of working with the true velocities , we work with closely related quantities and therefore define the conjugate momenta accordingly : _ * , ^ _ .
notice that this definition also differs from the usual one due to the presence of the factor @xmath68 ; this helps to completely remove the lapse from the relations . in horndeski theory , the primary constraint needed to remove the ostrogradsky ghost is automatically enforced by its construction , and it is @xmath69 . ''
means weak equality , i.e. equality on the phase space determined by constraints . ]
using the expression for beyond horndeski theories given in eqs .
( [ bh4-m ] ) and ( [ bh5-m ] ) , together with the expression of @xmath70 given in eq .
( [ dmuanuc ] ) , the conjugate momenta are obtained as @xmath71 using the properties of the matrix @xmath32 @xmath72 which stem from the fact that @xmath32 is constructed from the projection tensor @xmath73 and @xmath74 , we can derive the following identities @xmath75 hence , we can find a primary constraint of the form @xmath76 with the formulation of beyond horndeski theories given in section [ secbht ] , it is therefore very easy to show the existence of the primary constraint necessary to remove the ostrogradsky ghost .
it is important to notice that the constraint ( [ primary - bh ] ) is a linear combination of the conjugate momenta @xmath77 and @xmath78 . in this case
it is possible to remove @xmath49 by a suitable field redefinition and the system can be recast as a second order one . however there is no guarantee that the new system will be lorentz invariant . in order to better understand the properties of the conjugate momenta ( [ mom1full ] ) , ( [ mom2full ] ) and for the sake of the next subsection
, we now explain the above result focusing only on the highest order terms in the extrinsic curvature as given in eq .
( [ lagkin4 ] ) and ( [ lagkin ] ) .
the conjugate momenta simplify to @xmath79 from eqs .
( [ mom1 ] ) , ( [ mom2 ] ) it becomes easy to see that the key property of these momenta is that @xmath49 appears only in @xmath78 .
this implies that to build a constraint we need to eliminate @xmath49 taking a suitable linear combination of components of @xmath78 . using the properties of the three dimensional projection tensor , i.e. @xmath80 } \hat{a}_{\alpha } \hat{a}^{\mu}=\hat{a}^2 \hat{p}^{\beta}_{\nu}\ , , \qquad h^{\alpha}_{[\mu } h^{\beta}_{\nu } h^{{\gamma}}_{\rho]}\hat{a}_{\alpha } \hat{a}^{\mu}= \hat{a}^2 \hat{p}^{\beta}_{[\nu } \hat{p}^{{\gamma}}_{\rho ] } \,,\ ] ]
we can show the following relations for @xmath62 and @xmath58 in beyond horndeski theories @xmath81 these results imply that we can eliminate @xmath49 by contracting @xmath78 with @xmath82 @xmath83 then , it is straightforward to verify the primary constraint ( [ primary - bh ] ) . to conclude ,
let us give the redefinition of the extrinsic curvature that eliminate the cross term between @xmath49 and the extrinsic curvature @xmath84 we discuss in appendix [ app1 ] the cases when this redefinition is in fact a disformal transformation of the extrinsic curvature .
we now consider what happens to the primary constraint found in beyond horndeski , when we combine this theory with the horndeski one .
we will show that , when mixing actions of different orders , generically the primary constraint of the kind ( [ primary - bh ] ) is lost . in this regard ,
it is enough to consider only the highest order terms in the extrinsic curvature .
@xmath62 and @xmath58 in horndeski theories obey the following relations @xmath86 using these relations , the conjugate momenta in the presence of horndeski contributions become @xmath87 \hat{a}^{\mu } \hat{a}_{\alpha } \pi^{\alpha}_{\mu } & = & a _ * \left ( 2 \hat{a}^2 - a_*^2 \right ) \left ( 2 { \cal b}^{\alpha}_{\mu } k^{\mu}_{\alpha } + 3 { \cal b}^{\alpha \beta}_{\mu \nu } k^{\mu}_{\alpha } k^{\nu}_{\beta } \right ) \nonumber \\[1ex ] & & + \ , \frac{2 \left(2 g_{4x }
a^2 -g_4 \right)}{f_4 a_*}\ , { \cal b}^{{\alpha}}_{{\mu } } k^{{\mu}}_{{\alpha } } - \frac{g_{5x } a^2 } { f_5 a _ * } \ , { \cal b}^{\alpha \beta}_{\mu \nu } k^{\mu}_{\alpha } k^{\nu}_{\beta } \ , . \label{degmix1}\end{aligned}\ ] ] let us first consider the case in which we do not mix different orders . in these cases the primary constraint still persists . * quartic beyond horndeski @xmath85 quartic horndeski @xmath88 \pi _ * - \hat{a}^{\mu } \hat{a}_{\alpha}\ , \pi^{\alpha}_{\mu } \,\approx\ , 0 \,.\ ] ] * quintic beyond horndeski @xmath85 quintic horndeski @xmath89 \pi _ * - \hat{a}^{\mu } \hat{a}_{\alpha}\ , \pi^{\alpha}_{\mu } \,\approx\ , 0 \,.\ ] ] in the above cases , the redefinition of the extrinsic curvature that removes the mixing term is given respectively by k^ _
& = & |k^ _ - _ ^. [ hbh4 ] + k^ _ & = & |k^ _ - _ ^. [ hbh5 ] so far , so good .
problems arise when we combine together different orders .
this was already realised in @xcite for the case of quintic beyond horndeski @xmath85 quartic horndeski . here
we also provide the tools to analyse what happens joining together quartic beyond horndeski @xmath85 quintic horndeski , a system that is more technically challenging to deal with . in both of these cases
the primary constraint is lost .
indeed , the presence of horndeski terms of different order , obstructs any _ linear _ combination of momenta ( [ piamix1 ] ) and ( [ degmix1 ] ) that compensates and sets to zero the coefficients of both @xmath90 and @xmath91 .. for this particular value the quartic horndeski contribution vanishes .
( we will meet this special case also in the next section ) . ] of course the absence of the primary constraint which is present in isolated beyond horndeski theories , does not necessarily mean that a different constraint could not arise . surely it can not come from a linear combination of momenta is the only linear combination of components of @xmath78 that does not contain @xmath49 . ] , so in order to check its real absence , one should be able to prove the invertibility of the non - linear system of matrices equations ( [ piamix1])([degmix1 ] ) , which is not a simple task . to summarise , beyond horndeski theories do have the primary constraint needed to avoid the ostrogradsky instability , however this same constraint is spoiled when they are combined with contributions from different orders of horndeski .
in the previous section we found that , due to the linear nature of the primary constraint in beyond horndeski theories , a suitable redefinition of the extrinsic curvature can recast the theory into a manifestly second order system . in this section
( see also appendix [ app1 ] ) we study the relation between this field redefinition and a disformal transformation of the metric .
the ( generalised ) disformal transformation @xcite is given by @xmath92 in @xcite it was shown that horndeski theory is mapped into itself by a disformal transformation if @xmath93 is a function of the scalar field only and that the @xmath6 dependence of @xmath93 generates beyond horndeski terms . in the following
, we do not choose any gauge for our analysis ; for simplicity , we only consider the case where @xmath93 depends strictly on @xmath6 .
we can show that by a disformal transformation , the horndeski action is mapped to a combination of horndeski and beyond horndeski actions of the same order @xmath94 & = & { } ^{h}{\cal l}_4[g_4 ] + { } ^{bh}{\cal l}_4[d_4 ] \
, , \label{dish4 } \\ { } ^{h}{\cal \bar{l}}_5[\bar{g}_5 ] & = & { } ^{h}{\cal l}_5[g_5 ] + { } ^{bh}{\cal l}_5[d_5 ] \ , ; \label{dish5}\end{aligned}\ ] ] where @xmath95 g_5 & = & \frac{\bar{g}_5}{(1 + x \gamma)^{1/2 } } + \int \frac{\bar{g}_5 \left(\gamma + x \gamma_x \right)}{2(1 + x \gamma)^{3/2}}\ , dx \ , , \qquad d_5= \frac{x \bar{g}_{5\bar{x } } \gamma_x}{3(1 + x \gamma)^{5/2 } } \ , .
\label{g5d5}\end{aligned}\ ] ] on the other hand , beyond horndeski is mapped into itself only , with a different overall function @xmath96 & = & { } ^{bh}{\cal l}_4[f_4 ] \
, , \\ { } ^{bh}{\cal \bar{l}}_5[\bar{f}_5 ] & = & { } ^{bh}{\cal l}_5[f_5 ] \,;\end{aligned}\ ] ] where @xmath97 from these results , we can easily see that the combination of horndeski and beyond horndeski of the same order can be mapped to horndeski . indeed @xmath98
+ { } ^{bh}{\bar{\cal l}}_4 [ \bar{f}_4 ] & = & { } ^{h}{\cal l}_4[g_4 ] \ , , \label{dishbh4 } \\ { } ^{h}{\bar{\cal l}}_5[\bar{g}_5 ] + { } ^{bh}{\bar{\cal l}}_5 [ \bar{f}_5 ] & = & { } ^{h}{\cal l}_5[g_5 ] \ , , \label{dishbh5}\end{aligned}\ ] ] for a disformal transformation @xmath99 which satisfies @xcite @xmath100 \gamma_{5x } & = & \frac{3\,f_5}{3\,x^2 f_5 - x g_{5x}}\ , , \label{dis5}\end{aligned}\ ] ] where @xmath101 are respectively given in equations ( [ g4d4 ] ) , ( [ g5d5 ] ) and @xmath102 are given in ( [ f4f5 ] )
. thus , the absence of the ostrogradsky mode is ensured by the fact that these combinations can be mapped to horndeski . in the absence of a coupling with matter ,
the theory is just horndeski .
it is possible to check that , for ( [ dis4])([dis5 ] ) , the disformal transformation reproduces the redefinition of the extrinsic curvature that removes its mixing with @xmath49 , i.e. eqs . ( [ hbh4 ] ) , ( [ hbh5 ] ) . on the other hand ,
if there is _ only _ beyond horndeski , the disformal transformation can not map the theory to horndeski .
in fact in this limit @xmath103 and the disformal transformation becomes @xmath104 this is the projection tensor @xmath105 which satisfies @xmath106 , implying that it has a null eigenvalue and therefore is not invertible .
thus the disformal transformation required to map a combination of horndeski and beyond horndeski to horndeski , given by eqs .
( [ dis4])([dis5 ] ) , becomes singular in the limit @xmath107 .
that makes the transformation singular , i.e. @xmath108 .
this clarifies that the spacial case encountered in the former section , is actually related to an ill - defined disformal transformation .
] this is consistent with the fact that beyond horndeski is mapped into itself by the generalised disformal transformation .
surprisingly , as we showed in the previous section , there still exists a redefinition of the extrinsic curvature that removes the mixing ( dangerous ) term , eq .
( [ diag4 ] ) .
however , it can be shown ( see appendix [ app1 ] ) that this transformation is not a disformal transformation of the extrinsic curvature . accordingly to the generality of the disformal transformation @xcite , this implies that , when recast into a second order theory , beyond horndeski is different from horndeski .
it follows naturally therefore that beyond horndeski is equivalent ( up to a field redefinition ) to a lorentz breaking second order theory .
an important remark is that transforms as @xmath109 .
] , for the particular cases @xmath110 and @xmath111 , the generalised disformal transformation with an arbitrary function @xmath99 is a _ symmetry _ of the theories of beyond horndeski
. it would be interesting to investigate whether the existence of this symmetry is related to a first class constraint , i.e. if it is a gauge symmetry .
in this paper we examined the number of degrees of freedom present in the theory of beyond horndeski , in a gauge - invariant way .
we first rewrote beyond horndeski in terms of the extrinsic curvature of the constant scalar field hypersurface ; then we derived the relevant kinetic terms for both beyond horndeski and horndeski theories .
unlike horndeski , in beyond horndeski there is a mixing term between the second time derivative of the scalar field and the extrinsic curvature , which could give rise to the propagation of an additional ( ghost ) degree of freedom in absence of proper constraints .
we found that in beyond horndeski there exists a primary constraint that is able to remove the ostrogradsky instability .
moreover this constraint occurs in the form of a linear combination of conjugate momenta .
this allowed us to identify the transformation of the extrinsic curvature that eliminates the aforementioned mixing term .
this primary constraint still persists when we join together theories of beyond horndeski and horndeski of the same order . however , when mixing theories of different orders , the very same constraint is lost , as shown in @xcite for the case of quintic beyond horndeski @xmath85 quartic horndeski .
here we were able to show that this happens also in the other case , i.e. quartic beyond horndeski @xmath85 quintic horndeski .
nevertheless , the loss of that constraint does not exclude the possibility that , in these cases , a different ( non - linear ) primary constraint could still arise .
the study of this possibility however is well beyond the scope of this work .
notice that in ref .
@xcite a non - covariant method for removing the higher order time derivatives from the field equations was found .
this is valid even for mixed order beyond horndeski and horndeski combinations which would seem to contradict our results .
however , the fact that a higher order theory can be recast in a second order one in a non - covariant way , is not necessarily connected with the number of propagating dof .
nothing forbids that one of the ( second order ) modes is indeed a ghost , like in the ostrogradsky approach .
the hamiltonian analysis confirms itself as the unique tool for the correct counting of dof .
to identify the origin of the primary constraint , we studied the transformation of horndeski and beyond horndeski under a generalised disformal transformation which depends on the scalar field kinetic term , without fixing any gauge .
the results of this analysis are summarised in figure [ tab ] .
we confirmed that horndeski is disformally mapped to horndeski plus beyond horndeski of the same order @xcite . on the other hand , beyond horndeski
is mapped to itself by the generalised disformal transformation .
thus , in the absence of coupling with matter , the mixture of horndeski and beyond horndeski of the same order is nothing but horndeski itself .
we checked that the redefinition of the extrinsic curvature that removes the mixing term can be derived from the disformal transformation only when there are both horndeski and beyond horndeski together . on the other hand , the disformal transformation that maps horndeski plus beyond horndeski to horndeski is singular precisely in the limit that would leave the theory of beyond horndeski alone .
this is consistent with the fact that beyond horndeski is mapped to beyond horndeski itself . in this case
the transformation to remove the mixing term can not be derived from the disformal transformation .
this indicates that beyond horndeski is disformally disconnected from horndenski .
the fact that beyond horndeski can not be mapped into horndeski , makes this theory a covariant scalar - tensor theory that can be recast into a second order system which is not horndeski , therefore it should be lorentz breaking .
it is also interesting that the primary constraint is preserved only when we include a horndeski term of the same order , but it is spoiled when a different order horndeski term is added .
in the first case however the theory is just horndeski up to a generalised disformal transformation
. it will be important to continue the canonical analysis , performing the dynamical evolution of the primary constraint .
for the quartic case it has been recently shown that it leads to a secondary second class constraint @xcite ; for the quintic case however such an analysis is still missing . moreover , it would also be interesting to study couplings with matter .
we will come back to these issues in a forthcoming work .
we would like to thank cedric deffayet , david langlois , gustavo niz and jeremy sakstein for useful discussions .
mh is supported by a u.k .
science and technology facilities council ( stfc ) research studentship .
kk is supported by the uk science and technology facilities council grants st / k00090/1 and the european research council grant through 646702 ( costesgrav ) .
in this appendix we examine the relation between the transformation of the extrinsic curvature that removes the mixing term between itself and @xmath49 ( eq . ( [ diag4 ] ) ) , and the transformation of the extrinsic curvature under a generalised disformal transformation of the metric ( [ disftransf ] ) . for simplicity , and to avoid very long formulae that are not illuminating , we only give the expression of the conformal piece and of the term proportional to @xmath49 .
it is indeed this last one that plays the crucial role . in order to compare this transformation with the one coming from the analysis of the primary constraint ( [ diag4 ] ) ,
we first need to slightly manipulate the latter .
indeed , the transformation given in section [ sec - prim ] is not unique , we can always multiply the new extrinsic curvature by a generic function of @xmath6 without changing its properties .
moreover , since the disformal transformation acts on all the fields , and not just on the extrinsic curvature , we also need to accordingly transform the other objects present in ( [ diag4 ] ) .
once this has been done , we obtain that the transformation coming from the primary constraint is indeed of the form ( [ ktransf ] ) , but with b = ( ) ^1/2 , c = - . [ cdiag ] clearly in ( [ cdiag ] ) there is an additional piece proportional to @xmath49 that it is not present for the disformal transformation .
this is the key term .
when we add together beyond horndeski @xmath85 horndeski of the same order , using ( [ dis4])([dis5 ] ) in the analogous equation of ( [ cdiag ] ) coming from ( [ hbh4])([hbh5 ] ) , this extra piece vanishes and the transformation reduces to ( [ cdisf ] ) . on the other hand with beyond horndeski alone this extra term survives making the two transformations different .
we can therefore conclude that , since the transformation that is able to recast beyond horndeski into a second order theory is not a disformal transformation , beyond horndeski is equivalent to a lorentz breaking second order theory .
notice that the singular limit encountered in section [ sec - disf ] , i.e. @xmath103 , manifests also here : indeed for this value the extra piece does not cancel out as one would expect . | determining the most general , consistent scalar tensor theory of gravity is important for building models of inflation and dark energy . in this work
we investigate the number of degrees of freedom present in the theory of beyond horndeski .
we discuss how to construct the theory from the extrinsic curvature of the constant scalar field hypersurface , and find a simple expression for the action which guarantees the existence of the primary constraint necessary to avoid the ostrogradsky instability .
our analysis is completely gauge - invariant .
however we confirm that , mixing together beyond horndeski with a different order of horndeski , obstructs the construction of this primary constraint . instead ,
when the mixing is between actions of the same order , the theory can be mapped to horndeski through a generalised disformal transformation .
this mapping however is impossible with beyond horndeski alone , since we find that the theory is invariant under such a transformation .
the picture that emerges is that beyond horndeski is a healthy but isolated theory : combined with horndeski , it either becomes horndeski , or likely propagates a ghost
. 0.5 cm |
class of all subspaces of @xmath0 is denoted by @xmath1 and will be called a _ projective space_. the projective space @xmath1 is endowed with a subspace distance . given two subspaces @xmath4 and @xmath5 , @xmath6 the metric arises as a distance function in the field of error correction in random network coding . in this context , a subset of the projective space is called a _
code _ and the elements that belong to a code are called _
codewords_. koetter and kschischang showed that one can view transmission of messages over networks as transmission of elements of a code in a projective space@xcite . in their formulation , the amount of ` error ' between a transmitted subspace and a received subspace is measured by their subspace distance .
they showed that one can correct a fixed amount of errors in the network given that all the subspaces in the code are pairwise sufficiently far apart .
this approach mimicked the theory of classical error correcting codes . and
therefore the theory of error correction in random network coding draws some motivations from the field of classical error correcting codes .
classical error correcting codes are constructed over @xmath7 . here
a ` code ' is an arbitrary subset of @xmath7 and the ` codewords ' are @xmath8 length vectors whose co - ordinates are elements of @xmath9 . the distance , called the _ hamming distance _ , between two codewords is measured by the number of differing co - ordinates between the two vectors .
the space of @xmath7 which is equipped with a hamming distance is called the _ hamming space_. most of the theory on classical error correcting codes focus on a special class of codes called _
linear codes_. linear codes are just subspaces of @xmath7 .
a subspace , by definition , is an abelian group under vector addition . and
it turns out that the hamming distance is translation invariant with respect to this addition .
the geometry of the hamming distance combines with group structure to give a rich theory of classical binary error correcting codes @xcite .
however there are a few differences between error correction coding over projective spaces and hamming spaces .
if the volume of a set is defined as the number of elements in the set whenever the set is finite , one can study how the volume of a ball changes as the centre is translated .
in euclidean spaces and hamming spaces , the volume of a ball is independent of its centre .
but the volume of a ball is not independent of its centre in a projective space .
a technical way of restating this observation is that the geometry induced by the subspace distance @xmath10 on @xmath1 is not distance - regular ( unlike the hamming space ) and one can not mimic the tricks of sphere - covering / packing bounds for the hamming space @xcite . and
therefore the usual geometric intuitions fail in projective spaces . however , it turns out that restricting the projective spaces to a class of subspaces with a particular dimension gets this distance regular property .
therefore it becomes interesting to consider the set of all subspaces of particular dimension .
we shall use the notation @xmath11 to denote the set of all elements of @xmath1 with dimension @xmath12.the authors of @xcite lay the groundwork , derives bounds and construct a few codes on @xmath11 .
these codes are called _ constant dimension codes _ and recently there was a lot of activity in this area @xcite , @xcite , @xcite . on the other hand ,
our focus is on non - constant dimension codes .
even though the distance - regular property fails , one can recover a form of gilbert - varshamov bound for non - constant dimension codes @xcite and derive a form of singleton bound for the projective space @xcite .
the concepts of ` linearity ' and ` complements ' in projective spaces are examined in @xcite .
the definitions are motivated from their binary error correction coding counterpart .
the authors discuss different possibilities of complements and linearity and they state a few conjectures .
this paper is mainly focused on resolving one conjecture concerning the linearity . a code in @xmath1 which imitates the group structure of a binary vector space and has the translation invariant property on the subspace distance
is called a _ linear code in the projective space _ in @xcite .
henceforth , we will just call it a linear code .
one can construct such a linear code with @xmath3 codewords by identifying a basis as the generators of the group . a linear code constructed in this fashion
is called _ a code derived from a fixed basis_. it follows from @xcite that @xmath1 can not be a linear code .
so a maximal linear code is somewhere between these two extremes .
the main conjecture claims that the largest possible cardinality of a maximal linear code must be * equal * to the cardinality of a code derived from a fixed basis .
in other words , it states that the largest linear code in a projective space has a cardinality of @xmath3 . as a first step towards proving this result ,
a special case of a linear code over @xmath1 that contains @xmath0 is considered .
it is conjectured that a linear code that contains @xmath0 as a codeword can contain a maximum of @xmath13 @xmath12-dimensional codewords @xcite . in this paper , we prove this conjecture .
we further show that maximal linear codes that contains @xmath0 as a codeword must be derived from a fixed basis .
it is important to note that if we drop the condition of ` @xmath0 is a codeword ' , one can construct linear codes that are not derived from a fixed basis ( see section @xmath14 , example @xmath15 in @xcite ) .
the organization of our paper is as follows : in section [ overview ] we present the conjecture formally and state all the important definitions relating to the conjecture .
the proof of the conjecture and a few additional results are presented in section [ proofs ] .
we conclude the paper with general remarks about the conjecture and a few open problems in section [ conclusion ]
. _ notations _ : @xmath0 represents the @xmath8-dimensional vector space over the finite field @xmath16 . the class of all subspaces of @xmath0 is denoted by @xmath1
@xmath11 represents the set of all elements of @xmath1 with dimension @xmath12 .
the @xmath17 for a subset @xmath18 of @xmath0 stands for the linear span of the elements of @xmath18 .
occasionally , we identify the one - dimensional space @xmath19 with the vector @xmath20 . the trivial space and the null element of the vector space
are both represented by @xmath21 and they are meant to be distinguished from context . given two sets @xmath4 and @xmath5 , @xmath22 denotes the symmetric difference of @xmath4 and @xmath5 .
@xmath23 represents the direct sum of two disjoint subspaces @xmath24 and @xmath25 .
based on motivations to mimic the classical linear code structure on @xmath7 endowed with the hamming distance , a linear code on @xmath1 is defined in @xcite as follows : a code @xmath26 is a linear code if 1 .
there exists a binary operation @xmath27 on @xmath28 , such that @xmath29 is an abelian group .
2 . the additive identity of @xmath29 is the trivial space @xmath21 .
3 . for every @xmath30 , we have @xmath31 .
@xmath32 for all @xmath33 . a couple of remarks on the definition of a linear code in @xmath1 : the first three conditions of the definition allow us to view a linear code @xmath34 as a subspace of a vector space of characteristic @xmath35 .
the third property states that every element is an _
idempotent_. the last property in the definition is called the _ translation invariant property_. in words , this property states that the subspace distance between @xmath4 and @xmath5 does not change if both were ` translated ' by the same codeword @xmath36 .
the following proposition is an immediate corollary of the definition of a linear code .
[ dimensionofsum ] if @xmath4 and @xmath5 are two codewords of a linear code @xmath37 , then @xmath38 given two codewords @xmath4 and @xmath5 , using translation invariance and idempotence , we get : @xmath39 a linear code can be defined on every projective space @xmath1 as described in @xcite .
pick a basis @xmath40 for @xmath0 . define the code @xmath41 as follows : @xmath42 we can verify the simpler group properties . given a set @xmath43 ,
define @xmath44 . given two elements @xmath45 and
@xmath46 from @xmath41 , define @xmath47 the trivial space @xmath21 is spanned by the empty set . and
since the symmetric difference of a set with itself is empty , the @xmath27 operation is an involution .
the associativity and the translation invariance can be verified ( see @xcite ) .
the code @xmath41 has @xmath3 codewords ; one for each subset of @xmath48 . a linear code of this form will be called ` a code derived from a fixed basis ' .
clearly any code constructed like this is generated from the basis of @xmath8 one dimensional spaces .
however , we will see in theorem [ one - dim space ] that _ merely _ containing @xmath8 one dimensional spaces is sufficient to guarantee that the code is of the above type .
we record our formal definition for ` a code derived from a fixed basis ' below .
a linear code @xmath49 in @xmath1 is _ derived from a fixed basis _ if the group @xmath49 is generated by @xmath8 one - dimensional spaces in @xmath49 . as it will be seen in corollary [ fullspacenotlinear ]
, one can not define a group operation to make the entire space @xmath1 a linear code .
a natural question , therefore , is whether linear codes with cardinality larger than @xmath3 exist .
the authors in @xcite claim that it does not .
a special case of the conjecture asks if this conjecture can be proved for linear codes that contain @xmath0 .
the actual conjecture just asks this for the case of @xmath50 .
we record the conjecture formally as follows : if a linear code @xmath49 over @xmath1 contains @xmath0 as a codeword , then @xmath51 .
specifically @xmath52 it should be noted that the conjecture and our proof works for any field ! in other words , the largest abelian group on the subspaces of @xmath53 that contains the full space ( with the translation invariant property of the subspace metric ) can have at most @xmath3 elements in the group . in the particular case , when @xmath54 is infinite , there are infinitely many subspaces and yet our conditions on the group force its size to be at most @xmath3 .
we prove this conjecture in the next section .
we also prove that the equality @xmath55 holds if and only if @xmath49 is derived from a fixed basis .
in this section , we will formally prove the conjecture through a series of smaller results .
the following two lemmas are also proved in @xcite . and we record it here , because we will use them in the proof of the main theorem . if @xmath56 belong to a linear code @xmath49 and @xmath57 , then @xmath58 and @xmath59 .
[ cyclic sum ] this lemma follows from the property @xmath60 for all @xmath61 .
@xmath62 one can similarly prove @xmath63 the following lemma proves that adding two disjoint subspaces in the group gives the direct sum of the two subspaces .
if @xmath64 belong to a linear code @xmath49 and @xmath65 , then @xmath66 [ disjoint sum ] suppose @xmath67 , then from proposition [ dimensionofsum ] and @xmath68 , we have : @xmath69 from lemma [ cyclic sum ] , @xmath70 and from proposition [ dimensionofsum ] , @xmath71 therefore , @xmath72 and thus @xmath73 .
similarly we can show that @xmath74 .
thus @xmath75 which along with @xmath76 proves that @xmath77 .
the next lemma shows that given a codeword that is a subset of another codeword , the addition of these two codewords ` splits ' the larger codeword . in other words ,
the larger codeword can be generated from its smaller components .
if @xmath64 belong to a linear code @xmath49 and @xmath78 , then there exists a @xmath79 such that @xmath80 [ splitting ] let @xmath81 .
the dimension of @xmath36 can be calculated : @xmath82 from lemma [ cyclic sum ] , @xmath70 .
the above calculation of @xmath83 gives , @xmath84 therefore @xmath85 and thus @xmath86 . from lemma
[ disjoint sum ] , @xmath87 lemma [ splitting ] can be used to decompose larger dimensional codewords using the one - dimensional subspaces contained in the larger subspace . in the following lemma
, we will first show that the one - dimensional subspaces in a linear code must be linearly independent .
linear dependence of one dimensional spaces will contradict the uniqueness of a inverse . in the second part of the lemma
, we will show that the subspaces spanned by one dimensional codewords belong to the code .
let @xmath88 denote the set of all one dimensional spaces in @xmath49 .
assume @xmath89 and let @xmath90 represent distinct one dimensional spaces spanned by @xmath91 respectively for @xmath92 .
when we say the subspaces @xmath93s are linearly independent , it means that the corresponding @xmath94s are linearly independent
. let @xmath88 denote the set of all one dimensional codewords in @xmath49 .
then , 1 . @xmath88 is linearly independent .
@xmath95 [ subsetgensubspace ] proof of @xmath96 : the proof is by contradiction . without loss of generality , let us assume that @xmath97 is a minimally dependent set of vectors . since @xmath98s are distinct one dimensional spaces , we can assume @xmath99 .
by minimality of dependence , it follows that @xmath100 from lemma [ disjoint sum ] , we see that @xmath101 but by cancellation of @xmath102 on both sides , we get @xmath103 which contradicts the distinctness of the one dimensional spaces . therefore @xmath88 is linearly independent .
proof of @xmath104 : let @xmath105 .
we prove the claim by induction on @xmath106 .
the case of @xmath107 is easily verified .
partition @xmath18 as @xmath108 , then @xmath109 by induction hypothesis . from part @xmath96 of this proposition , it follows that @xmath110 is linearly independent of @xmath111 and therefore @xmath112 . applying lemma [ disjoint sum ]
, we get @xmath113 by closure property of @xmath49 , @xmath114 we shall now show that the group operation for every code derived from a fixed basis is the same .
we will specifically show that if @xmath24 and @xmath25 are two code words from such a code , then there exists a basis @xmath115 , such that @xmath116 is actually the span of the symmetric difference of some two subsets of @xmath88 .
let @xmath49 be a code derived from a fixed basis , and @xmath88 be the set of one dimensional codewords that generate @xmath28 .
then for any @xmath117 , there exists @xmath118 such that @xmath119 define @xmath120 since @xmath121 , we have @xmath122 .
we claim that @xmath123 . if not , @xmath124 .
from lemma [ splitting ] , @xmath125 and @xmath126 but @xmath115 generates @xmath49 and therefore there exists a one dimensional space @xmath127 that belongs to @xmath128 as well .
so @xmath129 belongs to @xmath130 but not to @xmath4 , contradicting @xmath131 .
thus we have @xmath123 and similarly , we have @xmath132 .
note that from lemma [ disjoint sum ] , the group addition of all the linearly independent one dimensional spaces belonging to @xmath4 is simply the span of @xmath4 . in the following expressions , the @xmath133 is with respect to the group operation .
therefore @xmath134 using the fact @xmath135 , we get : @xmath136 again using lemma [ disjoint sum ] and the fact that @xmath137 contain one - dimensional subspaces , @xmath138 we can choose a basis @xmath139 and consider the linear code as subspaces spanned by the subsets of @xmath139 where the group operation is defined by span of the symmetric difference of the generating sets .
the above proposition shows that all ` codes derived from a fixed basis ' are of this form . we will now prove a simple sufficient condition for the code to be derived from a fixed basis .
a consequence of proposition [ subsetgensubspace ] is that the subgroup generated by the one - dimensional codewords in a linear code is the collection of subspaces spanned by subsets of the one - dimensional spaces in @xmath28 .
however , the following theorem proves that the subgroup is the entire group if the subgroup _ merely contains _ @xmath8 one dimensional spaces .
alternatively , the theorem proves that if a code has @xmath8 one dimensional spaces , then the code is entirely generated from it .
[ one - dim space ] if a linear code @xmath49 in @xmath1 contains @xmath8 one - dimensional spaces then @xmath49 is derived from a fixed basis .
if @xmath28 contains @xmath8 one - dimensional spaces from @xmath1 , then it follows , from the first part of proposition [ subsetgensubspace ] , that the one dimensional spaces form a basis @xmath139 for @xmath0 .
we have to prove that these @xmath8 one dimensional spaces generate the code .
let @xmath18 be the set of all subspaces in @xmath28 that are spanned by a subset of @xmath139 .
we want to prove that @xmath140 .
the proof is by contradiction .
suppose @xmath141 is non - empty .
pick a minimal subspace @xmath142 in @xmath141 .
consider the following set : @xmath143 due to minimality and lemma [ splitting ] , @xmath142 does not contain any one - dimensional space in @xmath18 and therefore @xmath144 is non - empty .
let @xmath145 be a maximal subspace of @xmath144 .
we claim that @xmath146 we prove this claim by contradiction .
suppose @xmath147 , then there exists an element @xmath148 that does not belong to @xmath149 .
this means that @xmath150 and since @xmath145 is a maximal subspace that belongs to @xmath151 , @xmath152 . but
@xmath153 since @xmath154 and @xmath148 .
therefore we conclude @xmath155 . since @xmath154
, it must be spanned by a subset @xmath156 .
let @xmath157 , then @xmath158 we note that @xmath159 is non zero since @xmath160 . rearranging the equation to express @xmath161 , we get @xmath162 but this contradicts @xmath163 therefore @xmath164 by lemma [ disjoint sum ] , @xmath165
. however @xmath166 also satisfies @xmath167 .
cancellation laws hold for @xmath27 , and thus we have @xmath168 which contradicts the fact that @xmath169 ( since @xmath170 ) . hence @xmath171 i.e. only subspaces that can be spanned by a subset of the chosen basis @xmath139 can belong to the code . since @xmath18 contains spans of subsets of @xmath139 ,
the second part of proposition [ subsetgensubspace ] ensures the reverse inclusion @xmath172 .
thus we have established @xmath140 .
theorem [ one - dim space ] will be used to show that the equality holds in the conjecture only when the code is derived from a fixed basis . in other words ,
the equality of the conjecture holds only for codes that are designed by fixing a basis , then spanning all subsets of the fixed basis ( where the codewords are the span of the subsets ) .
therefore , the symmetric diffence of generating sets is the group operation that achieves the maximum size of a linear code , given that the @xmath0 belongs to the linear code .
theorem [ one - dim space ] also shows that the entire projective space @xmath1 can not be a linear code .
the following corollary can also be deduced from older results @xcite .
there does not exist a group operation @xmath27 such that @xmath1 is a linear code .
[ fullspacenotlinear ] from theorem [ one - dim space ] , it follows that @xmath1 is a linear code derived from a fixed basis .
however , the first part of proposition [ subsetgensubspace ] does not allow a linear code to have more than @xmath8 one dimensional spaces .
therefore @xmath1 can not be a linear code .
now we will prove a few lemmas that will be useful to prove the main theorem . at the heart of the proof , we need a technical lemma due to lovasz @xcite given below .
* theorem 4.9 ) if @xmath173 are @xmath174-dimensional spaces and @xmath175 are @xmath176-dimensional spaces with the property @xmath177 , then @xmath178 [ lovasz ] consider @xmath179 for all @xmath180 . the next lemma show that @xmath181 and @xmath182 satisfy the conditions of lemma [ lovasz ] , namely the two classes of subspaces have the same number of elements and under some ordering they satisfy the intersection conditions . in the first part of the following lemma , we shall establish that @xmath183 .
we shall also show , in the second part , that @xmath181 and @xmath182 satisfy the intersection conditions specified in lemma [ lovasz ] .
define a map @xmath184 as @xmath185 .
we claim the following : 1 .
2 . if @xmath186 and @xmath187 then @xmath188
if and only if @xmath189 . [ conditions ] the map is left multiplication by @xmath0 and since @xmath49 is closed under @xmath27 , @xmath190 is well - defined .
proof of @xmath96 : we will show that @xmath191 .
@xmath192 thus @xmath190 is an involution and therefore bijective . to show @xmath183 we prove that @xmath193 . in other words
, we will show that a @xmath186 gets mapped to a subspace of dimension @xmath194 by the following calculation : @xmath195 therefore @xmath196 and similarly @xmath197 .
thus @xmath198 .
since @xmath190 is an involution , we know that @xmath199 and therefore it follows that @xmath200 .
proof of @xmath104 : suppose @xmath201 .
applying lemma [ splitting ] , we get @xmath202 which implies @xmath188 . for the converse
, we note that @xmath188 and @xmath203 gives us @xmath202 and therefore @xmath204 .
we finally have the necessary lemmas to prove the conjecture .
the following theorem is the main theorem of this paper which proves the conjecture and characterizes the equality case . if @xmath49 is a linear code over @xmath1 that contains @xmath0 , then @xmath51 .
further the equality holds if and only if @xmath49 is derived from a fixed basis . the first part of lemma [ conditions
] shows that @xmath183 and along with second part of lemma [ conditions ] , @xmath181 and @xmath182 are collections of subspaces that satisfy the conditions of lemma [ lovasz ] . applying that lemma to @xmath181 and @xmath182 we get , @xmath205 the theorem
is finally proved by the following calculation : @xmath206 we note that the equality holds only if @xmath207 for all @xmath208 . in particular , this means @xmath209 and thus the code contains @xmath8 one dimensional subspaces . applying proposition [ one - dim space ] , we infer that @xmath28 is derived from a fixed basis .
a natural question is whether every maximal linear code contains the full space .
this would prove the conjecture that the size of a maximal linear code in @xmath1 is at most @xmath3 .
however , this strategy will fail because there are examples of maximal codes in the literature that do not contain the full space @xcite .
we have proved the conjecture on the bound of a linear code , in a projective space , that contains the full space as a codeword .
the maximal linear codes containing the full space are characterized as codes derived from a fixed basis .
however , all known examples of linear codes that does not contain the full space also have at most @xmath3 codewords .
this general conjecture would be an interesting problem to consider for future research .
further it can be shown that these proofs can be adapted in a more general framework of lattices .
we have not explored the various lattice connections in this paper . however hamming spaces are examples of distributive lattices and projective spaces are examples of modular lattices .
the codes that are derived from a fixed basis are embeddings of distributive lattices into the modular lattice of a projective space .
we conjecture that the size of the largest distributive sub - lattice of a geometric modular lattice of height @xmath8 must be @xmath3 .
the first author thanks prashant n. for stimulating discussions about projective spaces and gautam shenoy for verifying the proofs of the assertions .
m. gadouleau , z. yan , `` packing and covering properties of subspace codes for error control in random linear network coding'',_ieee transactions on information theory _ , vol .
56 , no . 5 , pp .
2097 2108 , may 2010 . | the set of all subspaces of @xmath0 is denoted by @xmath1 .
the subspace distance @xmath2 defined on @xmath1 turns it into a natural coding space for error correction in random network coding .
a subset of @xmath1 is called a code and the subspaces that belong to the code are called codewords . motivated by classical coding theory ,
a linear coding structure can be imposed on a subset of @xmath1 .
braun , etzion and vardy conjectured that the largest cardinality of a linear code , that contains @xmath0 , is @xmath3 . in this paper , we prove this conjecture and characterize the maximal linear codes that contain @xmath0 . |
cosmic structure forms hierarchically .
any object in the universe today , started with copious numbers of small progenitors at redshifts currently inaccessible to direct observations . traditionally in galaxy formation
@xcite = @xmath1 k halos are assumed to be the first cooling halos . nevertheless since the late 1960 s it has been known that molecular hydrogen , formed in the gas phase , can dominate cooling in objects of smaller virial temperature and mass @xcite . neglecting this early phase of h@xmath2 cooling
halos has been justified by arguing that h@xmath2 is destroyed via radiative feedback effects ( cf . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the photo - dissociation of h@xmath2 via the solomon process by an early soft ultraviolet background ( uvb ) is generally assumed as the main reason @xcite .
the mass scale of halos considered enters exponentially in the collapsed mass fraction and the abundance of halos .
figure [ fig : ps ] shows the predicted abundances of the earliest building blocks of galaxy formation as a function of redshift for the latest concordance cosmology using the sheth - tormen formalism @xcite .
the different lines correspond to different virial masses .
the solid line corresponds to halos with virial temperatures of @xmath1 k , the temperature at and above which atomic hydrogen line cooling is dominant . at redshift 30 ,
e.g. , the difference of abundances of @xmath3 and = @xmath1 k halos is five orders of magnitude .
even at redshift 10 this disparity is still a factor of a thousand . when studying reionization and chemical evolution of galaxies and the intergalactic medium , one needs to consider stellar feedback .
the simple fact that the binding energy of the gas of smaller mass halos is even less than the kinetic energy deposited by even one supernova ( sn ) is illustrated in figure [ fig : ps]b .
surely whether the atomic hydrogen line ( ly@xmath4 ) cooling halos are formed from pristine primordial gas or are mergers of many tens of progenitors that massive stars have enriched and expelled the gas from should make a significant change in their further evolution .
the minimum mass of star forming halos is undoubtedly an important issue independent of the techniques employed to study structure formation .
advances in cosmological hydrodynamics and its numerical methods @xcite allow now detailed investigations of all the relevant physical processes . modeling
the expected negative feedback from an early soft uvb is straightforward as a background flux only causes a spatially constant photo - dissociation rate in the chemical reaction network being solved when h@xmath2 does not exist at high enough abundances to self - shield .
used eulerian adaptive mesh refinement ( amr ) simulations to investigate the role of such a h@xmath2 dissociating ( lyman - werner ; lw ) background on the minimum mass of halos within which primordial gas can first cool for a variety of radiation amplitudes .
in addition to a lw background , the collapse of halos within relic regions can be either delayed or catalyzed .
@xcite used amr simulations with a short - lived 3 myr hydrogen ionizing uvb that simulates a nearby massive , metal - free ( pop iii ) star .
they found that halo collapses are prolonged if @xmath5 0.1 and catalyzed
if below this critical value , where is in units of @xmath6 erg s@xmath7 @xmath8 hz@xmath7 sr@xmath7 at a wavelength of 912 . in the case of a large uvb , the collapse is delayed due to lower gas densities and higher cooling times . in the small uvb regime ,
excess free electrons in the relic region accelerate h@xmath2 formation . in both cases , feedback in relic subsides after @xmath930% of a hubble time . strong suppression of h@xmath2 formation also occurs in @xmath10 halos with a lw background @xmath11 0.01 . similarly addressed this issue using smoothed particle hydrodynamics ( sph ) .
they found an additional effect on the minimum collapse mass of dynamical heating from the mass accretion history of the halo .
as the heat input increases , the virial temperature must rise before h@xmath2 cooling can start to dominate , and a cool phase develops in the center of the potential well .
self - consistent calculations in which the sources produce the radiation backgrounds which in turn affect the number of new sources are feasible so far only with semi - analytic approaches and small volume cosmological simulations at low spatial resolutions @xcite . from these studies
, one can derive realistic upper limits on the amplitude of the expected soft uvb . in all studies that include radiation sources in halos less than @xmath1 k halos , the largest the soft uvb flux can get before the @xmath12 halos
dominate the emission is @xmath13 ( cf .
? * ; * ? ? ?
* ; * ? ? ?
interestingly , for a lw intensity of @xmath14 , mba01 found that @xmath15 halos were still able to cool and collapse . on the other hand at that ,
yahs03 suggest negative feedback should become so strong that the critical h@xmath2 fraction for cooling can not be reached and cooling will not occur .
however , they did not explore this further with detailed higher resolution simulations to check whether their analytical expectation would hold .
we present a series of fourteen very high resolution eulerian amr simulations designed to see how the largest possible feedback may raise the minimum mass in which primordial gas will cool by molecular hydrogen .
the simulations techniques and details of the suite of calculations is the topic of the next section . in the following sections , we describe the results that show h@xmath2 cooling
can not be neglected in early structure formation . in the discussion
, we describe the nature of the uvb and why h@xmath2 cooling can occur in such large radiation backgrounds
. we also comment on the large range of questions in cosmological structure formation that this conclusion affects .
lccccc h2 & yes & yes & 0 & 29.7 & 31.1 + h2lw22 & yes & yes & 10@xmath16 & 28.3 & 27.5 + h2lw21 & yes & yes & 10@xmath17 & 24.4 & 24.7 + h2lw20 & yes & yes & 10@xmath18 & 20.5 & 22.4 + noe - h2 & yes & no & 0 & 18.7 & 23.4 + noe - h2lw20 & yes & no & 10@xmath18 & 16.8 & 21.4 + h+he & no & yes & 0 & 15.9 & 16.8
we use the eulerian amr hydrodynamic code _ enzo _ @xcite to study the importance of h@xmath2 cooling in early galaxy formation .
_ enzo _ uses an @xmath19-body adaptive particle - mesh solver @xcite to follow the dark matter ( dm ) dynamics .
we perform two cosmological realizations with different box sizes and random phases and wmap 1 year parameters of ( @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath19 ) = ( 0.72 , 0.73 , 0.27 , 0.024@xmath25 , 0.9 , 1 ) @xcite . the significantly different third year wmap results favor lesser small - scale power that delays high - redshift structure formation by @xmath940% and alters the statistical properties of dm halos @xcite .
the ratio @xmath22/@xmath23 also only lowered by 5% to 5.70 .
however these differences have no effect on the evolution and assembly of individual halos studied here that have typical mass accretion histories .
the initial conditions are the same as in @xcite .
both realizations have a top grid with a resolution of 128@xmath26 with three nested subgrids with twice finer resolution and are initialized at z = 129 ( 119 ) with the cosmics package @xcite .
the box size is 1.0 ( 1.5 ) comoving mpc .
the innermost grid has an effective resolution of 1024@xmath26 with dm particle masses of 30 ( 101 ) and a side length of 250 ( 300 ) comoving kpc .
we refine the amr grids when either the dm ( gas ) exceeds three times the mean dm ( gas ) density on the same level .
we also refine so that the local jeans length is resolved by at least 4 cells .
we focus on the region containing the most massive halo in the simulation box and follow its evolution until it collapses to an overdensity of @xmath27 that corresponds to a refinement level of 15 and a spatial resolution of @xmath93000 ( 4000 ) proper au .
we perform each realization with seven sets of assumptions .
table [ tab : sims ] summarizes them .
we use a nine - species ( h , h@xmath28 , he , he@xmath29 , he@xmath30 , e@xmath31 , h@xmath2 , h@xmath32 , h@xmath31 ) non - equilibrium chemistry model @xcite for all runs except the h+he runs that do not include h@xmath2 cooling .
the nine - species runs are specified by `` h2 '' and use the h@xmath2 cooling rates from @xcite . above number densities of @xmath33 or in an intense ultraviolet radiation field , the excited states of h@xmath2 become populated . the h@xmath2 collisional dissociation rates from @xcite are calculated in the ground state ; therefore we use a density dependent h@xmath2 dissociation rate from @xcite that considers this phenomenon .
runs with h@xmath2 dissociating ( lyman - werner ; lw ) radiation are denoted by `` lw '' followed by its negative log - flux .
we set to @xmath34 , @xmath6 , and @xmath35 erg s@xmath7 @xmath8 hz@xmath7 because the first two are typical values one finds in semi - analytic models of reionization and the latter investigates the case of an very large uvb .
we use the h@xmath2 photo - dissociation rate coefficient for the solomon process from @xcite of @xmath36 s@xmath7 .
we do not consider the self - shielding of lw photons . because the molecular core only becomes optically thick in the late stages of collapse and above column densities of 10@xmath37 @xmath8 @xcite
, we expect our results to not be drastically affected by neglecting lw self - shielding . additionally , lw self - shielding may be unimportant up to column densities of @xmath38 @xmath8 if the medium contains very large velocity gradients and anisotropies @xcite .
free electrons are necessary to form h@xmath2 in the gas phase . in order to restrict h@xmath2 formation to ly@xmath4 line cooling halos in our `` noe- '' calculations ,
we reduce the residual free electron fraction from @xmath39 @xcite to a physically low @xmath40 at the initial redshift .
this setup is designed to find the first halos that can collapse and form stars once free electrons from collisionally ionized hydrogen becomes available to catalyze h@xmath2 formation @xcite .
this work is an extension of the original work of mba01 , adding the calculations with = @xmath35 and ones in which h@xmath2 can not cool until ly@xmath4 cooling becomes efficient .
we consider these extreme cases to strengthen the point made in mba01 in which a uvb only increases the critical halo collapse mass , never completely suppressing the crucial importance of h@xmath2 formation and cooling .
our maximum spatial resolution in the finest amr level is a factor of four smaller than mba01 ; however , this does not cause any differences between our work and mba01 because these finest grid patches only exist in the dense , central core during the final 150 kyr of the collapse . in galaxy formation models ,
the virial temperature is a key quantity as it controls the cooling and star formation rates in a given halo .
we define a halo as the material contained in a sphere of radius @xmath41 enclosing an average dm overdensity @xmath42 of 200 . for an isothermal singular sphere , the virial temperature @xmath43 where @xmath44 is the circular velocity ( see * ? ? ?
* with @xmath45 = 1 ) .
here @xmath46 is the mean molecular weight in units of the proton mass @xmath47 , and @xmath48 is boltzmann s constant .
we use this definition of in this paper with @xmath46 = 0.59 .
we choose this value of @xmath46 to be consistent with the literature on galaxy formation even though the halos presented in this paper are neutral and have @xmath46 = 1.22 .
we first describe the halo properties at collapse .
then we compare them to previous studies of collapsing halos in the presence of a soft uvb .
figure [ fig : rho ] shows density - squared weighted projections of gas density and temperature when each calculation can cool and collapse to an overdensity of 10@xmath49 .
it illustrates the large difference in the sizes and morphologies of the collapsing halos in the various cases of negative feedback .
all panels have the same field of view of 1.2 proper kpc and same color scales .
it is clear from the relative sizes of the collapsing halos that the critical halo mass to cool increases with the amount of negative feedback .
the virial shock and numerous central shocks heat the gas to the virial temperature .
the central shocks create fine structure seen in the temperature projections . in all of the h@xmath2 cases , we see neither fragmentation nor large - scale disk formation . the internal structures of the halos with h@xmath2 cooling and residual free electrons are similar to previous studies of pop iii star forming halos ( mba01 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , exhibiting a turbulent medium with a radially monotonically decreasing density profile and a cool central core .
figure [ fig : mass]a depicts the halo mass and redshift when the halo collapses for all of the runs , and figure [ fig : mass]b shows their central temperature at the same epoch .
the collapse redshifts , @xmath50 and @xmath51 , are also listed in table [ tab : sims ] for simulations a and b , resepectively . as seen in other studies ( mba01 , yahs03 ) , the minimum dm halo mass to collapse increases with the background intensity .
the h+he case predictably collapses at @xmath52 k , and all of the halos with h@xmath2 cooling collapse at much smaller masses .
the temperature of the central core increases with halo mass from 300 k to 1000 k for halo masses @xmath53 and @xmath54
. restricting the data to models with residual electrons , the central temperature increases as a power - law , @xmath55 where @xmath56 , @xmath57 , and is in units of solar masses . this relationship is plotted in figure [ fig : mass ] . with neither residual electrons nor an uvb ( noe - h2 ) , the most massive halo collapses at @xmath58 at z = 18.7 ( 23.4 ) . here
h@xmath2 formation in the gas phase can only become important when sufficient free electrons are created by collisional ionization .
virial heating in the center of halos can increase temperatures up to twice the virial temperature @xcite that collisionally ionizes hydrogen in the central shocks and initiates h@xmath2 cooling @xcite in halos well below virial temperatures of 10@xmath59 k. these shocks are abundant throughout the central regions .
figure [ fig : efrac ] shows radial profiles of temperature and electron fraction for both simulations and depicts gas shock - heating up to @xmath60 k and raising electron fractions up to 10@xmath61 .
the electron fractions remain at unrealistically low values less than 10@xmath62 in low density regions where gas has not been collisionally ionized .
the higher density regions have condensed to densities above @xmath63 after free electrons in protogalactic shocks induced h@xmath2 cooling . a similar but extreme model , noe - h2lw20 , demonstrates that even in the presence of a very large uvb of = 10@xmath18 gas is able to form a cool and dense central molecular core at a mass of @xmath64 at redshift 16.8 ( 21.4 ) .
two major mergers in simulation a occur between z = 1721 , and the associated heating allows the halo to begin cooling by h@xmath2 .
a central core only forms once the system is adequately relaxed after the mergers , which causes the collapse mass difference between the realizations . by not fully resolving weak shocks in our main calculations , it is possible to underestimate the electron fraction .
we performed simb - h2lw20 with an additional refinement criterion that resolves the `` cooling length '' , @xmath65 , by at least 2 cells .
the large- and small - scale structure in the simulation is unchanged .
when we resolve these weak shocks , the increased electron fraction marginally accelerates the collapse , which occurs 780 kyr earlier at z = 22.5 .
the virial mass at this time is @xmath66 compared with @xmath67 .
hence we believe that the critical halo mass to collapse as a function of the lw background is independent of this refinement criterion . the combination of a recent major merger and collisional ionization produces complex structures as seen in the density and temperature projections of the sima - noe - lw20 and simb - noe- calculations in figure [ fig : rho ] , unlike the other h2 models with a single cool central core . through a series of amr calculations with varying uvb intensities
, mba01 found the minimum dm halo mass @xmath68 in order to cool and condense 4% of the baryons .
this fraction of cool and dense gas agrees with simulations of the formation of pop iii stars @xcite .
there is some scatter of @xmath90.5 dex in this threshold mass ( see also yahs03 ) .
for the uvb intensities used in our models ( = 0 , 10@xmath16 , 10@xmath17 , 10@xmath18 ) , the critical collapse masses are @xmath69 , @xmath70 , @xmath71 , and @xmath72 .
our calculations with h@xmath2 cooling and residual free electrons agree with the results of mba01 .
yahs03 studied the minimum collapse mass but also included the effects of self - shielding . through their sph simulations and arguments using equilibrium h@xmath2 abundances ,
they conclude that an uvb intensity of = 0.1 nearly prevents halo collapses below @xmath73 7000 k where ly@xmath4 cooling becomes efficient .
they also deduce that = 1.0 completely prevents any h@xmath2 cooling in these low - mass halos , based on h@xmath2 dissociation timescales .
we find the contrary in our h2lw21 and h2lw20 calculations where the most massive halo collapses with a mass of @xmath74 and @xmath75 , respectively . even in our noe- runs ,
the halo collapses when @xmath9 4000 k , i.e. before ly@xmath4 cooling becomes important , which is around the same mass scale that the h2lw20 runs condense .
we ignore self - shielding in our calculations , but this would only decrease the critical collapse mass and strengthens our main conclusion that h@xmath2 cooling is always dominant , even in the presence of a large lw flux . yahs03 used cosmological sph simulations and h@xmath2 formation and dissociation timescales to argue that a lw background intensity of @xmath76 suppresses h@xmath2 formation so halos can not cool before virial temperatures of 7000 k are reached . employing the same argument , we see that the h@xmath2 formation timescale @xmath77 with typical central values found in high - redshift halos before any radiative cooling becomes efficient ( see * ? ? ?
here @xmath78 and @xmath79 are the h@xmath2 and electron number fraction , respectively , and @xmath80 is the baryon number density . @xmath81
cm@xmath26 s@xmath7 is the h@xmath83 formation rate coefficient by electron photo - attachment at @xmath84 k @xcite .
this timescale is a factor of 1000 smaller than the value calculated in yahs03 because we use the quantities from the halo center as compared to the mean values .
the h@xmath2 dissociation timescale is @xmath85 = 23/ kyr , which is comparable with @xmath86 using the values above . the halo characteristics and the collapse redshift will likely depend on halo merger histories as seen in these two realizations . the better statistics of mba01 sampled this effect well .
here the scatter of threshold mass is @xmath90.5 dex and is smaller than the mass difference between halos with virial temperatures of 4000k and 10000k .
thus our limited sample of halos should not change our result of the importance of h@xmath2 cooling in halos well below = 10@xmath59 k , even with very large lw radiation backgrounds .
structure formation in the high - redshift universe is contained within shallow potential wells that are sensitive to negative feedback from a uvb .
additionally local positive and negative feedback will influence star formation and further complicate estimates of halo mass scales .
some examples include * _ positive feedback_ enhanced h@xmath2 formation in relic regions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and ahead of the ionization front @xcite , dust and metal line cooling @xcite , * _ negative feedback_ baryonic explusion from host halos @xcite , photo - evaporation @xcite , entropy floors @xcite .
these processes are not within the scope of this paper and will be considered in later publications that utilize three - dimensional radiation hydrodynamic simulations with pop iii star formation . here
we only focused on the effects of a uvb on low - mass halos .
the intensity of the uvb is a monotonically increasing function of redshift as more halos form stars .
the uvb increases on the order of a hubble time , which is much shorter than a dynamical time of a collapsing halo and justifies the use of a constant intensity in our calculations .
self - consistent studies that evolve the uvb according to star formation rates only find to be in the range of 0.01 and 0.1 at redshifts 1520 ( yahs03 , wa05 ) .
wa05 calibrated their model against the wmap1 measurement of @xmath87 = 0.17 . with the wmap3 result of the electron scattering optical depth @xmath87 = 0.09 and less small - scale power
, uvb intensities will be even lower at these redshifts .
we can relate reionization to lw radiation by equating in the lw band to a common quantity in reionization models , the ratio of emitted hydrogen ionizing photons to baryons , @xmath88 , where @xmath89 is the cosmic mean of the baryon number density . assuming that @xmath90 is constant in the lw band , the number density of lw photons is @xmath91 where @xmath92 is planck s constant and @xmath93 = @xmath94 hz , @xmath95 hz bound the lw band . to relate to @xmath96
, we must consider the intrinsic ionizing spectrum and absorption from the igm and host halo . at redshift 20 , the majority of star forming halos host pop iii stars that emit a factor @xmath97 more hydrogen ionizing photons than lw photons because of its @xmath98 k surface temperature . since the number density of sources exponentially increases with redshift , the majority of the early uvb at a given redshift originates from cosmologically nearby ( @xmath99 ) sources .
lyman line resonances absorb a fraction @xmath100 of the lw radiation in the intergalactic medium in this redshift range , producing a sawtooth spectrum @xcite .
additionally , absorption in the host halo reduces the number of ionizing photons that escape into the igm by a fraction @xmath101 . for pop iii halos ,
this factor is close to unity @xcite . by considering these multiplicative processes
, we now estimate @xmath102 this estimate is in agreement with the reionization models of @xcite and wa05 ( see also * ? ? ?
these models find that sources produce a large uvb of @xmath9 1 prior to reionization .
when pop iii stars dominate the uvb , the lw radiation will be small in comparison to the volume averaged hydrogen ionizing emissivity because of the intrinsically hard pop iii spectra that peaks at @xmath9300 .
hence high - redshift halos should not be exposed to a large uvb , i.e. @xmath5 0.1 , and h@xmath2 formation will remain important before reionization .
nearby star formation can boost the lw radiation over its background value , but these bursts are short - lived as pop iii lifetimes are only @xmath93 myr @xcite . for example , a 100 star produces 10@xmath103 lw photons s@xmath7 and will produce @xmath76 in the surrounding 3 proper kpc , neglecting any h@xmath2 self - shielding .
the lw background is uniform outside these spheres of influence .
the bursting nature of pop iii star formation does not affect the time evolution of the background .
the intensity only depends on the number of sources in a redshift range @xmath104 = 13.6 ev / 11.18 ev 1 , where the two energies bound the lw band , because any radiation redward of the lyman break contributes to the lw background . using a conservative minimum halo mass for pop iii star forming halos of 3 @xmath105 10@xmath106 at redshift 20 , there are @xmath942000 halos that have hosted a pop iii star in the volume contained within @xmath107 , using wmap3 parameters with sheth - tormen formalism .
clearly the background is uniform considering the sheer number of sources within this optically thin volume .
local perturbations from pop iii star formation should only affect the timing of nearby star formation but not the global star formation rate .
figure [ fig : h2form ] shows simb - lw20 twenty million years before the central core condenses . at this time , the core is just beginning to cool by h@xmath2 , catalyzed by the free electrons created in the central shocks . in these shocks , temperatures reach @xmath108 k and electron fractions up to @xmath109 exist there .
these conditions result in h@xmath2 formation timescales less than 25 kyr , which is necessary to cool in a uvb of @xmath9 1 . within the central 10 pc ,
hot and cold gas phases exist .
the hot phase exists behind the shocks that have lower densities around 10 and @xmath110 25 kyr .
this is where h@xmath2 cooling is catalyzed by collisional ionization in these shocks .
the cold phase has already cooled through h@xmath2 and has high densities and larger @xmath86 values .
both phases are apparent in the panels of figure [ fig : h2form ] .
similar conditions create h@xmath2 in the collapses in the `` noe- '' calculations , which have sufficient gravitational potential energy , resulting in temperatures above @xmath1 k in central shocks .
hence h@xmath2
formation is possible in the centers of high - redshift halos with virial temperatures below 10@xmath59 k , even with a uvb of intensity @xmath111 1 , larger than expected from semi - analytic models of reionization .
two consequences of a lower critical ly@xmath4 cooling halo mass are more frequent and earlier galaxy formation and higher mass fractions in cooling halos . at redshift 20 , e.g. , abundances of = 4000 k halos are an order of magnitude larger than = 10@xmath59 k halos , resulting from the exponential nature of press - schechter formalism . the mass fraction contained in these halos is three times higher than @xmath1 k halos . in semi - analytic models of reionization and chemical enrichment , the star formation rate ( sfr ) is linearly dependent on the collapsed mass fraction since the sfr is usually a product of mass fraction and star formation efficiency , which is the fraction of gas collapsing into stars ( e.g. * ? ? ?
the star formation efficiency for primordial stars is @xmath112 with a single massive star forming in dark matter halos with mass @xmath113 @xcite .
this fraction may rise to a few percent in dwarf galaxies as widespread star formation occurs @xcite .
various studies predict that a majority of the reionizing flux originates from dwarf galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if the mass contained in star forming halos is three times greater than previously thought , some of the predicted attributes , e.g. photon escape fractions and star formation efficiencies , of high - redshift dwarf galaxy will require appropriate adjustments to match observations , such as the wmap3 measurement of optical depth to electron scattering @xcite and gunn - peterson troughs at @xmath114 @xcite .
we conducted a suite of fourteen cosmology amr simulations that focus on the importance of h@xmath2 cooling with various degrees of negative feedback . we summarize the findings of each model below .
\2 . above
= @xmath6 , it had been argued that an h@xmath2 dissociating background would inhibit any h@xmath2 formation until the halo could cool through ly@xmath4 cooling .
we showed that central shocks provide sufficient free electrons from collisional ionization to drive h@xmath2 formation faster than dissociation rates even in a = @xmath35 background .
\3 . in our `` noe- '' models ,
we explored when collisional ionization becomes important and conducive for h@xmath2 formation .
this occurs at @xmath9 4000 k. recent major mergers above this mass scale create complex cooling structures , unlike the non - fragmented central cores in smaller halos .
@xcite independently studied halo collapses with _ enzo _ and similarly considered primordial gas chemistry and nine different uvb intensities ranging from zero to = 1 .
they agree with our conclusions in that primordial gas in @xmath116 halos can catastrophically cool and collapse even in models with @xmath117 .
they attribute the collapse to the increased h@xmath2 cooling rates at higher temperatures that is caused by greater dynamical heating in halos with @xmath118 .
the cooling rate per molecule is 100 times larger at 2000 k than at 500 k , typical of pop iii star - forming halos without an uvb .
most likely , the combination of the elevated h@xmath2 cooling rates and electron fractions from internal protogalactic shocks instigate the halo collapses in a strong uvb ( @xmath117 ) . in any case , h@xmath2 cooling triggers collapses in halos with virial temperatures well below 10@xmath59 k. the lower critical halo mass , corresponding to @xmath9 4000 k , increases mass fraction contained in these halos by three times at redshift 20 and the number density of high - redshift star forming halos by an order of magnitude ! by considering additional cases of extremely large negative feedback , we have strengthened the results of mba01 that h@xmath2 cooling plays a key role in high - redshift structure formation . we conclude that a uvb only delays and never completely suppresses h@xmath2 formation and cooling and subsequent star formation in these low - mass halos . | the first luminous objects in the concordance cosmology form by molecular hydrogen cooling in dark matter dominated halos of masses @xmath0 .
we use eulerian adaptive mesh refinement simulations to demonstrate that in the presence of a large soft ultraviolet radiation background , molecular hydrogen is the dominant coolant . even for very large radiation backgrounds , the halo masses that cool and collapse are up to two orders of magnitude smaller than the halos that cool via atomic hydrogen line cooling .
the abundance of cooling halos and the cosmic mass fraction contained within them depends exponentially on this critical mass scale .
consequently , the majority of current models of cosmological reionization , chemical evolution , supermassive black hole formation , and galaxy formation underestimate the number of star forming progenitors of a given system by orders of magnitude . at the highest redshifts ,
this disagreement is largest .
we also show that even in the absence of residual electrons , collisional ionization in central shocks create a sufficient amount of electrons to form molecular hydrogen and cool the gas in halos of virial temperatures far below the atomic cooling limit . |
we thank a. kostyuk , l. gerland , r. vogt , and l. csernai for many fruitful discussions . 9 j. schaffner - bielich , m. hanauske , h. stcker and w. greiner , astro - ph/0005490 .
r. b. wiringa , v. fiks and a. fabrocini , phys .
c * 38 * ( 1988 ) 1010 .
j. margueron , j. navarro , n. van giai and w. jiang , nucl - th/0110026 .
b. a. li , phys .
* 88 * ( 2002 ) 192701 ; b. a. li , nucl - th/0301039 . b. a. li , phys .
c * 67 * ( 2003 ) 017601 b. a. li , nucl .
a * 708 * ( 2002 ) 365 r. stock , phys .
* 135 * ( 1986 ) 259 .
s.a . bass et al .
42 * , 279 ( 1998 ) .
m. bleicher et al . , j. phys .
g * 25 * , 1859 ( 1999 ) .
m. di toro et al . , nucl - th/0301033 .
c. greiner and h. stcker , phys . rev .
d * 44 * ( 1991 ) 3517 .
afanasiev et al .
( na49 collab . ) , phys . rev .
c. * 66 * ( 2002 ) 054902 a. mischke et al .
( na49 collab . ) , j. phys .
g. * 28 * , 1761 ( 2002 ) .
a. mischke et al .
( na49 collab . ) , nucl - ex/0209002 , to be published in nucl .
a. the na49 collaboration , addendum-10 to proposal cern / spsc / p264 . w. cassing and e. l. bratkovskaya , phys .
rep . * 308 * , 65 ( 1999 ) .
h. stcker and w. greiner , phys . rep . *
137 * , 277 ( 1986 ) .
g. f. bertsch and s. das gupta , phys . rep . *
160 * , 189 ( 1988 ) .
w. cassing , v. metag , u. mosel , and k. niita , phys . rep . * 188 * , 363 ( 1990 ) . c. m. ko and g. q. li , j. phys .
g * 22 * , 1673 ( 1996 ) .
f. wang , h. liu , h. sorge , n. xu , and j. yang , phys .
c * 61 * , 064904 ( 2000 ) .
w. cassing , e. l. bratkovskaya , s. juchem , nucl .
a * 674 * , 249 ( 2000 ) .
j. geiss , w. cassing , and c. greiner , nucl .
a * 644 * , 107 ( 1998 ) .
h. weber , e.l .
bratkovskaya , and h. stcker , phys .
b * 545 * , 285 ( 2002 ) .
h. weber , e. l. bratkovskaya , w. cassing and h. stcker , phys .
c * 67 * ( 2003 ) .
w. ehehalt and w. cassing , nucl .
a * 602 * , 449 ( 1996 ) .
http://www.gsi.de / gsi - future / cdr/. l. ahle et al .
( e866 and e917 collab . ) , phys . lett .
b * 476 * , 1 ( 2000 ) ; phys .
b * 490 * , 53 ( 2000 ) .
m. belkacem et al . , phys .
c * 58 * , 1727 ( 1998 ) ; l.v .
bravina et al . , phys .
b * 434 * , 379 ( 1998 ) ; l.v .
bravina et al . , j. phys .
g * 25*,351 ( 1999 ) ; l.v .
bravina et al . , phys .
c * 60 * , 024904 ( 1999 ) .
w. cassing , nucl .
phys . a * 700 * , 618 ( 2002 ) .
r. rapp and e. v. shuryak , phys .
* 86 * , 2980 ( 2001 ) . c. greiner and s. leupold , j. phys .
g * 27 * , l95 ( 2001 ) .
s. ahmad et al .
( e891 collab . ) , phys .
b * 382 * , 35 ( 1996 ) ; c. pinkenburg et al .
( e866 collab . ) , nucl .
a * 698 * , 495c ( 2002 ) .
s. albergo et al . , ( e896 collab . ) , phys .
lett . * 88 * , 062301 ( 2002 ) .
f. antinori at al .
( wa97 collab . ) , nucl .
a * 661 * , 130c ( 1999 ) .
p. braun - munzinger , j. cleymans , h. oeschler and k. redlich , nucl .
a * 697 * , 902 ( 2002 ) .
l. ramello , ( na50 collab . ) , talk given at qm2002 c. spieles , r. vogt , l. gerland , s. a. bass , m. bleicher , h. stcker and w. greiner , phys .
c * 60 * ( 1999 ) 054901 w. cassing and e. l. bratkovskaya , nucl .
a * 623 * ( 1997 ) 570 | the isospin and strangeness dimensions of the equation of state are explored .
ria and the sis200 accelerator at gsi will allow to explore these regions in compressed baryonic matter . @xmath0sn+@xmath0sn and @xmath1sn+@xmath1sn collisions as well as the excitation functions of @xmath2 , @xmath3 and the centrality dependence of charmonium suppression from the urqmd and hsd transport models are presented and compared to data .
unambiguous proof for the creation of a novel phase of matter from strangeness and charm yields is not in sight .
the properties of neutron - rich matter at high densities and temperatures are largely unknown . collisions of heavy rare isotopes provide an excellent tool to explore this unknown region in the phase diagram of nuclear matter . the planned rare isotope accelerator ( ria )
( up to @xmath4 amev ) and the sis200 at gsi ( up to @xmath5 agev ) are the upcoming experimental facilities that will access new energy density regions for these collisions allowing to investigate the equation of state ( eos ) of compressed hadronic matter in regions ( 2 - 10 @xmath6 ) or even in directions ( @xmath7 ) which have not been experimentally accessible before . among the most interesting open questions
is the isospin and strangeness dependency of the eos , especially at high baryochemical potentials .
this has been the subject of intensive research , as it is not only important in the study of radioactive nuclei but bears consequences on various important astrophysical issues as well @xcite .
the eos of isospin asymmetric matter can be written as @xmath8 , where @xmath9 is the isospin asymmetry , @xmath10 is the energy density in isospin symmetric nuclear matter , and @xmath11 is a parametrisation for the isospin dependence of the nuclear symmetry energy @xcite .
parametrisations for @xmath11 fall into two main categories @xcite : ( i ) @xmath11 rises with @xmath12 , ( ii ) @xmath11 falls with increasing density @xmath12 .
the latter case has the interesting feature that symmetric matter may become unstable at high densities .
the ( @xmath13 ) ratio is particularly sensitive to the high density behaviour of @xmath14 @xcite .
this can be easily understood qualitatively within the @xmath15 resonance model for pion production @xcite : here first - chance independent nucleon - nucleon collisions @xcite yield a primordial ( @xmath13 ) ratio given by @xmath16 .
thus , the charged pion ratio yields information on the isospin asymmetry ( n / z ) in the dense participant matter , which depends on the high density behaviour of @xmath14 @xcite .
the ( @xmath13 ) ratio can therefore be used to probe the eos of neutron - rich nuclear matter . for our investigation of this effect the urqmd model v1.2 @xcite is applied to @xmath0sn+@xmath0sn and @xmath1sn+@xmath1sn collisions at @xmath17 agev and @xmath18 agev beam energy .
fig .
[ sn ] shows the time evolution of the ( n / p ) and ( @xmath13 ) ratios . for @xmath1sn+@xmath1sn ,
both the ( n / p ) and the @xmath19 ratios are roughly time independent and of order one , as expected for the symmetric system . for the asymmetric system @xmath0sn+@xmath0sn , the ( n / p ) ratio drops during the compression phase from its initial value of @xmath20 and then saturates . at 2 agev , this saturation occurs after 15 fm / c at a value of ( n / p ) @xmath21 , whereas at 11 agev the saturation occurs earlier ( 10 fm / c ) and to a lower final value of ( n / p ) @xmath22 .
the @xmath19 ratio rises slightly during the early phase of the reaction and saturates rather later than ( n / p ) ( @xmath23 fm / c at 2 agev and @xmath24 fm / c at 11 agev ) .
the saturation values of @xmath25 ( 2 agev ) and @xmath26 ( 11 agev ) fall far below the first - chance @xmath15 resonance model prediction @xmath27 .
this is evident , as it is not the initial but rather the local and dynamically changing ( n / p ) ratio that determines the @xmath19 ratio .
in addition , pion reabsorptions and rescatterings reduce the sensitivity of @xmath19 to ( n / p ) . at higher energies
, the increasing number of sequential nucleon - nucleon collisions drives the @xmath19 ratio towards @xmath28 via this mechanism . for isospin - asymmetric matter
, a distillation effect has recently been proposed @xcite , which is powered by chemical instabilities in the liquid - gas phase transition driving the system towards isospin symmetry .
this effect is similar to the well - known _ strangeness _ distillation mechanism in nuclear reactions , which has been proposed as a quark - gluon - plasma ( qgp ) signal @xcite .
the search for the qgp and its properties has been a driving force behind years of relativistic heavy ion physics .
the fact that the @xmath3 ratio and the @xmath29 ratio are found experimentally to exhibit a maximum at @xmath30=10 - 30 agev and drop to half that value at 160 agev has raised speculations about the appearance of the new phase transition. the na49 collaboration has recently started an energy scan at the sps .
first results are available now at 40 and 80 agev @xcite and support these findings .
further studies are underway at 30 and 20 agev @xcite .
let us discuss these maxima in the excitation function on the basis of conventional hadron transport theories do they exhibit analogous maxima ?
in that case , the interpretation of the data as a qgp signal becomes doubtful .
the various hadron spectra are conventionally calculated with non - equilibrium kinetic transport theory ( cf .
@xcite ) .
however , the calculated kaon to pion ratio from central nucleus - nucleus collisions turns out to vary by factors as large as 2 if different transport approaches are applied @xcite .
thus a unique interpretation of the data is questionable so far .
here we study the yields and ratios of protons , kaons , antikaons and hyperons from au+au ( or pb+pb ) collisions in the energy range from 1 agev to 160 agev @xcite .
the aim of this study is twofold : first , to find out the systematic differences between two currently used transport approaches ( urqmd @xcite and hsd @xcite ) , and second , to compare to related experimental data @xcite .
we provide predictions for experimental studies in the near future @xcite , which are also of relevance for the new gsi - proposal @xcite .
the two different transport models employed , i.e. the urqmd v1.3 ( here an updated version has been applied ) and hsd approach have been used to describe nucleus - nucleus collisions from sis to sps energies for several years . though different in the numerical realisation ,
both models are based on the same concepts : strings , quarks , diquarks ( @xmath31 ) accompanied by hadronic degrees of freedom .
it is important to stress that both approaches do not include any explicit phase transition to a quark - gluon plasma ( qgp ) . a common failure of both models in comparison to experimental data may
model independently indicate the appearance of a novel state of strongly interacting matter . as a general overview on the @xmath32 abundancies in central au+au and pb+pb collisions fig .
[ fig4 ] shows the @xmath33 and @xmath34 multiplicities at mid - rapidity as a function of the bombarding energy in comparison to the available data @xcite .
solid lines with open triangles show the results from the urqmd calculations while the solid lines with open squares and dashed lines stem from the hsd approach with and without potentials , respectively . at lower ags energies
the urqmd model gives slightly less pions then hsd ( with / without potential ) , but both models overpredict the mid - rapidity data . around 10
agev is the `` crossing point '' of both transport calculations and at sps energies the tendency turns around : urqmd gives more pions than hsd , so that hsd is now in a better agreement with the experimental data .
the @xmath35 and @xmath36 multiplicities at mid - rapidity are depicted in fig .
[ fig5 ] as a function of the beam energy in comparison to data from refs .
@xcite : the @xmath36 abundancies are well described by both transport models . the @xmath35 yield is slightly overestimated by urqmd at ags energies and underestimated at sps energies , whereas hsd is in reasonable agreement with the data .
thus , fig .
[ fig5 ] demonstrates that an underestimation of strangeness production is not the prevailing issue in comparison to the recent data from na49 @xcite .
both transport models can roughly describe - within their systematic range of uncertainties - the @xmath37 spectra and abundancies . fig .
[ fig6 ] contrasts the @xmath29 and @xmath38 ratios at mid - rapidity as a function of energy for central collisions of au+au ( ags ) or pb+pb ( sps ) with the data @xcite .
the excitation function of the @xmath38 ratio is roughly reproduced by both transport models , the maximum in the @xmath29 ratio seen experimentally is not described quantitatively by either hsd or urqmd . for the @xmath29 ratio both models give quite different results .
hsd gives a monotonous increase of this ratio with bombarding energy ( in qualitative disagreement with the data as pointed out in refs .
@xcite ) , whereas within urqmd the ratio shows a maximum around 10 agev and then saturates to a slightly lower value towards rhic energies , thus reproducing the behaviour of the data qualitatively .
in view of figs . [ fig4 ] and
[ fig5 ] this failure is not primarily due to a mismatch of strangeness production , but more due to an insufficient description of the pion abundancies .
this overestimation of pion abundancies can be attributed to a number of reasons : one problem of the transport approaches used here is that detailed balance is not implemented for @xmath39 transitions with @xmath40 @xcite .
thus multi - particle collisions might change the dynamical picture accordingly and lead to shorter chemical equilibration times @xcite .
in fact , the importance of @xmath41 transitions has been explored in an extended hsd transport approach @xcite for antiproton reproduction by meson fusion for @xmath42 collisions at the ags and sps . in order to achieve a more conclusive answer from transport studies
multi - particle interactions deserve further investigation in future generations of transport codes .
another reason for the overestimation of the pion yield might be that the pions in both transport models are treated as free particles , i.e. with their vacuum mass .
lattice qcd calculations and effective lagrangian approaches like the nambu - jona - lasinio ( njl ) model however , indicate an increase in the pion mass with temperature and density . so the overestimation of the pion yields in hsd and urqmd might be a signature for a dynamically larger pion mass
. moreover , in - medium changes of the strange hadron properties , as known from experimental studies at sis energies , may also show up in the compressed baryonic matter encountered in the ags and sps energy range .
thus , including all medium effects simultaneously in a consistent way might provide a more conclusive interpretation of the ratios in figs .
[ fig6 ] and [ fig7 ] .
however , such calculations require a precise knowledge about the momentum and density dependence of the hadron self - energies which is not available so far . fig .
[ fig7 ] shows the excitation functions of @xmath43-hyperons ( left ) at mid - rapidity as a function of the bombarding energy for central collisions of au+au ( ags ) or pb+pb ( sps ) in comparison to data @xcite . here
both models compare rather well with data . the @xmath44 ratios ( right ) at mid - rapidity are underestimated slightly which should again be attributed to the pion excess in the transport models ( as discussed above ) .
nevertheless , the experimentally observed maxima in the ratios are qualitatively reproduced by both models , indicating that with increasing bombarding energy @xmath45-quarks are more frequently produced within mesons ( @xmath46 ) rather than in associate production with baryons .
a similar trend is also found in statistical models @xcite .
what can we conclude from the common failures of the models studied in comparison to related experimental data and does this failure provide evidence for a new state of matter in view of a qgp ?
in view of the large systematic uncertainties in the discussed transport approaches we can quote an insufficient accuracy in the description of the pion degrees of freedom by both transport models .
thus , the deviations between theory and the data should not be seen as unambiguous signals for any kind of new physics .
another much - discussed possible qgp signal is charmonium production . here ,
new data on the @xmath47 end @xmath48 dependence of @xmath49 production in pb+pb collisions at 158 agev from the year 2000 na50 run has recently become available @xcite .
this data ( one analysis ) , along with urqmd results @xcite ( which are in line with hsd results @xcite ) is shown in fig .
[ jpsi ] .
note that the puzzling drop of the @xmath49 to drell - yan production at high @xmath50 which was present in the old , previously available na50 data is no longer present in the current na50 data .
the new na50 data fall very close to the published urqmd calculations by spieles et al .
@xcite . to our opinion
, both the @xmath47 dependence and the @xmath48 dependence do no longer justify to speak of an anomalous @xmath49 suppression in high energy heavy ion collisions .
we have shown that the exploration of the isospin and strangeness dimensions of the equation of state opens fascinating new areas of research .
especially the upcoming experimental possibilities like the planned rare isotope accelerator ria and the sis200 machine at gsi will allow to explore novel phenomena in compressed baryonic matter in great detail . while conclusive unambiguous proof for the creation of a novel phase of matter , e.g. quark gluon plasma , from strangeness and charm yields is still lacking , the hunt and quest for it has provided new insights and even opened novel fields of research .
it has become a driving force behind modern high energy nuclear physics . |
in 1939 , o. h. keller @xcite stated the famous conjecture which is called the jacobian conjecture : any polynomial mapping @xmath1 with nowhere vanishing jacobian is a polynomial automorphism .
the problem remains open today even for dimension 2 .
we call the smallest set @xmath2 such that the mapping @xmath3 is proper , the asymptotic set of @xmath4 .
the jacobian conjecture reduces to show that the asymptotic set of a complex polynomial mapping with nonzero constant jacobian is empty .
so the set of points at which a polynomial mapping fails to be proper plays an important role .
the second and third authors gave in @xcite a new approach to study the jacobian conjecture in the case of dimension 2 : they constructed a real pseudomanifold denoted @xmath5 , where @xmath6 , associated to a given polynomial mapping @xmath7 , such that the singular part of the variety @xmath8 is contained in @xmath9 where @xmath10 is the set of critical values of @xmath4 .
in the case of dimension 2 , the homology or intersection homology of @xmath8 describes the geometry of the singularities at infinity of the mapping @xmath4 .
our aim is to improve this result in the general case of dimension @xmath11 and compute the intersection homology of the associated pseudomanifold @xmath8 .
let @xmath12 be the leading forms of the components @xmath13 of the polynomial mapping @xmath14 .
we show ( theorem [ valettethuy2 ] ) that if the polynomial mapping @xmath15 has nowhere vanishing jacobian and if @xmath16 , then the condition of properness of @xmath4 is equivalent to the condition of vanishing homology or intersection homology of @xmath8 .
moreover , it is indeed more precise to compute intersection homology than homology . in order to compute the intersection homology of the variety @xmath8
, we show that it admits a stratification which is locally topologically trivial along the strata .
the main feature of intersection homology is to satisfy poincar duality that is more interesting in the case where the stratification has no stratum of odd real dimension .
we show that the variety @xmath8 admits a whitney stratification with only even dimensional strata .
it is well known that whitney stratification are locally topologically trivial along the strata .
it is our pleasure to thank j .-
brasselet for his interest and encouragements .
in this section we set - up our framework . all the considered sets in this article are semi - algebraic .
given a topological space @xmath17 , singular simplices of @xmath17 will be semi - algebraic continuous mappings @xmath18 , where @xmath19 is the standard @xmath20-simplex in @xmath21 . given a subset @xmath22 of @xmath23 we denote by @xmath24 the group of @xmath20-dimensional singular chains ( linear combinations of singular simplices with coefficients in @xmath25 ) ; if @xmath26 is an element of @xmath24 , we denote by @xmath27 its support . by @xmath28 and @xmath29 we denote respectively the regular and singular locus of the set @xmath22 . given @xmath30 , @xmath31 will stand for the topological closure of @xmath17 .
given a point @xmath32 and @xmath33 , we write @xmath34 for the ball of radius @xmath35 centered at @xmath36 and @xmath37 for the corresponding sphere , boundary of @xmath34 .
we briefly recall the definition of intersection homology ; for details , we refer to the fundamental work of m. goresky and r. macpherson @xcite ( see also @xcite ) .
let @xmath22 be a @xmath38-dimensional semi - algebraic set .
a * semi - algebraic stratification of @xmath22 * is the data of a finite semi - algebraic filtration [ dfn_stratification ] x = x_m x_m-1 x_0 x_-1 = such that for every @xmath20 , the set @xmath39 is either empty or a topological manifold of dimension @xmath20 .
a connected component of @xmath40 is called * a stratum * of @xmath22 .
[ @xcite ] one says that the * whitney @xmath41 condition * is realized for a stratification if for each pair of strata @xmath42 and for any @xmath43 one has : let @xmath44 be a sequence of points in @xmath45 with limit @xmath46 and let @xmath47 be a sequence of points in @xmath48 tending to @xmath46 , assume that the sequence of tangent spaces @xmath49 admits a limit @xmath50 for @xmath51 tending to @xmath52 ( in a suitable grassmanian manifold ) and that the sequence of directions @xmath53 admits a limit @xmath54 for @xmath51 tending to @xmath52 ( in the corresponding projective manifold ) , then @xmath55 .
we denote by @xmath56 the open cone on the space @xmath57 , the cone on the empty set being a point .
observe that if @xmath57 is a stratified set then @xmath56 is stratified by the cones over the strata of @xmath57 and a @xmath58-dimensional stratum ( the vertex of the cone ) .
a stratification of @xmath22 is said to be * locally topologically trivial * if for every @xmath59 , @xmath60 , there is an open neighborhood @xmath61 of @xmath36 in @xmath22 , a stratified set @xmath57 and a semi - algebraic homeomorphism @xmath62 such that @xmath63 maps the strata of @xmath61 ( induced stratification ) onto the strata of @xmath64 ( product stratification ) .
[ del pseudomanifolds ] a * pseudomanifold * in @xmath23 is a subset @xmath65 whose singular locus is of codimension at least 2 in @xmath22 and whose regular locus is dense in @xmath22 .
a * stratified pseudomanifold * ( of dimension @xmath38 ) is the data of an @xmath38-dimensional pseudomanifold @xmath22 together with a semi - algebraic filtration : @xmath66 which constitutes a locally topologically trivial stratification of @xmath22 .
[ dfn_pseudo_a_bord ] a * stratified pseudomanifold with boundary * is a semi - algebraic couple @xmath67 together with a semi - algebraic filtration @xmath66 such that : 1 .
@xmath68 is an @xmath38-dimensional stratified pseudomanifold ( with the filtration @xmath69 ) , 2 .
@xmath70 is a stratified pseudomanifold ( with the filtration @xmath71 ) , 3 .
@xmath70 has a * stratified collared neighborhood * : there exist a neighborhood @xmath72 of @xmath70 in @xmath22 and a semi - algebraic homeomorphism @xmath73\to u $ ] such that @xmath74)=u \cap x_j$ ] and @xmath75 .
a * perversity * is an @xmath76-uple of integers @xmath77 such that @xmath78 and @xmath79 .
traditionally we denote the zero perversity by @xmath80 , the maximal perversity by @xmath81 , and the middle perversities by @xmath82)$ ] ( lower middle ) and @xmath83)$ ] ( upper middle ) .
we say that the perversities @xmath84 and @xmath85 are * complementary * if @xmath86 .
given a stratified pseudomanifold @xmath22 , we say that a semi - algebraic subset @xmath87 is @xmath88-*allowable * if @xmath89 for all @xmath90 .
in particular , a subset @xmath87 is @xmath91-allowable if @xmath92 .
define @xmath93 to be the @xmath25-vector subspace of @xmath24 consisting of those chains @xmath94 such that @xmath95 is @xmath96-allowable and @xmath97 is @xmath98-allowable .
the * @xmath99 intersection homology group with perversity @xmath84 * , denoted by @xmath100 , is the @xmath99 homology group of the chain complex @xmath101 goresky and macpherson proved that these groups are independent of the choice of the stratification and are finitely generated @xcite .
_ for any orientable compact stratified semi - algebraic @xmath38-dimensional pseudomanifold @xmath22 , generalized poincar duality holds : @xmath102 where @xmath84 and @xmath85 are complementary perversities .
_ in the non - compact case the above isomorphism holds for borel - moore homology : @xmath103 where @xmath104 denotes the intersection homology with respect to borel - moore chains @xcite .
a relative version is also true in the case where @xmath22 has boundary .
[ proinvariance ] _ let x be a locally compact stratified pseudomanifold and @xmath84 a perversity , then the intersection homology groups @xmath105 and @xmath106 do not depend on the stratification of @xmath22 . _
let @xmath108 be a smooth submanifold .
we say that a differential form @xmath109 on @xmath110 is @xmath107 if there exists a constant @xmath111 such that for any @xmath112 : @xmath113 we denote by @xmath114 the cochain complex constituted by all the @xmath115-forms @xmath109 such that @xmath116 and @xmath117 are both @xmath107 .
the cohomology groups of this cochain complex are called the * @xmath107-cohomology groups of @xmath110 * and will be denoted by @xmath118 .
the third author showed that the @xmath107 cohomology of the differential forms defined on the regular part of a pseudomanifold @xmath22 coincides with the intersection cohomology of @xmath22 in the maximal perversity ( @xcite , theorem 1.2.2 ) : [ thm_intro_linfty ] _ let @xmath22 be a compact subanalytic pseudomanifold ( possibly with boundary ) .
then , for any @xmath115 : @xmath119 furthermore , the isomorphism is induced by the natural mapping provided by integration on allowable simplices .
_ let @xmath7 be a polynomial mapping .
we denote by @xmath2 the set of points at which the mapping @xmath4 is not proper , _
i.e. _ @xmath120 and call it the * asymptotic variety * or * jelonek set * of @xmath4 .
the geometry of this set was studied by jelonek in a series of papers @xcite .
jelonek obtained a nice description of this set and gave an upper bound for its degree . for the details and applications of these results we refer to the works of jelonek . in our paper
, we will need the following powerful theorems .
[ @xcite ] [ jelonek1 ] _ if @xmath121 is a generically finite polynomial mapping , then @xmath2 is either an @xmath122 pure dimensional @xmath123-uniruled algebraic variety or the empty set . _
[ @xcite ] [ jelonek2 ] _ if @xmath124 is a dominant polynomial map of smooth affine varieties of the same dimension then @xmath2 is either empty or is a hypersurface . _ here , by a @xmath123-uniruled variety @xmath22 we mean that through any point of @xmath22 passes a rational complex curve included in @xmath22 . in other words
, @xmath22 is @xmath123-uniruled if for all @xmath125 there exists a non - constant polynomial mapping @xmath126 such that @xmath127 . in the real case ,
the jelonek set is an @xmath25-uniruled semi - algebraic set but , if nonempty , its dimension can be any integer between 1 and @xmath122 @xcite .
the variety @xmath8 was constructed by the second and third authors in @xcite .
let us recall briefly this construction
. we will consider polynomial mappings @xmath7 as real ones @xmath128 .
by @xmath129 we mean the singular locus of @xmath4 that is the zero set of its jacobian determinant and we denote by @xmath10 the set of critical values of @xmath4 , _
i.e. _ the set @xmath130 .
we denote by @xmath131 the euclidean riemannian metric in @xmath132 .
we can pull it back in a natural way : @xmath133 define the riemannian manifold @xmath134 and observe that the mapping @xmath4 induces a local isometry nearly any point of @xmath135 .
[ @xcite ] [ valette1 ] _ there exists a finite covering of @xmath135 by open semi - algebraic subsets such that on every element of this covering , the mapping @xmath4 induces a diffeomorphism onto its image . _
[ @xcite ] [ valette2 ] _ let @xmath121 be a polynomial mapping .
there exists a real semi - algebraic pseudomanifold @xmath5 , for some @xmath136 , where @xmath137 such that @xmath138 and there exists a semi - algebraic bi - lipschitz mapping @xmath139 where @xmath140 is equipped with the riemannian metric induced by @xmath141 _
the variety @xmath8 is constructed as follows : let @xmath121 be a polynomial mapping .
thanks to lemma [ valette1 ] , there exists a covering @xmath142 of @xmath143 by open semi - algebraic subsets ( in @xmath132 ) such that on every element of this covering , the mapping @xmath4 induces a diffeomorphism onto its image . we may find some semi - algebraic closed subsets @xmath144 ( in @xmath135 ) which cover @xmath135 as well . thanks to mostowski s separation lemma ( see separation lemma in @xcite , page 246 ) , for each @xmath20 , @xmath145 , there exists a nash function @xmath146 , such that @xmath147 is positive on @xmath148 and negative on @xmath149 .
we define @xmath150 in order to prove @xmath151 is bi - lipschitz , we do as follows : choose @xmath152 , then there exists @xmath153 such that @xmath154 and the mapping @xmath155 is a diffeomorphism onto its image .
define , for @xmath156 , the following functions : _ i(y):=_i ( f_|u_j)^-1(y ) , for @xmath157 , and ( y):=(y,(y), ,(y ) ) .
we then have the formula h_f(x)=(f(x),_1(f(x)),
,_p(f(x)))= ( f(x ) ) .
as the map @xmath158 is bi - lipschitz , it is enough to show that @xmath159 is bi - lipschitz .
this amounts to prove that @xmath160 has bounded derivatives for any @xmath161 . in order to prove this , we chose the functions @xmath147 sufficiently small , by using ojasiewicz inequality in the following form : @xcite [ lojasiewicz ] let @xmath30 be a closed semi - algebraic set and @xmath162 a continuous semi - algebraic function .
there exist @xmath163 and @xmath164 such that for any @xmath165 we have @xmath166 in fact , we can choose the nash functions @xmath147 sufficiently small by multiplying @xmath147 by a huge power of @xmath167 which is a nash function ( see proposition 2.3 in @xcite ) .
thanks to ojasiewicz inequality , we also can choose the functions @xmath147 such that they tend to zero at infinity and near @xmath168 .
this is the reason why the singular part of @xmath8 is contained in @xmath169 .
moreover , the following diagram is commutative : [ dia1 ] where @xmath170 is the canonical projection on the first @xmath171 coordinates , and @xmath151 is bijective onto its image @xmath172 .
remark that the set @xmath8 is not unique , it depends on the covering of @xmath135 that we choose and on the choice of the nash function @xmath147 .
we see that in the complex case , even in the case @xmath173 , the real dimension of the variety @xmath8 is greater than 3 , so it is difficult to draw the variety @xmath8 in this case .
the natural question arises if the variety @xmath8 exists in the real case .
the answer is yes , but we note that in this case , the variety @xmath8 is not necessarily a pseudomanifold , because in the real case , the real dimension of the jelonek set of a polynomial mapping @xmath174 can be @xmath175 .
[ @xcite , @xcite ] [ thuyvalette0 ] _ let @xmath176 be a polynomial mapping . there exist _
\a ) a real semi - algebraic variety @xmath5 , for some @xmath177 where @xmath178 , such that @xmath179 \b ) a semi - algebraic bi - lipschitz mapping @xmath139 where @xmath140 is equipped with the riemannian metric induced by @xmath141 in order to understand better the variety @xmath8 , we give here an example in the real case . @xcite [ exthuy1 ] let @xmath180 be the polynomial mapping defined by @xmath181 let us construct the variety @xmath8 in this case . by an easy computation ,
we find : @xmath182 @xmath183 @xmath184 we see that @xmath185 is divided into four open subsets @xmath186 by @xmath129 ( see the figure 1a ) .
the mapping @xmath4 is a diffeomorphism on each @xmath186 , for @xmath187 .
observe that @xmath186 is closed in @xmath135 so that we can chose @xmath188 for @xmath189 ( see section [ valette1 ] ) .
there exist nash functions @xmath190 such that each @xmath191 is positive on @xmath186 and negative on @xmath153 if @xmath192
. since @xmath8 is the closure of @xmath193 where @xmath194 , then @xmath8 have 4 parts @xmath195 where @xmath196 is the closure of @xmath197 for @xmath189 .
again , an easy computation shows : @xmath198 @xmath199 each @xmath196 is @xmath200 embedded in @xmath201 but@xmath196 does not lie in the plane @xmath202 anymore , it is lifted " in @xmath201 .
however , the part contained in @xmath203 still remains in the plane @xmath202 since the functions @xmath147 tend to zero at infinity and near @xmath129 ( see the figure 1b ) .
now we want to know how the parts @xmath196 are glued together . using diagram ( [ dia1 ] ) ,
for any point @xmath204 the cardinal of @xmath205 is equal to the cardinal of @xmath206 since @xmath151 is bijective .
consider now the equation @xmath207 where @xmath208 we have [ equaex ] ^2 y^2 + 2 ^ 2y - = 0 .
as the reduced discriminant is @xmath209 , then \1 ) if @xmath210 , the equation ( [ equaex ] ) does not have any solution , \2 ) if @xmath211 , the equation ( [ equaex ] ) has two solutions .
then @xmath212 and @xmath213 are glued together along @xmath214 .
similarly , @xmath215 and @xmath216 are glued together along @xmath214 ( see the figure 1c ) .
[ dessinex1thuy ] .,title="fig : " ] -6 cm -3truecm
3truecm 4truecm [ @xcite ] [ valette3 ] _ let @xmath217 be a polynomial mapping .
there exists a natural stratification of the variety @xmath8 , by even ( real ) dimension strata , which is locally topologically trivial along the strata . _ in fact
, the stratification of the variety @xmath8 is showed in @xcite to be @xmath218 where @xmath219 .
[ @xcite ] [ valette4 ] _ let @xmath217 be a polynomial mapping with nowhere vanishing jacobian .
the following conditions are equivalent : _
\(1 ) @xmath4 is non proper ,
\(2 ) @xmath220 , \(3 ) @xmath221 for any perversity @xmath84 , \(4 ) @xmath221 for some perversity @xmath84 .
we notice that , for a given polynomial map @xmath121 , whatever the choice of @xmath8 , the homology and intersection homology groups @xmath222 and @xmath223 are either @xmath58 or not @xmath58 .
the following theorem generalizes the lemma [ valette3 ] and shows existence of suitable stratifications of the set @xmath2 in the case of a polynomial mapping @xmath224 + thanks to whitney s theorem ( theorem 19.2 , lemma 19.3 , @xcite ) , the set @xmath235 of points @xmath236 at which the whitney @xmath41 condition fails is contained in a complex algebraic variety of complex dimension smaller than @xmath175 , so @xmath237 .
we will define a filtration @xmath238 of @xmath132 by algebraic varieties and compatible with @xmath2 : @xmath239 by decreasing induction on @xmath240 .
assume that @xmath241 has been constructed .
if @xmath242 then we put @xmath243 otherwise we denote @xmath244 and @xmath245 .
we put now , consider the filtration @xmath248 of @xmath8 @xmath249 where @xmath250 and @xmath170 is the canonical projection from @xmath8 to @xmath132 , on the first @xmath171 coordinates ( see diagram ( [ dia1 ] ) ) .
let @xmath251 .
we claim that @xmath252 is proper .
this is obvious if @xmath253 is empty .
if @xmath253 is not empty , suppose that there exists a sequence @xmath254 in @xmath255 such that @xmath256 goes to a point @xmath257 in @xmath253 .
we have to show that the sequence @xmath254 does not go to infinity .
since @xmath258 , where @xmath259 , we have @xmath260 . if @xmath261 tends to infinity then @xmath262 , which is a contradiction .
let @xmath22 be a connected component of @xmath263 , where @xmath264 .
we have @xmath265 .
we claim that either @xmath266 or @xmath267 assume that there exist @xmath268 but @xmath269 and @xmath270 .
then we have @xmath271 , where @xmath236 .
there exists a curve @xmath272 where @xmath273 and @xmath274 , such that @xmath275 and @xmath276 .
let us call @xmath277 the first point at which @xmath278 meets @xmath279 .
thus , we have @xmath280 whenever @xmath281 and @xmath282 is in @xmath283 , for @xmath284
. moreover , @xmath285 tends to @xmath286 and @xmath282 tends to infinity as @xmath287 tends to @xmath288 .
hence , @xmath289 , so @xmath290 is in @xmath291 , contradicting @xmath292 .
let us show that @xmath293 is a smooth manifold , for all @xmath20 . because @xmath294 is proper , the restriction of @xmath170 to @xmath295 is proper .
consequently , @xmath170 is a covering map on @xmath296 .
this implies that @xmath296 is a smooth manifold .
observe that in the case where @xmath297 is nonempty then it is included in @xmath298 , if @xmath299 , since every point of @xmath297 is a point of @xmath300 .
as @xmath170 is a covering map on @xmath301 , this implies that @xmath302 is open in @xmath303 .
let us prove that the filtration @xmath248 defines a whitney stratification : at first , we prove that the stratification @xmath238 is a whitney stratification .
if the stratum @xmath251 is not empty , then by ( [ filtrationsf ] ) , we have @xmath304 this shows that the stratification @xmath238 satisfies whitney conditions .
we denote @xmath305 @xmath306 we now prove that if @xmath307 then @xmath308 . if @xmath309 then @xmath310 is the identity and thus @xmath22 belongs to @xmath311 . otherwise , @xmath312 .
assume that @xmath265 .
this implies that @xmath313 .
this amounts to say that @xmath314 .
thus @xmath315 . therefore , to show that @xmath308 , we have to check that @xmath316 is open and closed in @xmath317 . as @xmath170 is a local diffeomorphism at any point @xmath36 of @xmath22 ,
the set @xmath316 is a manifold of dimension @xmath318 , which is open in @xmath253 .
let us show that it is closed in @xmath253 .
take a sequence @xmath319 such that @xmath320 tends to @xmath321 .
let @xmath322 be such that @xmath323 .
since @xmath170 is proper , @xmath324 does not tend to infinity . taking a subsequence if necessary
, we can assume that @xmath324 is convergent .
denote its limit by @xmath36 .
as @xmath325 , then the point @xmath36 can not be in @xmath22 and thus belongs to @xmath291 since @xmath22 is closed in @xmath326 .
this implies that @xmath327 , as required .
let us consider a pair of strata @xmath328 of the stratification @xmath248 such that @xmath329 and let us prove that @xmath328 satisfies the whitney @xmath41 condition .
that is clear if @xmath330 .
if none of them is included in @xmath279 , then , as @xmath170 is a local diffeomorphism and whiney @xmath41 condition is a @xmath331 invariant , this is also clear .
therefore , we can assume that @xmath332 and @xmath333 ( if @xmath334 , then @xmath335 meets @xmath279 at the points of @xmath336 and then @xmath337 ) .
set for simplicity @xmath338 .
as @xmath335 is open in @xmath339 , there exists a subanalytic open subset @xmath340 of @xmath8 such that @xmath341 let @xmath342 .
we have @xmath343 ( see diagram ( [ dia1 ] ) ) .
consequently , the function distance @xmath344 nowhere vanishes on @xmath345 . as @xmath345 is a closed subset of @xmath132 , by ojasiewicz inequalty , multiplying the @xmath147 s by a huge power of @xmath346
, we can assume that on @xmath345 , for every @xmath20 now , in order to check that whitney @xmath41 condition holds , we take @xmath322 and @xmath348 tending to @xmath349 .
assume that @xmath350 and @xmath351 exist , we have to check that @xmath352 is included in @xmath353 . for every @xmath38 , @xmath324 belongs to @xmath354 for some @xmath115 . extracting a subsequence if necessary
, we may assume that it lies in the same @xmath354 .
on @xmath153 , @xmath170 is invertible and its inverse is @xmath355 where @xmath356 ( see section [ constructionnf ] , see also proposition 2.3 in @xcite ) . on one hand , the sets @xmath316 and @xmath365 belong to @xmath311 , they satisfy the whitney @xmath41 condition .
as a matter of fact @xmath366 ( extracting a sequence if necessary , we may assume @xmath367 is convergent ) .
we have @xmath368 on the other hand , observe that @xmath369 multypling @xmath370 by a huge power of @xmath371 , we can assume that the first order partial derivatives of @xmath372 at @xmath373 tend to zero as @xmath38 goes to infinity . then @xmath374 tends to @xmath375 .
but since @xmath376 , @xmath377 .
we now generalize theorem [ valette4 ] .
firstly we notice that a polynomial map @xmath378 can be written @xmath379 where @xmath380 is the part of degree @xmath381 in @xmath13 .
let @xmath382 the highest degree in @xmath13 , the leading form @xmath383 of @xmath13 is defined as @xmath384 [ @xcite ] [ valette5 ] let @xmath395 be a @xmath115-cycle and let @xmath391 be a compact semi - algebraic family of sets with @xmath396 for any @xmath287 .
assume that @xmath397 bounds a @xmath398-chain in each @xmath393 , @xmath399 small enough .
then @xmath400 bounds a chain in @xmath401 .
[ @xcite ] [ lemmathuy1 ] _ let @xmath405 be a polynomial mapping and @xmath406 the zero locus of @xmath407 , where @xmath408 is the leading form of @xmath13 for @xmath409 . if @xmath22 is a subset of @xmath23 such that @xmath410 is bounded , then @xmath411 is a subset of @xmath412 _ by definition , @xmath411 is included in @xmath413 .
we prove now that @xmath411 is included in @xmath406 .
in fact , given @xmath414 , then there exists a semi - algebraic curve @xmath415 \rightarrow x $ ] such that @xmath416 and @xmath417 .
then @xmath418 can be written as @xmath419 and @xmath420 where @xmath421 is the homogeneous degree of @xmath408 .
since @xmath410 is bounded , then @xmath13 can not tend to infinity when @xmath287 tends to @xmath288 , hence @xmath422 for all @xmath409 .
@xmath423 : by goresky - macpherson poincar duality theorem , we have @xmath424 where @xmath85 is the complementary perversity of @xmath84 . since @xmath221 for all perversities @xmath84 , then @xmath425 , for all perversities @xmath426 @xmath431 : assume that @xmath4 is not proper .
that means that there exists a complex puiseux arc @xmath432 , @xmath433 ( with @xmath35 negative integer and @xmath290 is an unit vector of @xmath132 ) tending to infinity in such a way that @xmath434 converges to a generic point @xmath435 .
let @xmath436 be an oriented triangle in @xmath132 whose barycenter is the origin .
then , as the mapping @xmath437 ( where @xmath438 extends continuously at @xmath58 , it provides a singular @xmath439-simplex in @xmath8 that we will denote by @xmath26 .
the support of @xmath443 lies in @xmath444 . by definition of @xmath8
, we have @xmath445 . since @xmath446 , the chain @xmath447 bounds a singular chain @xmath448 .
so @xmath449 is a @xmath450-allowable cycle of @xmath8 . by definition , @xmath457 and @xmath458
are subsets of @xmath459 .
observe that , in a neighborhood of infinity , @xmath17 coincides with the support of the puiseux arc @xmath278 .
the set @xmath457 is equal to @xmath460 ( denoting the orbit of @xmath461 under the action of @xmath462 on @xmath463 , @xmath464 ) .
let @xmath406 be the zero locus of the leading forms @xmath465 .
since @xmath466 and @xmath467 are bounded , by lemma [ lemmathuy1 ] , @xmath457 and @xmath458 are subsets of @xmath468 for @xmath469 large enough , the sphere @xmath470 with center 0 and radius @xmath469 in @xmath132 is transverse to @xmath17 and @xmath471 ( at regular points ) .
let @xmath472 after a triangulation , the intersection @xmath473 is a chain bounding the chain @xmath474
. denote by @xmath493 the closure of @xmath494 , and set @xmath495 . since the strong deformation retraction @xmath131 is the identity on @xmath496 $ ] , we see that @xmath497 ) = \s^1 .
a \subset v \cap \s^{2n-1}(0,1).\ ] ] the class of @xmath503 in @xmath460 is , up to a product with a nonzero constant , equal to the generator of @xmath460 .
therefore , since @xmath503 bounds the chain @xmath504 , the cycle @xmath460 must bound a chain in @xmath499 as well . by lemma [ valette5 ] , this implies that @xmath460 bounds a chain in @xmath505 which is included in @xmath506 .
the set @xmath406 is a projective variety which is an union of cones in @xmath132 . since @xmath507
then @xmath508 , so @xmath509 and @xmath510 the cycle @xmath511 thus bounds a chain in @xmath512 , which is a finite union of circles . a contradiction | in the paper `` _ geometry of polynomial mapping at infinity via intersection homology _ '' the second and third authors associated to a given polynomial mapping @xmath0 with nonvanishing jacobian a variety whose homology or intersection homology describes the geometry of singularities at infinity of the mapping .
we generalize this result . |
o oscilador harmnico um dos mais importantes sistemas em mecnica quntica porque ele apresenta soluo em forma fechada e isto pode ser til para gerar solues aproximadas ou solues exatas para vrios problemas .
o oscilador harmnico costumeiramente resolvido com o mtodo de soluo em sries de potncias ( veja , e.g. , @xcite ) e o mtodo algbrico ( veja , e.g. , @xcite ) , e tambm por meio de tcnicas de integrao de trajetria ( veja , e.g. , @xcite ) .
recentemente , o oscilador harmnico unidimensional foi abordado com os mtodos operacionais da transformada de fourier @xcite e da transformada de laplace @xcite . a equao de schrdinger com um potencial quadrtico acrescido de um termo inversamente quadrtico , conhecido como oscilador harmnico singular , tambm um problema exatamente solvel @xcite-@xcite
. a bem da verdade , o problema geral de espalhamento e estados ligados em potenciais singulares um tema antigo ( veja , e.g. , @xcite ) .
o caso do oscilador harmnico singular com parmetros do potencial dependentes do tempo tm sido alvo de investigao recente @xcite .
o oscilador singular se presta para a construo de modelos solveis de @xmath0 corpos @xcite , tanto quanto como base para expanses perturbativas e anlise
variacional para osciladores harmnicos acrescidos de termos com singularidades muito mais fortes que o termo inversamente quadrtico @xcite .
o oscilador singular tambm tem sido utilizado em mecnica quntica relativstica @xcite
. a exata solubilidade do oscilador singular pode ser constatada nas referncias @xcite e @xcite para o caso tridimensional , e tambm patente nas referncias @xcite e @xcite para o caso unidimensional restrito ao semieixo positivo .
duas referncias mais recentes abordam o problema unidimensional em todo o eixo @xcite-@xcite .
@xcite no h detalhes da soluo do problema nem meno s possveis degenerescncias , e l consta que , para um oscilador singular atrativo , a partcula colapsa para o ponto @xmath1 ( _ the particle collapses to the point @xmath1 _ ) .
@xcite , palma e raff esmiuam o problema com o potencial singular repulsivo , concluem apropriadamente sobre a degenerescncia e simplesmente afirmam que no h estado fundamental no caso do oscilador singular atrativo ( _ the attractive potential has no lower energy bound _ ) .
o intuito deste trabalho perscrutar a equao de schrdinger unidimensional com o oscilador harmnico singular .
veremos que , alm de uma crtica literatura concernente a um problema de interesse recente e j cristalizado em livros - texto largamente conhecidos , a abordagem dos estados ligados do oscilador harmnico singular que se presencia neste trabalho permite aos estudantes de mecnica quntica e fsica matemtica dos cursos de graduao em fsica o contato com equaes diferenciais singulares e o comportamento assinttico de suas solues , funo hipergeomtrica confluente , polinmios de laguerre e de hermite e outras funes especiais , valor principal de cauchy de integrais imprprias , condio de hermiticidade sobre operadores associados com grandezas fsicas observveis e o descarte de solues esprias , condies de contorno e analiticidade das solues na vizinhana de pontos singulares , paridade e extenses simtricas e antissimtricas de autofunes , degenerescncia em sistemas unidimensionais , transio de fase e surgimento de nveis intrusos , _
et cetera_. seguramente , a profuso de conceitos e tcnicas do interesse de estudantes e instrutores .
com o critrio de hermiticidade dos operadores associados com quantidades fsicas observveis , o tratamento do problema pe vista dos leitores que o oscilador singular , seja atrativo ou repulsivo , exibe um nmero infinito de solues aceitveis desde que o parmetro responsvel pela singularidade seja maior que um certo valor crtico .
veremos que o espectro de energia uma funo montona do parmetro responsvel pela singularidade e que a energia do estado fundamental do oscilador singular , independentemente do sinal de tal parmetro , sempre maior que dois teros da energia do estado fundamental do oscilador no - singular para o problema definido no semieixo , e sempre maior que o dobro da energia do estado fundamental do oscilador no - singular para o problema definido em todo o eixo .
mostramos que o problema definido em todo o eixo nos conduz dupla degenerescncia no caso do potencial singular e intruso de adicionais nveis de energia no caso do oscilador no - singular ( relacionados com as autofunes de paridade par ) .
o robusto critrio de hermiticidade do operador associado com a energia cintica ( ou potencial ) mostra - se suficiente para descartar solues ilegtimas e permite demonstrar que se o potencial singular for fracamente atrativo no h cabimento em se falar em colapso para o centro ou inexistncia de estado fundamental .
finalmente , mostramos que , seja o problema definido no semieixo ou em todo o eixo , o oscilador no - singular pode ser pensado como uma transio de fase do oscilador singular , e por causa disso a soluo do oscilador harmnico singular no pode ser obtida a partir da soluo do oscilador harmnico no - singular via teoria da perturbao .
a equao de schrdinger unidimensional para uma partcula de massa de repouso @xmath2 sujeita a um potencial externo @xmath3 dada por@xmath4aqui , @xmath5 a funo de onda , @xmath6 a constante de planck reduzida ( @xmath7 ) e @xmath8 o operador hamiltoniano @xmath9no dificil mostrar que@xmath10 , \]]e levando em considerao que o operador hamiltoniano um operador hermitiano , dito ser hermitiano se@xmath11onde @xmath12 e @xmath13 so duas funes de onda quaisquer que fazem @xmath14 .
em particular , as funes de onda devem ser quadrado - integrveis , viz .
@xmath15 @xmath16 . ]
temos o corolrio@xmath17a equao da continuidade@xmath18 satisfeita com a densidade de probabilidade@xmath19e a corrente@xmath20a equao da continuidade pode tambm ser escrita como@xmath21onde @xmath22 um ponto arbitrrio do eixo @xmath23 .
a forma integral da equao da continuidade , ( [ ec ] ) , permite interpretar inequivocamente a corrente @xmath24 como sendo o fluxo de probabilidade atravs de @xmath25 no instante @xmath26 .
no caso de potenciais externos independentes do tempo , a funo de onda @xmath27 admite solues particulares da forma @xmath28 onde @xmath29 obedece equao de schrdinger independente do tempo@xmath30neste caso , com condies de contorno apropriadas , o problema se reduz determinao do par caracterstico @xmath31 . a equao de
autovalor ( [ auto ] ) tambm pode ser escrita na forma @xmath32 com@xmath33a densidade e a corrente correspondentes soluo expressa por ( [ 2a ] ) tornam - se @xmath34em virtude de @xmath35 e @xmath36 serem independentes do tempo , a soluo ( [ 2a ] ) dita descrever um estado estacionrio .
note que , por causa da equao da continuidade ( [ con ] ) e ( [ ec ] ) , a corrente @xmath36 no to somente estacionria , mas tambm uniforme , i.e. @xmath37 . os estados ligados constituem uma classe de solues da equao de schrdinger que representam um sistema localizado numa regio finita do espao .
para estados ligados devemos procurar autofunes que se anulam medida que @xmath38 .
bvio que , em decorrncia deste comportamento assinttico , @xmath39 quando @xmath38 .
assim , a uniformidade da corrente dos estados estacionrios demanda que @xmath36 seja nula em todo o espao .
fato esperado em vista da interpretao de @xmath36 apresentada anteriormente .
tambm , neste caso podemos normalizar @xmath29 fazendo@xmath40com um potencial invariante sob reflexo atravs da origem ( @xmath41 ) , autofunes com paridades bem definidas podem ser construdas .
neste caso , as autofunes de @xmath8 so tambm autofunes do operador paridade , viz.@xmath42onde @xmath43 o operador paridade , @xmath44 e @xmath45 denota quaisquer outros nmeros qunticos .
como consequncia da hemiticidade de @xmath8 e @xmath43 , as autofunes satisfazem condio de ortogonalidade@xmath46por causa da paridade , podemos concentrar a ateno no semieixo positivo e impor condies de contorno na origem e no infinito .
naturalmente , a autofuno contnua .
no entanto , temos de considerar a condio de conexo entre a derivada primeira da autofuno direita e esquerda da origem , que deve ser obtida diretamente da equao de schrdinger independente do tempo .
normalizabilidade , conforme comentado de pouco , requer que @xmath47 quando @xmath38 .
autofunes com paridades bem definidas em todo o eixo podem ser construdas tomando combinaes lineares simtricas e antissimtricas de @xmath48 definida no lado positivo do eixo @xmath23 .
estas novas autofunes possuem a mesma energia , ento , em princpio , existe uma dupla degenerescncia ( @xmath49 ) .
notrio que o espectro de estados ligados de sistemas unidimensionais com potenciais regulares
no - degenerado ( veja , e.g. , @xcite e @xcite ) .
entretanto , se o potencial for singular na origem , por exemplo , tanto as autofunes pares quanto as autofunes mpares poderiam obedecer condio homognea de dirichlet na origem , e cada nvel de energia exibiria uma degenerescncia de grau dois .. ] a condio de conexo obedecida pela derivada primeira da autofuno , contudo , poderia excluir uma das duas combinaes lineares , e nesse caso os nveis de energia seriam no - degenerados .
seguindo a notao da ref .
@xcite , vamos agora considerar o potencial@xmath50o parmetro adimensional @xmath51 caracteriza trs diferentes perfis para o potencial , como est ilustrado na figura 1 . as linhas tracejada , contnua e pontilhada para os casos com @xmath52 negativo , nulo e positivo , respectivamente . ,
title="fig:",width=377 ] [ fig : fig1 ] para @xmath53 temos o potencial do oscilador harmnico regular ( poo simples ) , e para @xmath54 temos o caso de um poo duplo com uma barreira de potencial repulsiva e singular na origem ( @xmath55 ) ou o caso de um poo sem fundo puramente atrativo ( @xmath56 ) .
para o potencial ( [ pot ] ) , a equao de schrdinger independente do tempo assume a forma @xmath57 com@xmath58 quando @xmath53 , as solues da equao ( [ af ] ) so analticas em todo o eixo @xmath23 .
quando @xmath54 , todavia , @xmath48 pode manifestar singularidade na origem .
tal singularidade poderia comprometer a nulidade da corrente em @xmath1 , a existncia das integrais definidas nos intervalos ( @xmath59 ) e ( @xmath60 ) , e a hermiticidade dos operadores associados com as quantidades fsicas observveis .
por causa desta sensao de ameaa , comearemos a abordagem do problema pela averiguao do comportamento das solues de ( [ af ] ) na vizinhana da origem .
claro que o comportamento assinttico ( @xmath38 ) tambm meritrio .
na vizinhana da origem a equao ( [ af ] ) passa a ter duas formas distintas : @xmath61 de um jeito ou de outro , no semieixo positivo podemos escrever@xmath62onde@xmath63 soluo da equao algbrica indicial@xmath64na vizinhana da origem , o comportamento dos termos @xmath65comina a hermiticidade dos operadores associados com a energia cintica e com o potencial .
para @xmath66 , podemos escrever@xmath67 , \notag\end{aligned}\]]e para @xmath68@xmath69 .
\notag\end{aligned}\]]em ambos os casos @xmath70 .
vemos destas ltimas relaes que a hermiticidade do operador associado com a energia cintica ( ou potencial ) verificada somente se @xmath71 quando @xmath54 , o que equivale a dizer que o sinal negativo defronte do radical em ( [ beta ] ) deve ser descartado e @xmath51 deve ser maior que @xmath72 .
naturalmente , devemos considerar @xmath73 tanto quanto @xmath74 quando @xmath53 .
portanto , podemos afirmar que @xmath29 comporta - se na vizinhana da origem como@xmath75com@xmath76note ainda que a nulidade da corrente requer apenas que @xmath77 e @xmath78 , enquanto a condio de ortonormalizabilidade exige apenas que @xmath79 e @xmath78 .
a condio @xmath80 , que nos faz evitar um potencial atrativo com singularidade muito forte , relacionado com o problema da queda para o centro @xcite , tanto quanto a soluo apropriada da equao indicial ( [ indi ] ) , foram obtidas aqui de uma maneira extremamente simples , sem recorrer ao processo de regularizao do potencial na origem .
substitudo por @xmath81 para @xmath82 e depois de usar as condies de continuidade para @xmath29 e @xmath83 no _ cutoff _ tomamos o limite @xmath84 .
resulta que a soluo com @xmath85 suprimida em relao a essa envolvendo @xmath86 quando @xmath87 . ]
nota - se que , ainda que acrescido da exigncia de normalizabilidade , o processo de regularizao do potencial inapto para exluir o caso com @xmath88 @xcite . o critrio de hermiticidade do operador associado com a energia cintica ( ou potencial ) lcito e suficiente para descartar solues esprias .
e @xmath89 na vizinhana da origem , respectivamente . ]
resumidamente,@xmath90a condio de dirichlet homognea ( @xmath91 ) essencial sempre que @xmath54 , contudo ela tambm ocorre para @xmath53 quando @xmath92 mas no para @xmath93 .
em suma,@xmath94e@xmath95
a equao de schrdinger independente do tempo para o nosso problema , eq .
( [ af ] ) , tem o comportamento assinttico ( @xmath96e da sucede que a forma assinttica da soluo quadrado - integrvel dada por @xmath97 o comportamento assinttico de @xmath29 expresso por ( [ asym ] ) convida - nos a definir @xmath98 de forma que a autofuno para todo @xmath99 pode ser escrita como@xmath100onde a funo desconhecida @xmath101 soluo regular da equao hipergeomtrica confluente @xcite @xmath102com@xmath103a soluo geral de ( [ kum ] ) dada por @xcite @xmath104onde@xmath105 e @xmath106 so constantes arbitrrias , e @xmath107 , tambm denotada por @xmath108 , a funo hipergeomtrica confluente ( tambm chamada de funo de kummer ) expressa pela srie @xcite@xmath109onde @xmath110 a funo gama . a funo gama no tem razes e seus polos so dados por @xmath111 , onde @xmath45 um inteiro no - negativo @xcite .
a funo de kummer converge para todo @xmath99 , regular na origem ( @xmath112 ) e tem o comportamento assinttico prescrito por @xcite@xmath113haja vista que @xmath114 ,
e estamos em busca de soluo regular na origem , devemos tomar@xmath115em ( [ sg ] ) . a presena de
@xmath116 em ( [ asy ] ) deprava o
bom comportamento assinttico da autofuno j ditado por ( asym ) .
para remediar esta situao constrangedora devemos considerar os polos de @xmath117 , e assim preceituar que um comportamento aceitvel para @xmath107 ocorre somente se@xmath118neste caso , a srie ( [ ser ] ) truncada em @xmath119 e o polinmio de grau @xmath45 resultante proporcional ao polinmio de laguerre generalizado @xmath120 , com @xmath121 abr .
portanto , de ( [ k2 ] ) , ( [ l ] ) , ( [ ab ] ) e ( [ n ] ) podemos determinar que as energias permitidas so dadas por @xmath122e as autofunes definidas no semieixo positivo so@xmath123 na figura 2 ) em funo de @xmath124 no intervalo @xmath125 $ ] . as linhas contnuas para o oscilador singular , e os quadrados em @xmath126 para o oscilador no - singular.,title="fig:",width=377 ] [ fig : fig2 ] ilustramos os primeiros nveis de energia em funo de @xmath51 no intervalo @xmath125 $ ] .
observe que , qualquer que seja @xmath51 , ainda que @xmath127 , o espectro discreto e sempre positivo .
h um nmero infinito de nveis de energia igualmente espaados ( espaamento igual a @xmath128 , independentemente de @xmath51 ) , e o estado fundamental tem energia @xmath129 .
na figura 3 .
as linhas contnua , ponto - tracejada , tracejada e pontilhada para os casos com @xmath52 igual a @xmath130 , @xmath131 , @xmath132 e @xmath133 , respectivamente.,title="fig:",width=377 ] [ fig : fig3 ] ilustramos o comportamento da autofuno para o estado fundamental .
a normalizao foi realizada por mtodos numricos mas poderia ter sido obtida por meio de frmulas envolvendo os polinmios de laguerre associados constantes na ref .
@xcite . a comparao entre as quatro curvas mostra que a partcula tende a evitar a origem mais e mais medida que @xmath51 aumenta .
eis um resultado esperado no caso @xmath55 que ocorre tambm no caso @xmath56 .
observe que , qualquer que seja @xmath134 , as autofunes ( afu ) so fisicamente aceitveis , ainda que no intervalo@xmath135 elas possuam derivada primeira singular .
mesmo assim , contanto que o parmetro @xmath51 seja maior que @xmath72 , o par caracterstico @xmath136 constitui uma soluo permissvel do problema proposto . a partcula nunca colapsa para o ponto @xmath1 e certamente h
um estado fundamental . a autofuno definida para todo o eixo @xmath23 pode ser escrita como@xmath137onde@xmath138 a funo degrau de heaviside , @xmath139 o autovalor do operador paridade , e a autoenergia dada por@xmath140a hermiticidade do operador associado com a energia cintica ( ou potencial ) , por causa da singularidade em @xmath1 no caso @xmath56 ( @xmath141 ) , depende da existncia do valor principal de cauchy for singular na origem , a integral @xmath142 ser _
nonsense_. contudo , o valor principal de cauchy , @xmath143 , uma prescrio que pode atribuir um sentido proveitoso representao integral por meio da receita que se segue : @xmath144 ] da integral@xmath145obviamente , o valor principal de cauchy poderia consentir um afrouxamento das condies de contorno impostas sobre as autofunes .
autofunes mais singulares que essas anteriormente definidas no semieixo seriam toleradas se na vizinhana da origem os sinais de @xmath146 direita e esquerda da origem fossem diferentes para quaisquer @xmath139 e @xmath147 .
porm , temos@xmath148de modo que , no caso em que @xmath56 , a integral ( [ intk ] ) no seria finita para @xmath149 .
somos assim conduzidos a preservar a rigidez do critrio de hermiticidade j estabelecido no problema definido no semieixo .
a continuidade ( ou descontinuidade ) de @xmath83 na origem pode ser avaliada pela integrao de ( [ orig ] ) de @xmath150 para @xmath151 no limite @xmath152 . a frmula de conexo entre @xmath83 direita e @xmath83 esquerda da origem pode ser sumarizada por@xmath153tomando em considerao o valor principal de cauchy , no caso de @xmath154 antissimtrica com @xmath155 , podemos afirmar que as autofunes tm derivada primeira contnua na origem .
assim , o espectro do oscilador singular duplamente degenerado .
por causa da continuidade , no h autofuno mpar para @xmath93 e l ocorre a condio homognea de neumann ( @xmath156 ) , e no h autofuno par para @xmath92 .
portanto , o espectro do oscilador regular no - degenerado , como deveria ser .
e @xmath92 , em particular , os polinmios de laguerre @xmath157 e @xmath158 so proporcionais aos polinmios de hermite @xmath159 , respectivamente @xcite:@xmath160os polinmios de hermite so definidos no intervalo @xmath161 e gozam da propriedade @xmath162 .
assim , a soluo do oscilador no - singular pode ser escrita na forma que se costumou em termos dos polinmios de hermite:@xmath163com autofuno definida em todo o eixo expressa por@xmath164as constantes @xmath165 em ( [ afu ] ) e ( [ psi2 ] ) , chamadas de constantes de normalizao , podem ser determinadas por meio da condio de normalizao expressa por ( [ normaliza ] ) .
na figura 4 ) em funo de @xmath124 no intervalo @xmath125 $ ] .
as linhas contnuas para o oscilador singular , os quadrados e os crculos em @xmath166 para as solues mpares e pares do oscilador harmnico no - singular , respectivamente.,title="fig:",width=377 ] [ fig : fig4 ] ilustramos os primeiros nveis de energia em funo de @xmath51 no intervalo @xmath125 $ ] para o problema definido em todo o eixo .
para @xmath167 , o espectro exatamente igual a esse do problema definido no semieixo .
quando a singularidade nula , entanto , o espaamento dos nveis @xmath168 .
esta mudana de espaamento de nveis quando @xmath53 devida aos nveis intrusos que surgem por causa emergncia da condio de contorno homognea de neumann , em adio condio de contorno homognea de dirichlet j existente no problema definido no semieixo .
esta invaso de novas solues com @xmath169 tem como consequncia imediata um drstico efeito sobre a localizao da partcula .
os resultados apresentados neste trabalho mostram com transparncia que o oscilador harmnico singular unidimensional , seja definido no semieixo ou em todo o eixo , seja repulsivo ou atrativo , exibe um conjunto infinito de solues aceitveis , em claro contraste com os ditames estampados nas referncias @xcite e @xcite , e transcritos na introduo .
sim , para um potencial singular fracamente atrativo o colapso para o centro no tem nada a ver , e certamente h um estado fundamental com energia maior que @xmath170 .
a generalizao dos resultados do problema definido no semieixo para o oscilador harmnico singular tridimensional pode ser feita com facilidade por meio da substituio de @xmath51 por @xmath171 , onde @xmath172 o numero quntico orbital , e pela concomitante substituio de @xmath29 pela funo radial @xmath173 .
para o problema definido no semieixo , a transio de fase manifesta - se apenas por meio do comportamento da derivada primeira da autofuno na origem .
derivada primeira infinita para @xmath56 , constante para @xmath53 , e nula para @xmath55 .
para o problema definido em todo o eixo , entretanto , a transio de fase manifesta - se por meio da degenerescncia , pelo comportamento da autofuno e sua derivada primeira na origem , e pela localizao da partcula .
uma outra assinatura da transio de fase o espaamento dos nveis de energia .
quando o potencial singular na origem , as autoenergias so igualmente espaadas com passo igual a @xmath128 .
admirvel que esse espaamento de nveis independente da intensidade do parmetro responsvel pela singularidade do potencial .
quando a singularidade nula , entanto , o espaamento dos nveis @xmath168 .
esta brusca mudana de espaamento de nveis quando @xmath51 passa por @xmath53 devida aos nveis intrusos que surgem por causa emergncia da condio de contorno homognea de neumann , em adio condio de contorno homognea de dirichlet j existente para @xmath54 .
esta intruso permite o surgimento de polinmios de hermite pares e seus autovalores associados , que se entremeiam entre os autovalores pr - existentes associados com os polinmios de hermite mpares .
os polinmios de hermite pares tm @xmath174 e esta condio de contorno nunca permitida quando a singularidade est presente , ainda que @xmath51 seja muito pequeno .
esta invaso sbita dos polinmios pares tem um brusco efeito sobre a localizao da partcula .
poder - se - ia tambm tentar compreender tal transio sbita partindo de um potencial no - singular ( @xmath53 ) , quando a soluo do problema envolve os polinmios de hermite pares e mpares , e ento adicionar o potencial singular como uma perturbao do potencial com @xmath53 .
agora , por sua natureza o potencial singular perturbativorepulsivo , demanda que @xmath175 e assim ele mata naturalmente a soluo envolvendo os polinmios de hermite pares .
ademais , no h degenerescncia no espectro para o caso de @xmath53 .
lathouwers @xcite considerou o caso unidimensional do oscilador harmnico singular como o oscilador harmnico no - singular perturbado .
acontece que , por causa dos distintos comportamentos da derivada primeira da autofuno na origem para @xmath53 ( @xmath176 ) e @xmath54 ( @xmath177 para @xmath56 , e @xmath178 para @xmath55 ) , nossos comentrios finais desfavorecem tal aspirao , ainda que haja continuidade do espectro na vizinhana de @xmath53 no caso associado com autofunes mpares do oscilador no - singular . k.m .
case , phys .
rev . * 80 * , 797 ( 1950 ) ; f.l .
scarf , phys .
* 109 * , 2170 ( 1958 ) ; a. pais e t.t .
wu , phys .
rev . * 134 * , b1303 ( 1964 ) ; w.m .
frank , d.j .
land e r.m .
spector , rev . mod
* 43 * , 36 ( 1971 ) .
p. camiz , et al .
, j. math phys .
* 12 * , 2040 ( 1971 ) ; v.v .
dodonov , i.a .
malkin , e v.i .
manko , phys .
a * 39 * , 377 ( 1972 ) ; v.v .
dodonov , v.i .
manko e l. rosa , phys .
a * 57 * , 2851 ( 1998 ) ; j.r .
choi , j. korean phys .
soc . * 44 * , 223 ( 2004 ) .
hall , n. saad , e a. von keviczky , j. math phys . * 39 * , 6345 ( 1998 ) ; r.l .
hall e n. saad , j. phys .
a * 33 * , 5531 ( 2000 ) ; r.l .
hall e n. saad , j. phys .
a * 33 * , 569 ( 2000 ) ; r.l .
hall e n. saad , j. phys .
a * 34 * , 1169 ( 2001 ) ; r.l . hall , n. saad , e a. von keviczky , j. math phys . * 43 * , 94 ( 2002 ) ; r.l .
hall , n. saad , e a. von keviczky , j. phys .
a * 36 * , 487 ( 2003 ) .
nagiyev , e.i .
jafarov , e r.m .
imanov , j. phys . a * 36 * , 7813 ( 2003 ) ; a.s .
de castro , ann .
( n.y . ) * 311 * , 170 ( 2004 ) ; a. de souza dutra e c .- s
. jia , phys .
a * 352 * , 484 ( 2006 ) ; t.r .
cardoso , l.b .
castro , e a.s . de castro , j. phys .
a * 45 * , 075302 ( 2012 ) . | investiga - se a equao de schrdinger unidimensional com o oscilador harmnico singular .
a hermiticidade dos operadores associados com quantidades fsicas observveis usada como critrio para mostrar que o oscilador singular atrativo ou repulsivo exibe um nmero infinito de solues aceitveis , contanto que o parmetro responsvel pela singularidade seja maior que um certo valor crtico , em discordncia com a literatura . o problema definido em todo o eixo exibe dupla degenerescncia no caso do oscilador singular e intruso de adicionais nveis de energia no caso do oscilador no - singular .
outrossim , mostra - se que a soluo do oscilador singular no pode ser obtida a partir da soluo do oscilador no - singular via teoria da perturbao .
* palavras - chave : * oscilador harmnico , potencial singular , degenerescncia , colapso para o centro |
measure - valued markov chains , or more generally measure - valued markov processes , arise naturally in modeling the composition of evolving populations and play an important role in a variety of research areas such as population genetics and bioinformatics ( see , e.g. , @xcite ) , bayesian nonparametrics @xcite , combinatorics @xcite and statistical physics @xcite . in particular , in bayesian nonparametrics there has been interest in measure - valued markov chains since the seminal paper by @xcite , where the law of the dirichlet process has been characterized as the unique invariant measure of a certain measure - valued markov chain . in order to introduce the result by @xcite , let us consider a polish space @xmath0 endowed with the borel @xmath1-field @xmath2 and let @xmath3 be the space of probability measures on @xmath0 with the @xmath1-field @xmath4 generated by the topology of weak convergence . if @xmath5 is a strictly positive finite measure on @xmath0 with total mass @xmath6 , @xmath7 is a @xmath0-valued random variable ( r.v . ) distributed according to @xmath8 and @xmath9 is a r.v .
independent of @xmath7 and distributed according to a beta distribution with parameter @xmath10 then , theorem 3.4 in @xcite implies that a dirichlet process @xmath11 on @xmath0 with parameter @xmath5 uniquely satisfies the distributional equation @xmath12 where all the random elements on the right - hand side of ( [ ftchain1 ] ) are independent .
all the r.v.s introduced in this paper are meant to be assigned on a probability space @xmath13 unless otherwise stated . in @xcite , ( [ ftchain1 ] ) is recognized as the distributional equation for the unique invariant measure of a measure - valued markov chain @xmath14 defined via the recursive identity @xmath15 where @xmath16 is arbitrary , @xmath17 is a sequence of @xmath0-valued r.v.s independent and identically distributed as @xmath7 and @xmath18 is a sequence of r.v.s , independent and identically distributed as @xmath9 and independent of @xmath17 .
we term @xmath14 as the feigin
tweedie markov chain . by investigating the functional markov chain , with @xmath19 for any @xmath20 and for
any measurable linear function @xmath21 , @xcite provide properties of the corresponding linear functional of a dirichlet process .
in particular , the existence of the linear functional @xmath22 of the dirichlet process @xmath11 is characterized according to the condition @xmath23 ; these functionals were considered by @xcite and their existence was also investigated by @xcite who referred to them as moments , as well as by @xcite and @xcite .
further developments of the linear functional markov chain @xmath24 are provided by @xcite and more recently by @xcite .
starting from the distributional equation ( [ ftchain1 ] ) , a constructive definition of the dirichlet process has been proposed by @xcite . if @xmath11 is a dirichlet process on @xmath0 with parameter @xmath25 , then @xmath26 where @xmath27 is a sequence of independent r.v.s identically distributed according to @xmath28 and @xmath29 is a sequence of r.v.s independent of @xmath27 and derived by the so - called stick breaking construction , that is , @xmath30 and @xmath31 for @xmath32 , with @xmath33 being a sequence of independent r.v.s identically distributed according to a beta distribution with parameter @xmath10 .
then , equation ( [ ftchain1 ] ) arises by considering @xmath34 where now @xmath35 and @xmath36 for @xmath37 .
thus , it is easy to see that @xmath38 is also a dirichlet process on @xmath0 with parameter @xmath5 and it is independent of the pairs of r.v.s @xmath39 .
if we would extend this idea to @xmath40 initial samples , we should consider writing @xmath41 where @xmath42 and @xmath43 is a dirichlet process on @xmath0 with parameter @xmath5 independent of the random vectors @xmath44 and @xmath45 .
however , this is not an easy extension since the distribution of @xmath9 is unclear , and moreover @xmath9 and @xmath46 are not independent . for this reason , in @xcite an alternative distributional equation has been introduced .
let @xmath5 be a strictly positive finite measure on @xmath0 with total mass @xmath6 and let @xmath47 be a @xmath0-valued plya sequence with parameter @xmath5 ( see @xcite ) , that is , @xmath47 is a sequence of @xmath0-valued r.v.s characterized by the following predictive distributions @xmath48 and @xmath49 , for any @xmath50 .
the sequence @xmath47 is exchangeable , that is , for any @xmath51 and any permutation @xmath1 of the indexes @xmath52 , the law of the r.v.s @xmath53 and @xmath54 coincide ; in particular , according to the celebrated de finetti representation theorem , the plya sequence @xmath47 is characterized by a so - called de finetti measure , which is the law of a dirichlet process on @xmath0 with parameter @xmath5 . for a fixed integer @xmath55 , let @xmath56 be a random vector distributed according to the dirichlet distribution with parameter @xmath57 , @xmath58 , and let @xmath9 be a r.v . distributed according to a beta distribution with parameter @xmath59 such that @xmath27 , @xmath60 and @xmath9 are mutually independent .
moving from such a collection of random elements , lemma 1 in @xcite implies that a dirichlet process @xmath61 on @xmath0 with parameter @xmath5 uniquely satisfy the distributional equation @xmath62 where all the random elements on the right - hand side of ( [ ftchain3 ] ) are independent . in order to emphasize the additional parameter @xmath40 , we used an upper - script @xmath63 on the dirichlet process @xmath11 and on the random vector @xmath64 .
it can be easily checked that equation ( [ ftchain3 ] ) generalizes ( [ ftchain1 ] ) , which can be recovered by setting @xmath65 . in the present paper ,
our aim is to further investigate the distributional equation ( [ ftchain3 ] ) and its implications in bayesian nonparametrics theory and methods .
the first part of the paper is devoted to investigate the random element @xmath66 in ( [ ftchain3 ] ) which is recognized to be the random probability measure ( r.p.m . ) at the @xmath40th step of a measure - valued markov chain defined via the recursive identity @xmath67 where @xmath68 is a sequence of independent r.v.s , each @xmath69 distributed according a beta distribution with parameter @xmath70 , @xmath71 for @xmath72 and @xmath55 and the sequence @xmath68 is independent from @xmath73 .
more generally , we observe that the measure - valued markov chain defined via the recursive identity ( [ nuovacatena ] ) can be extended by considering , instead of a plya sequence @xmath73 with parameter @xmath5 , any exchangeable sequence @xmath74 characterized by some de finetti measure on @xmath3 and such that @xmath68 is independent from @xmath74 .
asymptotic properties for this new class of measure - valued markov chains are derived and some linkages to bayesian nonparametric mixture modelling are discussed . in particular , we remark how it is closely related to a well - known recursive algorithm introduced in @xcite for estimating the underlying mixing distribution in mixture models , the so - called newton s algorithm . in the second part of the paper , by using finite and asymptotic properties of the r.p.m .
@xmath66 and by following the original idea of @xcite , we define and investigate from ( [ ftchain3 ] ) a class of measure - valued markov chain @xmath75 which generalizes the feigin tweedie markov chain , introducing a fixed integer parameter @xmath40 .
our aim is in providing features of the markov chain @xmath75 in order to verify if it preserves some of the properties characterizing the feigin
tweedie markov chain ; furthermore , we are interested in analyzing asymptotic ( as @xmath76 goes to @xmath77 ) properties of the associated linear functional markov chain @xmath78 with @xmath79 for any @xmath20 and for any function @xmath21 such that @xmath80 . in particular , we show that the feigin tweedie markov chain @xmath14 sits in a larger class of measure - valued markov chains @xmath75 parametrized by an integer number @xmath40 and still having the law of a dirichlet process with parameter @xmath5 as unique invariant measure .
the role of the further parameter @xmath40 is discussed in terms of new potential applications of the markov chain @xmath75 with respect to the the known applications of the feigin tweedie markov chain . following these guidelines , in section [ sec2 ] we introduce a new class of measure - valued markov chains @xmath81 defined via exchangeable sequences of r.v.s ; asymptotic results for @xmath81 are derived and applications related to bayesian nonparametric mixture modelling are discussed . in section [ secthree ] , we show that the feigin tweedie markov chain @xmath14 sits in a larger class of measure - valued markov chains @xmath75 , which is investigated in comparison with @xmath14 . in section [ sec4 ] , some concluding remarks and future research lines are presented .
let @xmath68 be a sequence of independent r.v.s such that @xmath82 almost surely and @xmath69 has beta distribution with parameter @xmath70 for @xmath83 . moreover , let be a sequence of @xmath0-valued exchangeable r.v.s independent from @xmath68 and characterized by some de finetti measure on @xmath3 .
let us consider the measure - valued markov chain @xmath81 defined via the recursive identity @xmath84 in the next theorem , we provide an alternative representation of @xmath85 and show that @xmath86 converges weakly to some limit probability @xmath87 for almost all @xmath88 , _ that is _
, for each @xmath89 in some set @xmath90 with @xmath91 .
in short , we use notation @xmath92 a.s.-@xmath93 .
[ asconv ] let @xmath94 be the markov chain defined by ( [ eqprec ] ) . then : a. an equivalent representation of @xmath95 , @xmath96 is @xmath97 where @xmath98 , @xmath99 has dirichlet distribution with parameter @xmath100 , and @xmath101 and @xmath74 are independent .
b. there exists a r.p.m .
@xmath102 on @xmath13 such that , as @xmath103 , @xmath104 where the law of @xmath102 is the de finetti measure of the sequence @xmath74 . as far as ( i ) is concerned , by repeated application of the recursive identity ( [ eqprec ] )
, it can be checked that , for any @xmath55 , @xmath105 where @xmath106 almost surely and @xmath107 is defined to be 1 when @xmath108 . defining @xmath109 , @xmath72 , it is straightforward to show that @xmath99 has the dirichlet distribution with parameter @xmath100 and @xmath58 , so that ( [ eqqrepr ] ) holds .
regarding ( ii ) , by the definition of the dirichlet distribution , an equivalent representation of ( [ eqqrepr ] ) is @xmath110 where @xmath111 is a sequence of r.v.s independent and identically distributed according to standard exponential distribution , independent from @xmath74 .
let @xmath112 be any bounded continuous function , and consider @xmath113 the expression in the denominator converges almost surely to 1 by the strong law of large numbers .
as far as the numerator is concerned , let @xmath102 be the r.p.m .
defined on @xmath13 , such that the r.v.s @xmath114 are independent and identically distributed conditionally on @xmath102 ; the existence of such a random element is guaranteed by the de finetti representation theorem ( see , e.g. , @xcite , theorem 1.49 ) .
it can be shown that @xmath115 is a sequence of exchangeable r.v.s and , if @xmath116 , @xmath117 where @xmath118 , @xmath88 , @xmath119 , is a r.p.m . with trajectories in @xmath120 , and
@xmath121 denotes the random distribution relative to @xmath122 .
this means that , conditionally on @xmath122 , @xmath115 is a sequence of r.v.s independent and identically distributed according to the random distribution ( evaluated in @xmath123 ) @xmath124 of course , @xmath125 since @xmath126 is bounded .
as in @xcite , example 7.3.1 , this condition implies @xmath127 so that @xmath128 a.s.-@xmath93 . by theorem 2.2 in @xcite
, it follows that @xmath92 a.s.-@xmath93 as @xmath129 . throughout the paper
, @xmath5 denotes a strictly positive and finite measure on @xmath0 with total mass @xmath130 , unless otherwise stated .
if the exchangeable sequence @xmath74 is the plya sequence with parameter @xmath5 , then by theorem [ asconv](i ) @xmath81 is the markov chain defined via the recursive identity ( [ nuovacatena ] ) ; in particular , by theorem [ asconv](ii ) , @xmath92 @xmath131 where @xmath102 is a dirichlet process on @xmath0 with parameter @xmath5 .
this means that , for any fixed integer @xmath55 , the r.p.m .
@xmath85 can be interpreted as an approximation of a dirichlet process with parameter @xmath5 . in appendix
[ appa1 ] , we present an alternative proof of the weak convergence ( convergence of the finite dimensional distribution ) of @xmath81 to a dirichlet process on @xmath0 with parameter @xmath5 , using a combinatorial technique . as a byproduct of this proof , we obtain an explicit expression for the moment of order @xmath132 ) of the @xmath133-dimensional plya distribution
. a straightforward generalization of the markov chain @xmath134 can be obtained by considering a nonparametric hierarchical mixture model .
let @xmath135 be a kernel , that is , @xmath136 is a measurable function such that @xmath137 is a density with respect to some @xmath1-finite measure @xmath138 on @xmath0 , for any fixed @xmath139 , where @xmath140 is a polish space ( with the usual borel @xmath1-field ) .
let @xmath81 be the markov chain defined via ( [ eqprec ] ) .
then for each @xmath141 we introduce a real - valued markov chain @xmath142 defined via the recursive identity @xmath143 where @xmath144 by a straightforward application of theorem 2.2 in @xcite , for any fixed @xmath141 , when @xmath145 is continuous for all @xmath141 and bounded by a function @xmath146 , as @xmath103 , then @xmath147 where @xmath102 is the limit r.p.m . in theorem [ asconv ] .
for instance , if @xmath102 is a dirichlet process on @xmath0 with parameter @xmath5 , @xmath148 is precisely the density in the dirichlet process mixture model introduced by @xcite .
when @xmath146 is a @xmath138-integrable function , not only the limit @xmath149 is a random density , but a stronger result than ( [ convdens ] ) is achieved .
[ rigo ] if @xmath145 is continuous for all @xmath141 and bounded by a @xmath138-integrable function @xmath146 , then @xmath150 where @xmath102 is the limit r.p.m . in theorem [ asconv ] . the functions @xmath151 and @xmath152 , defined on @xmath153 , are @xmath154-measurable , by a monotone class argument . in fact , by kernel s definition , @xmath155 is @xmath156-measurable .
moreover , if @xmath157 , @xmath158 and @xmath159 , then @xmath160 is @xmath154-measurable .
let @xmath161 .
since @xmath162 contains the rectangles , it contains the field generated by rectangles , and , since @xmath162 is a monotone class , @xmath163 .
the assertion holds for @xmath148 of the form @xmath164 since there exist a sequence of simple function on rectangles which converges pointwise to @xmath133 .
therefore , @xmath165 does not converge to @xmath166 .
then , by fubini s theorem , @xmath167 hence , @xmath168 where @xmath169 is the set of @xmath89 such that @xmath170 does not converge to @xmath171 . for any @xmath89 fixed in @xmath169 , it holds @xmath172 , @xmath138-a.e .
, so that by the scheff s theorem we have @xmath173 the theorem follows since @xmath168 .
we conclude this section by remarking an interesting linkage between the markov chain @xmath81 and the so - called newton s algorithm , originally introduced in @xcite for estimating the mixing density when a finite sample is available from the corresponding mixture model .
see also see also @xcite and @xcite .
briefly , suppose that @xmath174 are @xmath40 r.v.s independent and identically distributed according to the density function @xmath175 where @xmath136 is a known kernel dominated by a @xmath1-finite measure @xmath138 on @xmath0 ; assume that the mixing distribution @xmath176 is absolutely continuous with respect to some @xmath1-finite measure @xmath177 on @xmath140 .
@xcite proposed to estimate @xmath178 as follows : fix an initial estimate @xmath179 and a sequence of weights @xmath180 .
given @xmath174 independent and identically distributed observations from @xmath181 , compute @xmath182 for @xmath183 and produce @xmath184 as the final estimate .
we refer to @xcite , and @xcite for a recent wider investigation of the newton s algorithm . here
we show how the newton s algorithm is connected to the measure - valued markov chain @xmath81 .
let us consider @xmath40 observations from the nonparametric hierarchical mixture model , that is , @xmath185 and @xmath186 where @xmath102 is a r.p.m .
if we observed @xmath187 , then by virtue of ( ii ) in theorem [ asconv ] , we could construct a sequence of distributions @xmath188 for estimating the limit r.p.m .
@xmath102 , where @xmath189 is a sequence of independent r.v.s such that @xmath82 almost surely and @xmath190 has beta distribution with parameters @xmath191 .
this approximating sequence is precisely the sequence ( [ eqprec ] ) .
therefore , taking the expectation of both sides of the previous recursive equation , and defining @xmath192 $ ] , @xmath193=1/i$ ] , we have @xmath194 which can represent a predictive distribution for @xmath195 , and hence an estimate for @xmath102 .
however , instead of observing the sequence @xmath196 , it is actually the sequence @xmath197 which is observed ; in particular , we can assume that @xmath174 are @xmath40 r.v.s independent and identically distributed according to the density function ( [ mixmod ] ) .
therefore , instead of ( [ eqqtilde ] ) , we consider @xmath198 where @xmath199 in ( [ eqqtilde ] ) has been substituted ( or estimated , if you prefer ) by @xmath200 . finally , observe that , if @xmath201 is absolutely continuous , with respect to some @xmath1-finite measure @xmath177 on @xmath140 , with density @xmath202 , for @xmath72 , then we can write @xmath203 which is precisely a recursive estimator of a mixing distribution proposed by @xcite when the weights are fixed to be @xmath204 for @xmath72 and the initial estimate is @xmath205 $ ] .
in this section our aim is to define and investigate a class of measure - valued markov chain which generalizes the feigin
tweedie markov chain introducing a fixed integer parameter @xmath40 , and still has the law of a dirichlet process with parameter @xmath5 as the unique invariant measure .
the starting point is the distributional equation ( [ ftchain3 ] ) introduced by @xcite ; see appendix [ appa2 ] for an alternative proof of the solution of the distributional equation ( [ ftchain3 ] ) .
all the proofs of theorems in this section are in appendix [ appa3 ] for the ease of reading . for a fixed integer @xmath55 ,
let @xmath206 be a sequence of independent r.v.s with beta distribution with parameter @xmath59 , @xmath207 , with @xmath208 for any @xmath209 , be a sequence of independent r.v.s identically distributed according to a dirichlet distribution with parameter @xmath57 and @xmath210 , @xmath211 be sequence of independent r.v.s from a plya sequence with parameter @xmath5 . moving from such collection of random elements , for each fixed integer @xmath55 we define the measure - valued markov chain @xmath75 via the recursive identity @xmath212 where @xmath213 is arbitrary . by construction , the markov chain @xmath14 proposed by @xcite and defined via the recursive identity ( [ ftchain2 ] )
can be recovered from @xmath75 by setting @xmath65 . following the original idea of @xcite , by equation ( [ ftchain4 ] ) we have defined the markov chain @xmath75 from a distributional equation having as the unique solution the dirichlet process .
in particular , the markov chain @xmath75 is defined from the distributional equation ( [ ftchain3 ] ) which generalizes ( [ ftchain1 ] ) substituting the random probability measure @xmath214 with the random convex linear combination @xmath215 , for any fixed positive integer @xmath40 .
observe that@xmath215 is an example of the r.p.m .
@xmath95 defined in ( [ eqqrepr ] ) and investigated in the previous section , when @xmath216 is given by the plya sequence @xmath217 with parameter @xmath5 .
in particular , theorem [ asconv ] shows that @xmath95 a.s .- converges to the dirichlet process @xmath11 when @xmath40 goes to infinity ; however here we assume a different perspective , that is , @xmath40 is fixed . as for the case @xmath65 ,
the following result holds .
[ thminvariantpn ] the markov chain @xmath75 has a unique invariant measure @xmath218 which is the law of a dirichlet process @xmath11 with parameter @xmath5 .
another property which still holds in the more general case when @xmath219 is the harris ergodicity of the functional markov chain @xmath78 , under assumption ( [ eqcondlog ] ) below .
this condition is equivalent to the finiteness of the r.v .
@xmath220 ; see also @xcite .
[ thmharrisfunct ] let @xmath21 be any measurable function . if @xmath221 then the markov chain @xmath78 is harris ergodic with unique invariant measure @xmath222 , which is the law of the random dirichlet mean @xmath223 .
we conclude the analysis of the markov chain @xmath224 by providing some results on the ergodicity of the markov chain @xmath225 and by discussing on the rate of convergence .
let @xmath226 and let @xmath75 be the markov chain defined by ( [ ftchain4 ] ) . in particular , for the rest of the section , we consider the mean functional markov chain @xmath227 defined recursively by @xmath228 where @xmath229 is arbitrary and @xmath40 is a given positive integer . from theorem [ thmharrisfunct ] , under the condition @xmath230 , the markov chain
@xmath227 has the distribution @xmath231 of the random dirichlet mean @xmath232 as the unique invariant measure .
it is not restrictive to consider only the chain @xmath227 , since a more general linear functionals @xmath233 of a dirichlet process on an arbitrary polish space has the same distribution as the mean functional of a dirichlet process with parameter @xmath234 , where @xmath235 for any @xmath236 .
[ thmgeoergfunct ] the markov chain @xmath227 satisfies the following properties : a. @xmath227 is a stochastically monotone markov chain ; b. if further @xmath237=\int_{\mathbb{r}}|x|\alpha_0(\mathrm{d}x)<+\infty,\ ] ] then @xmath227 is a geometrically ergodic markov chain ; c. if the support of @xmath5 is bounded then @xmath227 is an uniformly ergodic markov chain .
recall that the stochastic monotonicity property of @xmath227 allows to consider exact sampling ( see @xcite ) for @xmath232 via @xmath238 .
[ remjarner ] condition can be relaxed . if the following condition holds @xmath239=\int_{\mathbb{r}}|y|^{s}\alpha_{0}(\mathrm{d}x)<+\infty\qquad\mbox{for some } 0<s<1,\ ] ] then the markov chain @xmath240 is geometrically ergodic .
see appendix for the proof . if , for instance , @xmath28 is a cauchy standard distribution and @xmath6 , condition is fulfilled so that @xmath240 will turn out to be geometrically ergodic for any fixed integer @xmath40 . from theorem [ asconv ]
, @xmath241 converges in distribution to the random dirichlet mean @xmath232 as @xmath103 ; so it is clear that , for a fixed integer @xmath40 , the law of @xmath242 approximates the law of @xmath232 and that the approximation will be better for @xmath40 large . if we reconsider ( [ ftchain11 ] ) , written as @xmath243 since the innovation term @xmath242 is an approximation in distribution of the limit ( as @xmath244 ) r.v .
@xmath232 , it is intuitive that the rate of convergence will increase as @xmath40 gets larger .
this is confirmed by the description of small sets @xmath245 in ( [ ftchain15 ] ) ( in the proof of theorem [ thmgeoergfunct ] ) .
in fact , under ( [ eqfinitmeanalpha0 ] ) or ( [ eqjar ] ) , the markov chain @xmath227 is geometrically or uniformly ergodic since it satisfies a foster
lyapunov condition @xmath246 for a suitable function @xmath247 , a small set @xmath245 and constants @xmath248 , @xmath249 . in particular , the small sets
@xmath245 generalize the corresponding small set @xmath250 obtained in theorem 1 in @xcite which can be recovered by setting @xmath65 , that is , @xmath251 $ ] where @xmath252}{\lambda - a/(1+a)}.\ ] ] here the size of the small set @xmath245 of @xmath227 can be controlled by an additional parameter @xmath40 , suggesting the upper bounds of the rate of convergence of the chain @xmath227 depends on @xmath40 too .
however , if we would establish an explicit upper bound on the rate of convergence , we would need results like theorem 2.2 in @xcite , or theorems 5.1 and 5.2 in @xcite .
all these results need a minorization condition to hold for the @xmath253th step transition probability @xmath254 for any @xmath119 and @xmath255 , for some positive integer @xmath253 and all @xmath256 in a small set ; in particular , if @xmath257 , where @xmath258 is the density of @xmath259 and @xmath260 is some density such that @xmath261 , then @xmath262 where @xmath263 is a probability measure on @xmath264 . in order to check the validity of our intuition that the rate of convergence will increase as @xmath40 gets larger , the function @xmath265 should be increasing with @xmath40 in order to prove that the uniform error ( when the support of the @xmath266 s is bounded ) in total variation between the law of @xmath267 given @xmath268 and its limit distribution decreases as @xmath40 increases .
if @xmath269 is the density of @xmath242 , which exists since , conditioning on @xmath270 s , @xmath242 is a random dirichlet mean , then @xmath271 therefore , the density function corresponding to @xmath272 is @xmath273 unfortunately , the explicit expression of @xmath269 , which for @xmath65 reduces to the density of @xmath28 if it exists , is not simple ; from proposition 5 in @xcite for instance , for @xmath274 , @xmath275 where , when @xmath276 for @xmath72 , @xmath277 here @xmath278 is the distribution of @xmath45 which , by definition , can be recovered by the product rule @xmath279 with @xmath280 and @xmath281}(y_{i}).\ ] ] however , some remarks on the asymptotic behavior of @xmath265 can be made under suitable conditions . since , @xmath282 if the support of @xmath266 s is bounded ( for instance equal to @xmath283 $ ] ) and the derivative of @xmath269 is bounded by some constant @xmath284 , then , by taylor expansion of @xmath269 , we have @xmath285\\ [ -8pt ] & \geq&\frac{n}{a+n}-kk^{(n)}(\lambda)\int_{0}^{1}(1-t)^{a}t^{n-2}\,\mathrm{d}t\nonumber\\ & = & \frac{n}{a+n}-\frac{akk^{(n)}(\lambda)}{n-1}.\nonumber\end{aligned}\ ] ] for a large enough @xmath286 , if we fix @xmath138 equal to some positive constant @xmath250 which is grater than @xmath287 for all @xmath288 , then @xmath289 is bounded above by @xmath290}{c - a/(a+n_{0})}.\ ] ] the second term in ( [ eqqualeps ] ) is negligible with respect to the first term , which increase as @xmath40 increases .
as we mentioned , when the support of the @xmath266 s is bounded , from theorem 16.2.4 in @xcite it follows that the error in total variation between the @xmath76th transition probability of the markov chain @xmath240 and the limit distribution @xmath231 is less than @xmath291 .
this error decreases for @xmath40 increasing greater than @xmath292 .
so far we have provided only some qualitative features on the rate of convergence ; however , the derivation of the explicit bound of the rate of convergence of @xmath267 to @xmath232 for each fixed @xmath40 , via @xmath293 and @xmath265 , is still an open problem .
some examples confirm our conjecture that the convergence of the markov chain @xmath240 improves as @xmath40 increases .
nonetheless , we must point out that simulating the innovation term @xmath242 for @xmath40 larger than 1 will be more computationally expensive , and also that this cost will be increasing as @xmath40 increases .
in fact , if @xmath40 is greater that one , @xmath294 more r.v.s must be drawn at each iteration of the markov chain ( @xmath295 more from the plya sequence and @xmath40 more from the finite - dimensional dirichlet distribution ) . moreover , we compared the total user times of the r function simulating @xmath296 .
we found that all these times were small , of course depending on @xmath297 , but not on the total mass parameter @xmath130 ( all the other values being fixed ) .
the total user times when @xmath298 were about @xmath299 greater than those for @xmath65 , while they were about 5 , 10 and 50 times greater when @xmath300 and 100 , respectively , for a number of total iterations equal to 500 . from the following examples
, we found that values of @xmath40 between @xmath301 and @xmath302 are a good choice between a fast rate of convergence and a moderate computational cost .
[ ex1 ] let @xmath28 be a uniform distribution on @xmath303 and let @xmath130 be the total mass .
in this case @xmath304=1/2 $ ] so that for any fixed integer @xmath40 , the chain will be geometrically ergodic ; moreover , it can be proved that @xmath303 is small so that the chain is uniformly ergodic .
when @xmath305 , @xcite showed that the convergence of @xmath306 is very good and there is no need to consider the chain with @xmath307 .
we consider the cases @xmath308 , @xmath309 and @xmath310 , and for each of them we run the chain for @xmath65 , @xmath301 , @xmath311 and @xmath302 .
we found that the trace plots do not depend on the initial values . in figure
[ figfigure1bis ] , we give the trace plots of @xmath267 when @xmath312 .
observe that convergence improves as @xmath40 increases for any fixed value of @xmath130 ; however the improvement is more glaring from the graph for large @xmath130 . when @xmath313 the convergence of the chain for @xmath65 seems to occur at about @xmath314 , while for @xmath315
the convergence is at about a value between @xmath309 and @xmath316 .
for these values of @xmath40 , the total user times to reach convergence was 0.038 seconds for the former , and 0.066 seconds for the latter .
moreover , the total user times to simulate 500 iterations of @xmath240 were 0.05 , 0.071 , 0.226 , 0.429 , 2.299 seconds when @xmath317 , respectively.=1 with @xmath28 the uniform distribution on @xmath303 , @xmath318 and @xmath65 ( solid blue line ) , @xmath298 ( dashed red line ) , @xmath319 ( dotted black line ) and @xmath315 ( dot - dashed violet line ) . ]
this behaviour is confirmed in the next example , where the support of the measure @xmath5 is assumed to be unbounded .
[ ex2 ] let @xmath28 be a gaussian distribution with parameter @xmath303 and let @xmath308 .
the behavior of @xmath320 has been considered in @xcite .
figure [ figfigura2 ] displays the trace plots of @xmath321 for three different initial values ( @xmath322 ) , with @xmath323 .
also in this case , it is clear that the convergence improves as @xmath40 increases .
as far as the total user times are concerned , we drew similar conclusions than in example [ ex1 ] .
with @xmath28 the gaussian distribution with parameter @xmath303 , @xmath308 , @xmath323 and @xmath324 ( dashed red line ) , @xmath325 ( solid blue line ) and @xmath326 ( dotted black line ) . ]
the next is an example in the mixture models context .
[ ex3 ] let us consider a gaussian kernel @xmath327 with unknown mean @xmath9 and known variance equal to 1 .
if we consider the random density @xmath328 , where @xmath11 is a dirichlet process with parameter @xmath5 , then , for any fixed @xmath256 , @xmath329 is a random dirichlet mean .
therefore , if we consider the measure - valued markov chain @xmath330 defined recursively as in ( [ ftchain4 ] ) , we define a sequence of random densities @xmath331 , where @xmath332 . in each panel of figure
[ figfigura3 ] , we drew @xmath333 for different values of @xmath76 when @xmath40 is fixed .
in particular , we fixed @xmath297 to be a gaussian distribution with parameter @xmath303 , and let @xmath313 ; in this case , since the `` variance '' of @xmath11 is small , the mean density @xmath334(x)=\int_{\mathbb{r } } k(x,\theta)\alpha_0(\mathrm{d}\theta)$ ] ( gaussian with parameter @xmath335 ) will be very close to the random function @xmath329 , so that it can be considered an approximation of the `` true '' density @xmath329 . from the plots ,
it is clear that the convergence improves as @xmath40 increases : when @xmath65 , only @xmath336 is close enough to the mean density @xmath334(x)$ ] , while , if @xmath315 , @xmath337 , as well as the successive iterations , is a good approximation of @xmath329 .
as in example [ ex3 ] ; @xmath313 , @xmath338 , @xmath339 and @xmath340 . ] in any case , observe that even if the `` true '' density @xmath329 is unknown , as when @xmath305 , the improvement ( as @xmath40 increases ) is clear as well ; see figure [ figfigura4 ] , where 5 draws of @xmath341 , @xmath342 , are plotted for different values of @xmath40 . as in example
[ ex3 ] ; @xmath305 , @xmath338 , @xmath339 and @xmath340 . ]
the paper @xcite constitutes , as far as we know , the first work highlighting the interplay between bayesian nonparametrics on the one side and the theory of measure - valued markov chains on the other . in the present paper , we have further studied such interplay by introducing and investigating a new measure - valued markov chain @xmath343 defined via exchangeable sequences of r.v.s . two applications related to bayesian nonparametrics
have been considered : the first gives evidence that @xmath343 is strictly related to the newton s algorithm of the mixture of dirichlet process model , while the second shows how @xmath81 can be applied in order to a define a generalization of the feigin
tweedie markov chain .
an interesting development consists in investigating whether there are any new applications related to the feigin
tweedie markov chain apart from the well - known application in the field of functional linear functionals of the dirichlet process ( see , e.g. , @xcite ) .
the proposed generalization of the feigin
tweedie markov chain represents a large class of measure - valued markov chains @xmath344 maintaining all the same properties of the feigin
tweedie markov chain ; in other terms , we have increased the flexibility of the feigin
tweedie markov chain via a further parameter @xmath345 .
we believe that a number of different interpretations for @xmath40 can be investigated in order to extend the applicability of the feigin tweedie markov chain . in this respect ,
an intuitive and simple extension is related to the problem of defining a ( bivariate ) vector of measure - valued markov chains @xmath346 , where , for each fixed @xmath76 , @xmath347 is a vector of dependent random probabilities , @xmath348 being fixed positive integers .
marginally , the two sequences @xmath349 and @xmath350 are defined via the recursive identity ( [ ftchain4 ] ) ; the dependence is achieved using the same plya sequence @xmath351 @xmath352 and assuming dependence in @xmath353 or between @xmath354 and @xmath355 .
for instance , if , for each @xmath76 , @xmath356 are independent r.v.s , @xmath357 with an exponential distribution with parameter @xmath130 , @xmath358 with an exponential distribution with parameter 1 , we could define @xmath359 , @xmath360 .
of course , the dependence is related to the difference between @xmath361 and @xmath362 .
work on this is ongoing .
a proof of the weak convergence of the sequence @xmath363 of r.p.m.s on @xmath0 to a dirichlet process @xmath11 is provided here , when the @xmath364 s are a plya sequence with parameter measure @xmath5 .
the result automatically follows from theorem [ asconv](ii ) , but this proof is interesting _ per se _ , since we use a combinatorial technique ; moreover an explicit expression for the moment of order @xmath132 ) of the @xmath133-dimensional plya distribution is obtained .
let @xmath95 defined in ( [ eqprec ] ) , where @xmath47 are a plya sequence with parameter @xmath5 .
then @xmath365 where @xmath11 is a dirichlet process on @xmath0 with parameter @xmath5 . by theorem 4.2 in @xcite ,
it is sufficient to prove that for any measurable partition @xmath366 of @xmath0 , @xmath367 characterizing the distribution of the limit . for
any given measurable partition @xmath366 of @xmath0 , by conditioning on @xmath368 , it can be checked that @xmath369 is distributed according to a dirichlet distribution with empirical parameter @xmath370 , and @xmath371\\ [ -8pt ] & & \quad={n\choose j_{1}\cdots j_{k}}\frac{(\alpha(b_{1}))_{j_{1}\uparrow1 } \cdots(\alpha(b_{k}))_{j_{k}\uparrow1}}{(a)_{n\uparrow1 } } , \nonumber\end{aligned}\ ] ] where @xmath372 with @xmath373 . for any @xmath133-uple of nonnegative integers @xmath374
, we are going to compute the limit , for @xmath103 , of the moment @xmath375\nonumber\\ [ -8pt]\\ [ -8pt ] & & \hspace*{-5pt}\quad= \sum_{(j_{1},\ldots , j_{k})\in\mathcal{d}^{(0)}_{k , n } } { n\choose j_{1}\cdots j_{k } } \frac{(\alpha(b_{1}))_{j_{1}\uparrow1 } \cdots(\alpha(b_{k}))_{j_{k}\uparrow1}}{(a)_{n\uparrow1 } } \frac{(j_{1})_{r_{1}\uparrow1}\cdots(j_{k})_{r_{k}\uparrow1 } } { ( n)_{(r_{1}+\cdots+r_{k})\uparrow1}},\nonumber\end{aligned}\ ] ] where in general @xmath376 denotes the pochhammer symbol for the @xmath40th factorial power of @xmath256 with increment @xmath5 , that is @xmath377 .
we will show that , as @xmath378 , @xmath379\\ & & \quad\rightarrow{\mathbb{e}}\biggl[(p(b_1)\big)^{r_1}\cdots(p(b_{k-1}))^{r_{k-1 } } \biggl ( 1-\sum_{i=1}^{k-1}p(b_{i})\biggr)^{r_k}\biggr],\end{aligned}\ ] ] where @xmath11 is a dirichlet process on @xmath380 with parameter measure @xmath5 , that is , the r.v .
@xmath381 has dirichlet distribution with parameter @xmath382
. this will be sufficient to characterize the distribution of the limit @xmath122 , because of the boundedness of the support of the limit distribution .
first of all , we prove the convergence for @xmath383 , which corresponds to the one - dimensional case .
in particular , we have @xmath384\\ & & \quad= \frac{1}{(n)_{(r_{1}+r_{2})\uparrow1}}\sum_{(j_{1},j_{2})\in\mathcal{d}^{(0)}_{2,n } } { n\choose j_{1},j_{2 } } \frac{(\alpha(b_{1}))_{j_{1}\uparrow1}(\alpha(b_{2}))_{j_{2}\uparrow1 } } { ( a)_{n\uparrow1 } } ( j_{1})_{r_{1}\uparrow1}(j_{2})_{r_{2}\uparrow1}\\ & & \quad=\frac{1}{(n)_{(r_{1}+r_{2})\uparrow1}}\sum_{t_{1}=1}^{r_{1}}|s(r_{1},t_{1})|\sum_{s_{1}=1}^{t_{1}}s(t_{1},s_{1 } ) \sum_{t_{2}=1}^{r_{2}}|s(r_{2},t_{2})|\sum_{s_{2}=1}^{t_{2}}s(t_{2},s_{2})\\ & & \qquad{}\times\sum_{(j_{1},j_{2})\in\mathcal{d}^{(0)}_{2,n } } { n\choose j_{1},j_{2 } } \frac{(\alpha(b_{1}))_{j_{1 } \uparrow1}(\alpha(b_{2}))_{j_{2}\uparrow1}}{(a)_{n\uparrow1 } } ( j_{1})_{s_{1}\downarrow1}(j_{2})_{s_{2}\downarrow1},\end{aligned}\ ] ] where @xmath385 and @xmath386 and @xmath387 are the stirling number of the first and second kind , respectively .
let us consider the following numbers , where @xmath388 , @xmath389 are nonnegative integers and @xmath390 @xmath391 and prove they satisfy a recursive relation .
in particular , @xmath392 so that the following recursive equation holds @xmath393 therefore , starting from @xmath394 , @xmath395 , we have @xmath396 and by ( [ rec1 ] ) we obtain @xmath397 .
thus , @xmath398\\ & & \quad = \sum_{t_{1}=0}^{r_{1}}|s(r_{1},t_{1})|\sum_{s_{1}=0}^{t_{1}}s(t_{1},s_{1})\sum_{t_{2}=0}^{r_{2}}|s(r_{2},t_{2})|\sum_{s_{2}=0}^{t_{2}}s(t_{2},s_{2})\lim_{n\rightarrow+\infty}\frac{c_{n}^{(s_{1},s_{2})}}{(n)_{(r_{1}+r_{2})\uparrow1 } } \\ & & \quad = |s(r_{1},r_{1})|s(r_{1},r_{1})|s(r_{2},r_{2})|s(r_{2},r_{2})\frac{(\alpha(b_{1}))_{r_{1}\uparrow1 } ( \alpha(b_{2}))_{r_{2}\uparrow1 } } { ( a)_{(r_{1}+r_{2})\uparrow1}}\\ & & \quad=\frac{(\alpha(b_{1}))_{r_{1}\uparrow1}(\alpha(b_{2}))_{r_{2}\uparrow1 } } { ( a)_{(r_{1}+r_{2})\uparrow1}}.\end{aligned}\ ] ] the last expression is exactly @xmath399 $ ] , where @xmath400 has beta distribution with parameter @xmath401
. this proof can be easily generalized to the case @xmath402 .
analogously to the one - dimensional case , we can write @xmath403\\ & & \quad=\frac{1}{(n)_{(r_{1}+\cdots+r_{k})\uparrow1}}\sum_{(j_{1},\ldots , j_{k})\in \mathcal{d}^{(0)}_{k , n } } { n\choose j_{1}\cdots j_{k } } \frac{(\alpha(b_{1}))_{j_{1}\uparrow1}\cdots(\alpha(b_{k}))_{j_{k}\uparrow1}}{(a)_{n\uparrow1 } } ( j_{1})_{r_{1}\uparrow1}\cdots(j_{k})_{r_{k}\uparrow1}\\ & & \quad=\frac{1}{(n)_{(r_{1}+\cdots+r_{k})\uparrow1}}\sum_{t_{1}=0}^{r_{1}}|s(r_{1},t_{1})|\sum_{s_{1}=0}^{t_{1}}s(t_{1},s_{1 } ) \cdots\sum_{t_{k}=0}^{r_{k}}|s(r_{k},t_{k})|\sum_{s_{k}=0}^{t_{k}}s(t_{k},s_{k})\\ & & \qquad { } \times\sum_{(j_{1},\ldots , j_{k})\in\mathcal{d}^{(0)}_{k , n } } { n\choose j_{1}\cdots j_{k } } \frac{(\alpha(b_{1}))_{j_{1}\uparrow1}\cdots(\alpha(b_{k}))_{j_{k}\uparrow1}}{(a)_{n\uparrow1 } } ( j_{1})_{s_{1}\downarrow1}\cdots(j_{k})_{s_{k}\downarrow1}\end{aligned}\ ] ] and , as before , define @xmath404 and prove they satisfy a recursive relation . observe that @xmath405 is the moment of order @xmath374 of the @xmath133-dimensional plya distribution by definition . therefore , for @xmath402 @xmath406 & & \hphantom{\sum_{j_k=0}^{n}}{}\times\sum_{(j_{1},\ldots , j_{k-1})\in\mathcal{d}^{(0)}_{k-1,n - j_{k } } } { n - j_{k}\choose j_{1}\cdots j_{k-1 } } \frac{(\alpha(b_{1}))_{j_{1}\uparrow1}\cdots(\alpha(b_{k-1}))_{j_{k-1}\uparrow1}}{(a-\alpha(b_{k}))_{(n - j_{k})\uparrow1 } } \\[-2pt ] & & \hspace*{10pt}\hphantom{\sum_{j_k=0}^{n}\sum_{(j_{1},\ldots , j_{k-1})\in\mathcal{d}^{(0)}_{k-1,n - j_{k } } } } { } \times(j_{1})_{s_{1}\downarrow1}\cdots(j_{k-1})_{s_{k-1}\downarrow1}\\[-2pt ] & = & \sum_{j_{k}=0}^{n}{n\choose j_{k}}\frac{(\alpha(b_{k}))_{j_{k}\uparrow1}(a-\alpha(b_{k}))_{(n - j_{k})\uparrow1 } ( j_{k})_{s_{k}\downarrow1 } } { ( a)_{n\uparrow1 } } \\[-2pt ] & & \hphantom{\sum_{j_{k}=0}^{n}}{}\times\frac{(\alpha(b_{1}))_{s_{1}\uparrow1}\cdots ( \alpha(b_{k-1}))_{s_{k-1}\uparrow1 } ( n - j_{k})_{(s_{1}+\cdots+ s_{k-1})\downarrow1}}{(a-\alpha(b_{k}))_{(s_{1}+\cdots+s_{k-1})\uparrow1}},\end{aligned}\ ] ] where the last equality follows by induction hypothesis ( we already proved the base case @xmath383 in ( [ rec1 ] ) ) .
then , following the same steps of the one - dimensional case , we can recover a recursive equation for @xmath407 , @xmath408 starting from @xmath409 , @xmath410 and @xmath411 & = & \frac{\gamma(n+1)}{\gamma(a+n)}\frac{(\alpha(b_{1}))_{s_{1}\uparrow1}\cdots ( \alpha(b_{k-1}))_{s_{p-1}\uparrow1}\alpha(b_{k } ) } { ( a-\alpha(b_{k}))_{(s_{1}+\cdots+s_{k-1})\uparrow1}}\sum_{i=0}^{n-1}\frac{\gamma(a+i)}{\gamma(1+i)}c_{i}^{(1,\ldots,1,0 ) } \\[-2pt ] & = & \frac{\gamma(n+1)}{\gamma(a+n)}\frac{(\alpha(b_{1}))_{s_{1}\uparrow1}\cdots ( \alpha(b_{k-1}))_{s_{k-1}\uparrow1}\alpha(b_{k } ) } { ( a-\alpha(b_{k}))_{(s_{1}+\cdots+s_{k-1})\uparrow1}}\\[-2pt ] & & { } \times(-1)^{k+1}\sum_{i=0}^{n-1}\frac{\gamma(a+i)}{\gamma(1+i ) } \frac{(a-\alpha(b_{k}))_{i\uparrow1 } ( -i)_{(k-1)\uparrow1}}{(a)_{i\uparrow1}}\\[-2pt ] & & \hphantom { { } \times(-1)^{k+1}\sum_{i=0}^{n-1}}{}\times\frac{\gamma(-a+\alpha(b_{k})-i+1)\gamma(-k - a+2 ) } { \gamma(-a - i+1)\gamma(-a+\alpha(b_{k})+2-k ) } \\[-2pt ] & = & \frac{\alpha(b_{1})\cdots\alpha(b_{k})(n)_{k\downarrow1}}{(a)_{(k\uparrow1)}},\end{aligned}\ ] ] by repeated application of ( [ rec2 ] ) , we obtain @xmath412 thus , @xmath413\\ & & \quad= \sum_{t_{1}=0}^{r_{1}}|s(r_{1},t_{1})|\sum_{s_{1}=0}^{t_{1}}s(t_{1},s_{1})\cdots\sum_{t_{k}=0}^{r_{k}}|s(r_{k},t_{k})|\sum_{s_{k}=0}^{t_{k}}s(t_{k},s_{k } ) \lim_{n\rightarrow+\infty}\frac{c_{n}^{(s_{1},\ldots , s_{k})}}{(n)_{(r_{1}+\cdots+r_{k})\uparrow1}}\\ & & \quad=|s(r_{1},r_{1})|s(r_{1},r_{1})\cdots|s(r_{k},r_{k})|s(r_{k},r_{k } ) \frac{(\alpha(b_{1}))_{r_{1}\uparrow1}\cdots(\alpha(b_{k}))_{r_{k}\uparrow1 } } { ( a)_{(r_{1}+\cdots+r_{k})\uparrow1}}\\ & & \quad=\frac{(\alpha(b_{1}))_{r_{1}\uparrow1}\cdots(\alpha(b_{k}))_{r_{k}\uparrow1 } } { ( a)_{(r_{1}+\cdots+r_{k})\uparrow1}}\\ & & \quad={\mathbb{e}}\biggl[(p(b_1))^{r_1}\cdots(p(b_{k-1}))^{r_{k-1 } } \biggl ( 1-\sum_{i=1}^{k-1}p(b_{i})\biggr)^{r_k}\biggr],\end{aligned}\ ] ] where @xmath11 is a dirichlet process with parameter @xmath5 . here , we provide an alternative proof for the solution of the distributional equation ( [ ftchain3 ] ) introduced by @xcite . for any fixed integer @xmath55 ,
the distributional equation @xmath414 has the dirichlet process with parameter @xmath5 as its unique solution , assuming the independence between @xmath61 , @xmath415 and @xmath45 in the right - hand side . from skorohod s theorem
, it follows that there exist @xmath40 independent r.v.s @xmath416 such that @xmath417 has beta distribution with parameter @xmath418 for @xmath72 and @xmath419 and @xmath420 for @xmath421 ; in particular , by a simple transformation of r.v.s , it follows that @xmath56 is distributed according to a dirichlet distribution function with parameter @xmath57 .
further , since @xmath422 a.s .
, then @xmath423 a.s . and
it can be verified by induction that @xmath424 let @xmath366 be a measurable partition of @xmath0 .
we first prove that conditionally on @xmath425 , the finite dimensional distribution of the r.p.m .
@xmath215 is the dirichlet distribution with the empirical parameter @xmath426 . actually , since @xmath427 then , conditionally on the r.v.s @xmath425 , the r.v .
@xmath428 , @xmath429@xmath430 is distributed according to a dirichlet distribution with parameter @xmath431 @xmath432 , where @xmath433 for @xmath434 .
conditionally on @xmath425 , the finite dimensional distributions of the right - hand side of ( [ ftchain3 ] ) are dirichlet with updated parameter @xmath435 .
this argument verifies that the dirichlet process with parameter @xmath5 satisfies the distributional equation ( [ ftchain3 ] ) .
this solution is unique by lemma 3.3 in @xcite ( see also @xcite , section 1 ) .
proof of theorem [ thminvariantpn ] the proof is based on the `` standard '' result that properties ( e.g. , weak convergence ) of sequences of r.p.m.s can be proved via analogous properties of the sequences of their linear functionals .
first of all , we prove that if @xmath21 is a bounded and continuous function , then @xmath78 with @xmath436 is a markov chain on @xmath264 with unique invariant measure @xmath222 . from ( [ ftchain4 ] )
, it follows that @xmath437 is a markov chain on @xmath264 restricted to the compact set @xmath438 $ ] and it has at least one finite invariant measure if it is a weak feller markov chain .
in fact , for a fixed @xmath274 @xmath439 since the distribution of @xmath440 has at most a countable numbers of atoms and @xmath441 is absolutely continuous . from proposition 4.3 in @xcite
, if we show that @xmath442 is @xmath443-irreducible for a finite measure @xmath443 , then the markov chain is positive recurrent and the invariant measure @xmath444 is unique .
let us consider the following event @xmath445 .
then for a finite measure @xmath443 we have to prove that if @xmath446 , then @xmath447 for any @xmath448 .
we observe that @xmath449 therefore , since @xmath450 , using the same argument in lemma 2 in @xcite , we conclude that @xmath451 for a suitable measure @xmath443 such that @xmath446 . finally , we prove the aperiodicity of @xmath225 by contradiction . if the chain is periodic with period @xmath452 , the exist @xmath453 disjoint sets @xmath454 such that @xmath455 for all @xmath456 and for @xmath457 .
this implies @xmath458 for almost every @xmath459 with respect to the lebesgue measure restricted to @xmath303 .
thus , @xmath460 for @xmath461 . for generic @xmath5 and @xmath126 ,
this is in contradiction with the assumption @xmath452 . by theorem 13.3.4(ii ) in @xcite , @xmath462 converges in distribution for @xmath444-almost all starting points @xmath463 .
in particular , @xmath78 converges weakly for @xmath222-almost all starting points @xmath448 . from the arguments above
, it follows that , for all @xmath126 bounded and continuous , there exists a r.v .
@xmath233 such that @xmath464 as @xmath244 for @xmath222-almost all starting points @xmath448 .
therefore , for lemma 5.1 in @xcite there exists a r.p.m .
@xmath11 such that @xmath465 as @xmath244 and @xmath466 for all @xmath467 .
this implies that the law of @xmath11 is an invariant measure for the markov chain @xmath75 .
then , as @xmath244 , @xmath468 and the limit is unique for any @xmath467 . since for any random measure @xmath469 and @xmath470 we know that @xmath471 if and only if @xmath472 for any @xmath467 ( see theorem 3.1 . in @xcite ) , the invariant measure for the markov chain @xmath75 is unique . by the definition of @xmath75 , it is straightforward to show that the limit @xmath11 must satisfy ( [ ftchain3 ] ) so that @xmath11 is the dirichlet process with parameter @xmath5 .
proof of theorem [ thmharrisfunct ] the proof is a straightforward adaptation of the proof of theorem 2 in @xcite , using @xmath473 & \leq & \sum_{i=1}^{n}\log\bigl(1+|g(y_{1,i})|\bigr)+ \log\bigl(1+(1-\theta_{1})\big|g^{(n)}_{0}|\bigr)\end{aligned}\ ] ] instead of their inequality ( 8) . proof of theorem [ thmgeoergfunct ] as regards ( i ) , given the definition of stochastically monotone markov chain , we have that for @xmath474 , @xmath475 , @xmath476 & \geq & \mathbb{p}\biggl(\theta_{1}\sum_{i=1}^{n}q_{1,i}^{(n)}y_{1,i}+(1-\theta_{1})z_{2}<a\biggr)\\[-2pt ] & = & p^{(n)}_{1}(z_{2},(-\infty , s)).\end{aligned}\ ] ] as far as ( ii ) is concerned , we first prove that , under condition ( [ eqfinitmeanalpha0 ] ) , the markov chain @xmath227 satisfies the foster
lyapunov condition for the function @xmath477 .
this property implies the geometric ergodicity of the @xmath227 ( see @xcite , chapter 15 ) .
we have @xmath478\\[-2pt ] & \leq & 1+{\mathbb{e}}[\theta_{1}]\sum_{i=1}^{n}e\bigl[\bigl|q_{1,i}^{(n)}y_{1,i}\bigr|\bigr]+|x|{\mathbb{e}}[1-\theta_{1}]\\[-2pt ] & \leq & 1+\frac{n}{n+a}\sum_{i=1}^{n}{\mathbb{e}}\bigl[\bigl|q_{1,i}^{(n)}y_{1,i}\bigr|\bigr]+\frac{a}{n+a}|x|=1+\frac{n}{n+a}{\mathbb{e}}[|y_{1,1}|]+\frac{a}{n+a}|x|.\end{aligned}\ ] ] therefore , we are looking for the small set @xmath245 such that the foster
lyapunov condition holds , that is , a small set @xmath245 such that @xmath479+\frac{a}{n+a}|x|\leq\lambda(1+|x|)+b\mathbbl{1}_{c^{(n)}}(x)\ ] ] for some constant @xmath248 and @xmath249 .
if @xmath480 $ ] , where @xmath481}{\lambda - a/(n+a)},\vadjust{\goodbreak}\ ] ] then , condition ( [ ftchain15 ] ) holds for all @xmath482.\ ] ] as in the proof of theorem [ thminvariantpn ] , we can prove that the markov chain @xmath227 is weak feller ; then , since @xmath245 is a compact set , it is a small set ( see @xcite ) . as regards ( iii )
, the proof follows by standard arguments .
see , for instance , the proof of theorem 1 in @xcite .
proof of remark [ remjarner ] as we have already mentioned , the geometric ergodicity follows if a foster
lyapunov condition holds .
let @xmath483 ; then , if @xmath484<+\infty$ ] , it is straightforward to prove that the foster
lyapunov condition @xmath485 holds for some constant @xmath248 , and @xmath138 such that @xmath486=\frac{\gamma(a+s)\gamma(a+n)}{\gamma(a)\gamma(a+s+n)}<\lambda<1\ ] ] and for some compact set @xmath487 .
of course ( [ eqjar ] ) implies @xmath484<+\infty;$ ] in fact , conditioning on the random number @xmath488 of distinct values @xmath489 in @xmath490 , @xmath491 , we have @xmath492 since @xmath493 are independent and identically distributed according to @xmath28 , then @xmath494\leq\int_{0}^{+\infty}y^{s}n(a_{0}(y))^{n-1}\alpha_{0}(\mathrm{d}y ) \leq
n\int_{0}^{+\infty}y^{s}\alpha_{0}(\mathrm{d}y)\leq n{\mathbb{e}}[|y_{1,i}|^{s}]<+\infty,\end{aligned}\ ] ] where @xmath495 is the distribution corresponding to the probability measure @xmath28 , and this is equivalent to @xmath484<+\infty$ ] .
the authors are very grateful to patrizia berti and pietro rigo who suggested the proof of theorem [ rigo ] , and to eugenio regazzini for helpful discussions .
the authors are also grateful to an associate editor and a referee for comments that helped to improve the presentation .
the second author was partially supported by miur grant 2006/134525 . | measure - valued markov chains have raised interest in bayesian nonparametrics since the seminal paper by ( _ math .
proc .
cambridge philos .
soc . _ * 105 * ( 1989 ) 579585 ) where a markov chain having the law of the dirichlet process as unique invariant measure has been introduced . in the present paper , we propose and investigate a new class of measure - valued markov chains defined via exchangeable sequences of random variables .
asymptotic properties for this new class are derived and applications related to bayesian nonparametric mixture modeling , and to a generalization of the markov chain proposed by ( _ math .
proc .
cambridge philos .
soc .
_ * 105 * ( 1989 ) 579585 ) , are discussed .
these results and their applications highlight once again the interplay between bayesian nonparametrics and the theory of measure - valued markov chains . , |
the disordered equiatomic binary alloys of the xy ( x= fe , co y= pd ,
pt ) type crystallize in the fcc structure and the magnetization is along the [ 111 ] axis.@xcite at low temperatures these alloys tend to order in the @xmath0 layered - ordered structure and in this case the spontaneous magnetization tends to align perpendicular to the layer stacking explaining the behavior of copt films during the magnetic annealing.@xcite the strong perpendicular magnetic anisotropy ( pma ) is due to the highly anisotropic @xmath0 structure and it makes them very attractive for magnetic recording devices.@xcite the latter structure can be also obtained by molecular beam epitaxy ( mbe ) of alternating layers of pure x and y atoms due to the substrate induced constraints .
the first observation of the @xmath1 long - range order for a [ 001 ] copt film grown by mbe was made by harp _
et al._@xcite in 1993 and for a [ 001 ] fept film by cebollada _
et al._@xcite lately other techniques have been also employed to develop films presenting pma .
copt films were grown by sputtering by visokay _
et al._@xcite and by evaporation by lin and gorman,@xcite and fept films were grown by various sputtering techniques.@xcite the magneto - crystalline anisotropy energy ( mca ) can be probed by many techniques such as torque or ferromagnetic resonance measurements .
both these methods describe the mca in terms of phenomenological anisotropy constants .
it has been demonstrated by weller _
et al._@xcite that x - ray magnetic circular dichroism ( xmcd ) is also a suitable technique for probing the mca , _ via _ the determination of the anisotropy of the orbital magnetic moment on a specific shell and site .
the x - ray absorption spectroscopy ( xas ) using polarized radiation probes element specific magnetic properties of alloys by applying the xmcd sum rules to the experimental spectra.@xcite however for itinerant systems , in particular to low symmetry systems , the use of the sum rules is debated because they are derived from an atomic theory.@xcite lately angle - dependent xmcd experiments have been used to provide a deeper understanding for the relation between mca and the orbital magnetic moments.@xcite the x - ray absorption for copt multi - layers has been already studied experimentally by nakajima _
et al._,@xcite koide _ et al._,@xcite regg _
et al._,@xcite and schltz _
et al._@xcite nakajima _
et al._@xcite revealed a strong enhancement of the cobalt orbital moment when pma was present , koide _
et al._@xcite showed that with decreasing cobalt thickness the easy axis rotates from in - plane to out - of - plane , and regg _
et al._@xcite that platinum polarization increases also with decreasing cobalt thickness .
hlil _ et al._@xcite showed by x - ray absorption spectroscopy that modifications of platinum edges in different compounds are correlated to the change in the number of holes .
several _ ab - initio _ calculations have already been performed to investigate the xmcd.@xcite the @xmath2- and @xmath3-edges involving electronic excitations of 2@xmath4-core electrons towards @xmath5-valence states have primarily attracted much attention due to dependence of the dichroic spectra on the exchange - splitting and the spin - orbit coupling of both initial core and final valence states . for 5@xmath5 elements dissolved in 3@xmath5 transition metals , the spin - orbit coupling of the initial 2@xmath4-core states is large and the resulting magnetic moment is small , while the opposite is true for the 3@xmath5 elements .
this can lead to a pronounced dichroic spectra as seen by schtz in the case of 5@xmath5 elements dissolved in iron.@xcite in this work we study the correlation between the quantization axis dependent xmcd and the magnetic properties of both ordered alloys fept and copt .
our method is based on an all - electron relativistic and spin - polarized full - potential muffin - tin orbital method ( lmto)@xcite in conjunction with both the von barth and hedin parameterization to the local density approximation ( lsda)@xcite and the generalized gradient approximation ( gga)@xcite to the exchange correlation potential . the implementation of the calculation of the xmcd spectra has been presented in a previous work.@xcite in section 2 we present the details of the calculations , while in sections 3 and 4 we discuss our mca and the magnetic spin and orbital moments , respectively . in section 5
we present our calculated xmcd as a function of the spin quantization axis , and in section 6 we discuss the interpretation of the mca using the band structure and the total density of states anisotropy .
to compute the electronic properties of copt and fept we used the experimental lattice constants ( @xmath6=3.806 and @xmath7/@xmath6=0.968 for copt@xcite ; @xmath6= 3.861 and @xmath7/@xmath6=0.981 for fept@xcite ) and a unit cell containing one atom of cobalt(iron ) and one of platinum .
the @xmath0 structure can be seen as a system of alternating cobalt(iron ) and platinum layers along the [ 001 ] direction .
the mca and the xmcd are computed with respect to the angle @xmath8 between the [ 001 ] axis and the spin quantization axis on the ( 010 ) plane .
so @xmath8=0@xmath9 corresponds to the [ 001 ] axis and @xmath8=90@xmath9 to the [ 100 ] axis .
mca can be computed directly using _ ab - initio _ methods ; it is defined as the difference between the total energy for two different spin quantization axis . the spin - orbit coupling contribution to mca
is implicitly included in our _ ab - initio _ calculations , and we do not take into account the many - body interactions of the spin - magnetic moments@xcite since their contribution to the mca is negligible.@xcite the number of * k * points for performing the brillouin zone ( bz ) integration depends strongly on the interplay between the contributions to the mca from the fermi surface and the remaining band structure contribution to the total energy.@xcite when the former contribution to the mca is important , a large number of * k*-points is needed to describe accurately the fermi surface .
for the two studied systems we found that 6750 * k*-points in the bz are enough to converge the mca within 0.1 mev . to perform the integrals over the bz we use a gaussian broadening method which convolutes each discrete eigenvalue with a gaussian function of width 0.1 ev .
this method is known to lead to a fast and stable convergence of the spin and charge densities compared to the standard tetrahedron method . to develop the potential inside the mt spheres we calculated a basis set of lattice harmonics including functions up to @xmath10 , while for the fft we used a real space grid of 16@xmath1116@xmath1120 .
we used a double set of basis functions , one set to describe the valence states and one for the unoccupied states . for the valence electrons we used a basis
set containing 3@xmath12 , 3@xmath13 and 2@xmath14 wave functions , and for the unoccupied states 2@xmath12 , 2@xmath13 and 2@xmath14 wave functions .
copt and fept films are known to present a strong uniaxial mca , because of the high anisotropic @xmath0 inter - metallic phase .
experimentally the magnetization axis is found to be along the [ 001 ] direction.@xcite in a first step we performed calculations with 250 * k*-points in the bz .
this number of * k*-points is large enough to produce accurate total energy when gaussian smearing is used for the integration in the brillouin zone but not enough accurate to compute the mca . in figure [ fig12 ]
we present calculations with respect to the angle @xmath8 between the [ 001 ] axis and the spin quantization axis on the ( 010 ) plane for the copt compound within the lsda .
we observe that total energy value increases with the angle @xmath8 .
so the ground state corresponds to @xmath8=0@xmath9 , _
i.e. _ the [ 001 ] axis .
the same behavior occurs for the fept compound . because 250 * k*-points are not enough to produce an accurate value for the mca , we present in figure [ fig2 ] the convergence of the mca with the number of * k*-points for both compounds within lsda .
the mca is the difference in the total energy between the in - plane axis [ 100 ] and [ 110 ] and the easy axis [ 001 ] .
we have found that the total energy difference between the [ 100 ] and the [ 110 ] directions is negligible compared to the difference between the in - plane axis and the [ 001 ] axis .
our values are converged up to 6750 * k*-points and are given per unit cell ( one atom of x and one of platinum ) .
the mca converges to 2.2 ev for copt and 3.9 ev for fept .
these behavior confirms the assumption that the system is isotropic inside the plane .
the gga mca calculations converged to 1.9 mev and 4.1 mev for copt and fept , respectively .
the gga results seem to be in good agreement with the lsda results .
daalderop _ et al._@xcite performed calculations for mca in copt and fept by means of an lmto method in the atomic sphere approximation ( asa ) within lsda@xcite using the force theorem.@xcite the easy magnetization axis found by this latter calculation is the [ 001 ] for both systems in agreement with our lsda and gga results .
their mca value is 2 mev for copt and is 3.5 mev for fept .
et al._@xcite used the same structure within a real - space green s function technique framework to find also that the magnetization is along the [ 001 ] axis . including the spin - orbit interaction for all the atoms , they found a copt mca value of 2.3 mev and a fept value of 3.4 mev .
our computed value agrees also with the value of 1.5 mev for copt and 2.8 mev for fept , obtained by sakuma@xcite using lmto method in the atomic sphere approximation ( asa ) in conjunction with the force theorem.@xcite the drawback of the lmto - asa method is that it accounts only for the spherical part of the potential and ignores the interstitial region .
furthermore , the force theorem does not account directly for the exchange - correlation contribution to the mca . finally , oppeneer used the augmented spherical waves method ( asw ) in the atomic sphere approximation and found a mca value of 2.8 mev for fept and 1.0 mev for copt,@xcite which are the smallest among all the _ ab - initio _ calculations .
grange _ et al._,@xcite using a torque measurement for a mbe deposited copt film on a mgo(001 ) substrate , obtained a mae of 1.0 mev .
an early measurement of a monocristal of copt , by eurin and pauleve , produced a value of 1.3 mev.@xcite the large value of eurin and pauleve is due to the fact that their sample was completely ordered . for fept the first experiment of ivanov _ et al._@xcite produced an anisotropy value of 1.2 mev and showed that for a thin film the shape anisotropy would be one order of magnitude smaller compared to the mca .
et al._@xcite and thiele _
et al._@xcite found for mbe deposited fept films on a mgo(001 ) substrate an anisotropy value of 1.8 mev .
these films were highly ordered ( more than 95% of the atoms were in the correct site ) .
all experiments have been carried out at room temperature , which explain at some extend the difference between the calculated and experimental mca values ( the mca decreases with temperature@xcite ) .
it is worth mentioning that the experimental mca for fept is much larger than that of copt in agreement with our calculations . for thick films volume shape anisotropy ( vsa )
contributes also to the mae and it favors always an in - plane magnetization axis .
we can estimate the vsa as @xmath15 in c.g.s
. units , where @xmath16 is the mean magnetization density,@xcite and obtain a value of -0.1 mev for fept and -0.06 mev for copt .
these values are one order of magnitude smaller than the mca values in agreement with the speculation of ivanov _
in figure [ fig3 ] we present the cobalt - projected partial density of states for three spin quantization axis corresponding to angles @xmath8= 0@xmath9 , 45@xmath9 and 90@xmath9 calculated with 6750 * k*-points .
the 3-@xmath5 states dominate the electronic structure of cobalt .
the spin - up band is practically totally occupied while the spin down band is almost half - occupied .
the general form of the cobalt dos , as well as the iron dos , does not seem to change appreciably with the angle @xmath8 .
the platinum projected density of states show similar behavior for both fept and copt compounds .
bulk platinum is paramagnetic and the small changes in the dos come from the polarization of the 5@xmath5 electrons via hybridization with the 3@xmath5 electrons of cobalt ( iron ) .
it is worth mentioning that the dos is calculated inside each muffin - tin and the interstitial region is not taken into account . to minimize the contribution of the interstitial region we use almost touching muffin - tin spheres .
in addition , we have find that remaining region has a negligible spin polarization .
the spin magnetic moments are isotropic with respect to the spin quantization axis as expected .
they were calculated by attributing all the charge inside each muffin - tin sphere to the atom located in that sphere .
as outlined above we have found that the interstitial contribution to the spin magnetic moment is one order of magnitude smaller than that of the platinum site . in copt
we found a cobalt spin magnetic moment of 1.79 @xmath17 and a platinum moment of 0.36 @xmath17 within the lsda .
the gga cobalt spin magnetic moment of 1.83 @xmath17 is slightly larger than the value within lsda .
this is due to a more atomic like description of the atoms in a solid within the gga compared to the lsda .
although the gga underestimates the hybridization between cobalt and platinum @xmath5 valence electrons compared to the lsda , the larger cobalt spin moment leads to a slightly larger gga platinum spin moment of 0.37 @xmath17 .
our calculated values are in good agreement with the experimental values of grange _ et al . _
( 1.75@xmath17 for cobalt and 0.35@xmath17 for platinum).@xcite previous experiments by van laar on a powder sample gave a value of 1.7 @xmath17 for the cobalt atom and 0.25 for the platinum atom.@xcite the spin magnetic moments have been previously calculated by solovyev _ et al._@xcite ( 1.72 for cobalt and 0.37 for platinum ) , by daalderop _ et al_@xcite ( 1.86 for cobalt ) , by sakuma@xcite ( 1.91 for cobalt and 0.38 for platinum ) , and finally by kootte _ et al._@xcite by means of a localized spherical wave method ( 1.69 for cobalt and 0.37 for platinum ) .
all previous calculations are in good agreement with our full - potential results .
as expected the iron spin moments are much larger than the cobalt ones .
the lsda produced a value of 2.87 @xmath17 while the gga produced a slightly larger value , 2.96 @xmath17 .
the hybridization between the iron 3@xmath5 states and the platinum 5@xmath5 states is less intense than in the case of copt resulting in a smaller platinum moment for fept .
the lsda platinum spin moment is 0.33 @xmath17 ( compared to 0.36 @xmath17 in copt ) and the gga platinum spin moment is 0.34 @xmath17 ( compared to 0.37 @xmath17 in copt ) .
the spin magnetic moments have been previously calculated by solovyev et al@xcite ( 2.77 for iron and 0.35 for platinum ) , by daalderop et al@xcite ( 2.91 for iron ) , by sakuma@xcite ( 2.93 for iron and 0.33 for platinum ) , and finally by osterloch _ et al._@xcite ( 2.92 for iron and 0.38 for platinum ) . here again all previous calculations are in good agreement with our full - potential results .
all the methods produced a smaller induced spin moment for platinum atom in fept than for copt , verifying that the hybridization effect in copt is much stronger than in fept .
contrary to spin moments , the orbital moments are anisotropic .
figure [ fig4 ] presents the behavior of the orbital moments as a function of the angle @xmath8 between the [ 001 ] direction and the spin quantization axis within lsda for both copt and fept compounds ( the lines are guide to the eye ) .
the orbital moments decrease with respect to the angle @xmath8 but the values for the [ 100 ] axis does not follow this general trend .
all four lines seem to have the same behavior , but it is interesting to notice that for the copt and magnetization along the [ 100 ] axis cobalt orbital moment is smaller than the platinum one ( for the values see tables [ table1 ] and [ table2 ] ) .
the cobalt moments decrease faster than platinum moments in copt . for the iron and platinum atoms in fept the two lines are practically parallel .
the platinum orbital moments in fept are smaller than for the copt with a factor that varies from 73% for @xmath8=0@xmath9 to 67% for @xmath8=90@xmath9 .
the ratio of iron and platinum orbital moments in fept varies from 1.56 for @xmath8=0@xmath9 down to 1.35 for for @xmath8=90@xmath9 , which is considerably smaller than the ratio of cobalt and platinum orbital moments in copt .
we notice here that in our calculations we can estimate only the projection of the total orbital moment on the spin quantization axis and we have no information concerning the real value of the total magnetic moment or its direction in space .
it seems that for the magnetization along the [ 100 ] direction , the direction of the orbital magnetic moment undergoes a discontinuous jump , resulting in a large projection on the spin quantization axis with respect to @xmath8 at the vicinity of 90@xmath9 .
this behavior is also reproduced by the gga . in table
[ table1 ] we present the values of the orbital moments of iron and cobalt within both the lsda and gga as a function of the angle @xmath8 .
the gga values seem to be smaller than the lsda ones but follow exactly the same trends .
the orbital moment anisotropy is more important in the case of cobalt .
the lsda cobalt orbital moment changes by 0.048 @xmath17 and the gga moment by 0.027 @xmath17 as we pass from the easy axis [ 001 ] to the hard axis [ 100 ] .
the lsda iron moment changes by 0.002 @xmath17 and the gga moment is the same for the two high symmetry directions . in the case of cobalt
the difference between values calculated within the two functionals , lsda and gga , becomes smaller when the angle increases and for 75@xmath9 it changes sign . in the case of iron
this difference decreases only slightly with the angle but since the difference is considerably smaller than in copt ( less than 0.004 ) , we conclude that the orbital moment in the lsda and the gga are roughly the same . in table
[ table2 ] we present the values for the platinum orbital moment within both functionals .
we see that the gga produces larger moments than the lsda for platinum in fept contrary to copt .
the platinum moments are in general smaller than the moments of the 3@xmath5 ferromagnets , and the difference between the values calculated within lsda and gga are small .
the absolute values for platinum are comparable to cobalt(iron ) orbital moments even though the spin moments on platinum are one order of magnitude smaller than for cobalt(iron ) .
the large orbital moments for platinum are due to a much larger spin - orbit coupling for the @xmath5 electrons of the platinum compared to the 3@xmath5 ferromagnets .
the orbital moments of fept and copt have been previously calculated by daalderop and collaborators@xcite and by solovyev and collaborators@xcite for the [ 001 ] direction using the lsda .
the orbital moment of the cobalt site was found to be 0.12 @xmath17 by daalderop and 0.09 @xmath17 by solovyev .
the value of daalderop is closer to our lsda value of 0.11 @xmath17 .
for iron site daalderop found a value of 0.08 @xmath17 and solovyev 0.07 @xmath17 in good agreement with our lsda value .
the platinum orbital moment has been calculated by solovyev .
he found a value of 0.06 @xmath17 for platinum in copt and 0.044 @xmath17 for platinum in fept , close to our values , 0.06 @xmath17 and 0.05 @xmath17 , respectively . on the other hand ,
experimental data are available for copt by grange;@xcite obtained by applying the sum rules to the experimental xmcd spectra .
the sum rules give the moments per hole in the @xmath5-band .
to compare experiment with theory we calculated the number of @xmath5-holes by integrating the @xmath5 projected density of states inside each muffin - tin sphere .
we found 2.63 @xmath5 and 2.48 @xmath5 holes for cobalt and platinum , respectively .
the cobalt orbital moment varies from 0.26 @xmath17 for @xmath8=10@xmath9 down to 0.11 @xmath17 for @xmath8=60@xmath9 .
the measured values are also available for two other angles : 0.24 @xmath17 for @xmath8=30@xmath9 and 0.17 @xmath17 for @xmath8=45@xmath9 .
our theory reproduces qualitatively the experimental trends but underestimates the absolute values by more than 50% ( see tables [ table1 ] and [ table2 ] for all the values ) .
the calculated values show a less sharp decrease with the angle than the experimental ones . for platinum
the experimental data are available only for two angles 10@xmath9 and 60@xmath9 . for @xmath8=10@xmath9
the orbital moment is 0.09 @xmath17 and for @xmath8=60@xmath9 it is 0.06 @xmath17 .
for the platinum site the calculated values are in much better agreement than for the cobalt site but we must keep in mind that the sum rules have their origin in an atomic theory and their use for 5@xmath5 itinerant electrons like in platinum is still debated . the discrepancy between the theory and the experiment comes mainly from the approximation to the exchange and correlation . both the lsda and the gga approximations to the density functional theory
are known to underestimate the orbital moment values , because the orbital moment is a property directly associated with the current in the solid and a static image is not sufficient .
but until now a dft formalism like the current and spin density functional theory ( cs - dft),@xcite which can treat at the same footing the kohn - sham and the maxwell equations is too heavy to implement in a full - potential _ ab - initio method_. the other problem is that there is no form of the exchange energy of a homogeneous electronic gas in a magnetic field known and this is the main quantity entering the cs - dft formalism .
brooks has also developed an _
ad hoc _ correction to the hamiltonian to account for the orbital polarization but this correction originates from an atomic theory and its application to itinerant systems is not satisfactory.@xcite
xmcd spectroscopy became popular after the development of the sum rules that enable the extraction of reliable information on the micro - magnetism directly from the experimental spectra.@xcite the great advantage of xmcd is that we can probe each atom and orbital in the system so to obtain information on the local magnetic properties . lately angle - dependent xmcd experiments allowed the determination of magnetic properties for different spin quantization axis .
all experimental spectra for copt have been obtained by grange _
et al_,@xcite and all calculated spectra presented in this section were obtained using the lsda .
the gga produced the same results and are not presented .
figure [ fig5 ] presents the xmcd spectra for cobalt and iron atoms for the [ 100 ] and [ 001 ] magnetization axis .
we convoluted our theoretical spectra using a lorentzian width of 0.9 ev and a gaussian width of 0.4 ev as proposed by ebert,@xcite in the case of iron to account for the core hole effect and the experimental resolution , respectively .
the energy difference between the @xmath3 and @xmath2 peaks is given by the spin - orbit splitting of the @xmath18 and @xmath19 core states .
it is larger in the case of cobalt , 14.8 ev , than for iron , 12.5 ev .
the intensities of the peaks are comparable for both atoms , but the cobalt peak - intensities are larger for the [ 001 ] axis contrary to the iron spectrum .
the most important feature of these spectra is the integrated @xmath3/@xmath2 branching ratio as it enters the sum rules .
it is larger in the case of cobalt site .
this is expected since the sum rules predict that a larger @xmath3/@xmath2 ratio is equivalent to a larger orbital moment .
for both atoms the integrated @xmath3/@xmath2 branching ratio is larger for the [ 001 ] axis .
but the ratio anisotropy is larger for the cobalt site ( 1.32 for * m*@xmath20[001 ] and 1.17 for * m*@xmath20[100 ] ) than for the iron site ( 1.15 for * m*@xmath20[001 ] and 1.14 for * m*@xmath20[100 ] ) reflecting the larger orbital moment anisotropy of cobalt compared to iron ( see table [ table1 ] ) . especially for iron both orbital moment and integrated @xmath3/@xmath2 branching ratio are practically the same for both magnetization axis . as these changes in the xmcd signal are related to the change in the orbital magnetic moment between the two magnetization directions .
the xmcd anisotropy should be roughly proportional to the underlying mca but no relation exist that connects these two anisotropies .
however they are connected indirectly through the orbital moment anisotropy.@xcite in figure [ fig9 ] we have plotted the platinum xmcd spectra for two angles @xmath8=10@xmath9 and 60@xmath9 for both compounds .
the life time of the core - hole in platinum is smaller than for cobalt so the broadening used to account for its life time should be larger .
we used both a lorentzian ( 1 ev ) and a gaussian ( 1 ev ) to represent this life time and a gaussian of 1 ev width for the experimental resolution .
as in the case of cobalt and iron the platinum xmcd spectra change with the angle and depend on the surrounding neighbors .
the peak intensities are larger in copt .
also the difference in the intensity due to the anisotropy has different sign in the two compounds .
the intensities of fept spectra for @xmath8=10@xmath9 are much larger than for @xmath8=60@xmath9 contrary to the copt behavior . as is the case for cobalt in copt , the platinum site integrated @xmath3/@xmath2 branching ratio shows larger anisotropy than for platinum in fept . in copt
it is 1.49 for @xmath8=10@xmath9 and 1.19 for @xmath8=60@xmath9 , while for fept it is 1.20 for @xmath8=10@xmath9 and 1.14 for @xmath8=60@xmath9 .
here again the xmcd follows the anisotropy of the orbital magnetic moment in these compounds .
the energy difference between the two peaks is 1727 ev for both compounds .
the 2@xmath4 electrons of platinum are deep in energy and are little influenced by the local environment , so that their spin - orbit splitting does not depend on the neighboring atoms of platinum .
we expect a better agreement between the theoretical and the experimental xmcd spectra for the platinum site than for the cobalt site , because the core hole is deeper and would effect less the final states of the photo - excited electron . in figure [ fig6 ] ,
we have plotted the absorption and the xmcd spectra of cobalt for @xmath8=0@xmath9 .
we have scaled our spectra in a way that the experimental and theoretical @xmath3 peaks in the absorption spectra have the same intensity .
the energy difference between the @xmath3 and @xmath2 peaks is in good agreement with experiment .
but the intensity of the @xmath2 peak is larger than the corresponding experimental peak .
the high intensity of the calculated @xmath2 edge makes the theoretical xmcd integrated @xmath3/@xmath2 branching ratio of 1.32 much smaller than the experimental ratio of 1.72 .
this is because the lsda fails to represent the physics of the core hole photo - excited electron recombination . in the case of 3@xmath5 ferromagnets
the core hole is shallow and influences the final states seen by the photo - excited electron .
a formalism that can treat this electron - hole interaction has been proposed by schwitalla and ebert,@xcite but it failed to improve the @xmath3/@xmath2 branching ratio of xas of the late transition metals .
benedict and shirley@xcite have also developed a scheme to treat this phenomena but its application is limited only to crystalline insulators . in figure [ fig7 ]
we have plotted experimental and theoretical total absorptions for the platinum atom in copt .
for both @xmath2 and @xmath3 edges the theory gives a sharp peak which does not exist in experiment . as expected the @xmath3 peak is much more intense than the @xmath2 .
in contrast to what is obtained for cobalt , the results for the platinum xmcd ( see figure [ fig8 ] ) show better agreement with experiment , due to the fact that the core hole effect is less intense ( core hole much deeper compared to cobalt ) .
the experimental and theoretical @xmath2 and @xmath3 edges are separated by a spin - orbit splitting of the @xmath21 core states of 1709 and 1727ev respectively .
the width of both @xmath2 and @xmath3 edges is comparable to experiment , but the calculated @xmath2 edge is much larger .
this produces a calculated integrated branching ratio of 1.49 which is much smaller than the experimental ratio of 2.66 . here
again the theory is underestimating the branching ratio .
in figure [ fig10 ] we present the band structure along the [ 001 ] and [ 100 ] axis in the reciprocal space for different angles @xmath8 for the copt compound within the lsda . we know that it is essentially the area around the fermi level that changes with respect to the spin quantization axis . for this reason
we have enlarged a region of @xmath221 ev around the fermi level . in the first panel
we plot the relativistic band structure for @xmath8=0@xmath9 . in the second and third panel
we have plotted the relativistic band structure for @xmath8=45@xmath9 and @xmath8=90@xmath9 .
we remark that as the angle increases there are bands that approach the fermi level and cross it .
however this information concerns just two high symmetry directions . for this reason we limited ourselves to the changes in the total dos . in figure [ fig11 ]
we notice that just below the fermi level the dos for the hard axis is lower than for the easy axis which seems to favor the hard axis .
this means that the anisotropy does not originate from the changes ate the vicinity of the fermi level but from that of the whole dos .
it is difficult to investigate this phenomena by inspection of the changes at the vicinity of the fermi surface and to explain the sign of the mca .
our results confirm the work of daalderop and collaborators@xcite that argued that not only states in the vicinity of the fermi surface contribute to the mae , as originally thought,@xcite but states far away make an equally important contribution .
we have performed a theoretical _ ab - initio _ study of the magnetic properties of the ordered copt and fept fct alloys systems .
the calculated easy axis is the [ 001 ] for both compounds in agreement with other calculations and with experiments on films which found a strong perpendicular magnetization axis .
the density of states is found to change very little with the direction of the spin quantization axis , and hence the magnetic moments are isotropic with respect to the magnetization axis .
contrary to the spin moments , the orbital magnetic moments decrease with the angle @xmath8 up to 75@xmath9 ( @xmath8 is the angle between the spin quantization axis and the [ 001 ] axis ) .
the calculated x - ray magnetic circular dichroism ( xmcd ) for all the atoms reflect the behavior of the orbital moments .
especially platinum resolved spectra present large differences between the two compounds .
cobalt xmcd spectra are in agreement with experiment but as usual the @xmath3/@xmath2 ratio is underestimated by the theory .
the platinum site shows better agreement with experiments because the core - hole is much deeper than in the case of cobalt .
finally we showed that all the occupied electronic states contribute to the magneto - crystalline anisotropy and not just states near the fermi level for both copt and fept .
we thank j. m. wills for providing us with his fplmto code .
i.g . is supported by an european union grant n@xmath9 erbfmxct96 - 0089 .
the calculations were performed using the sgi origin-2000 supercomputer of the universit louis pasteur de strasbourg and the ibm sp2 supercomputer of the cines under grant gem1917 .
k. r. coffey , m. a. parker , and j. k. howard , ieee trans .
magn . * 31 * , 2737 ( 1995 ) ; s. h. charap , p. -l .
lu , and y. he , ieee trans .
magn . * 33 * , 978 ( 1997 ) ; n. li and b. m. lairson , ieee trans . magn . * 35 * , 1077 ( 1999 ) .
c. brouder and m. hikam , phys .
b * 43 * , 3809 ( 1991 ) ; c. brouder , m. alouani , and k. h. bennamann , phys . rev .
b * 54 * , 7334 ( 1996 ) ; g. y. guo , phys . rev .
b * 57 * , 10 295 ( 1998 ) ; a. ankudinov and j. j. rehr , phys . rev .
b * 56 * , 1712 ( 1997 ) ; a. ankudinov , b. ravel , j. j. rehr , and s. conradson , phys . rev .
b * 58 * , 7565 ( 1998 ) ; g. van der laan and b. t. thole , phys . rev .
b * 43 * , 13 401 ( 1991 ) . .calculated lsda and gga cobalt(iron ) orbital magnetic moments with respect to the angle @xmath8 between the [ 001 ] axis and the spin quantization axis in the ( 010 ) plane . for both atoms
the orbital moments decrease with the angle but values for @xmath8=90@xmath9 do not follow this trend .
both lsda and gga produce the same trends .
the moments at the cobalt site are larger than at the iron site .
the experimental orbital moments for the cobalt site are taken from the ref . .
the theory underestimates the experimental orbital moments by more than 50% but reproduces the correct trends .
[ cols="^,^,^,^,^,^,^,^",options="header " , ] | we have performed a theoretical study of the magnetic circular dichroism in the x - ray absorption spectra ( xmcd ) of the equiatomic copt and fept ordered alloys as a function of the spin quantization axis .
we found that the magnetization axis is along the [ 001 ] direction and the magneto - crystalline anisotropy energy ( mca ) for the fept compound is twice as large as that of the copt compound in agreement with experiment .
the band structure and the total density of states confirm that all electronic states contribute to the mca , and not just the states at the vicinity of the fermi level .
the orbital magnetic moments decrease with respect to the angle between the [ 001 ] axis and the spin quantization axis , and are much larger for the copt compound .
we show that the orbital moment anisotropy is reflected in the xmcd signal . |
quantum field theory ( qft ) on the anti - de sitter ( ads ) space - time has come today to the general attention because of the role that ads geometry plays at several places in modern theoretical physics @xcite .
ads qft is also believed to provide an infrared regularization @xcite that can be useful for instance to understand the longstanding problems of gauge qft s . from a more conceptual viewpoint , this regain of interest has led some authors to realize the need for a deeper setting of ads qft and to investigate the consequences of general principles which , in spite of the difficulties generated by the peculiar geometry of ads , might lead to a reasonable approach to the interacting fields on this spacetime ; such an approach should of course include the case of ads free fields , whose preliminary versions were given in early works @xcite @xcite . in this spirit three recent works on general ads qft can be mentioned .
@xcite uses the framework of observable algebras , but without assuming local commutativity from the outset .
instead the property of `` passivity '' ( in the sense of @xcite ) is postulated for the vacuum , and a remarkable proportion of the more standard properties is shown to follow .
indeed we argue in our final remarks that , heuristically speaking , their assumptions imply those of the present paper .
two other works @xcite and @xcite ( motivated by @xcite ) independently exhibit explicit relationships of a general type between ads qft and conformal qft in the minkowskian boundary of ads : while @xcite also pertains to the framework of local observable algebras , @xcite relies on a limited use of analytic @xmath0-point functions of local fields in a wightman - type approach ( see @xcite ) .
it is the complete setting of such a wightman - type framework for ads qft and the derivation of a number of general results for interacting fields belonging to that framework which are the purpose of the present paper . 0.1 cm it may be useful to start by recalling how ads qft
is rendered difficult by the lack of the global hyperbolicity property of the underlying manifold .
this manifests itself in two ways : there exist closed timelike curves on the ads manifold ( in particular geodesics ) and there is a `` boundary '' at spacelike infinity .
the first problem is commonly avoided by considering the covering of the manifold ; the second fact implies that the standard procedure of canonical quantization for free fields ( see e.g. @xcite ) can not be used , since there does not exist a global cauchy surface and information can flow in from spacelike infinity . to construct a viable qft under these circumstances
one may need to specify suitable boundary conditions at infinity . to this end ,
the nice idea in @xcite was to use the conformal embedding of the ads manifold in the einstein static universe , which is a globally hyperbolic space - time .
it has been then possible ( with some restrictions ) to produce a class of boundary conditions that render the resulting ads qft well defined .
unfortunately the procedure is very special and tricky , and can work at best for free field theories . generally speaking , a well known problem when studying qft on gravitational backgrounds
is the absence of a criterion to select the physically meaningful states .
there are indeed infinitely many inequivalent representations of the same field algebra and it is in general impossible to characterize the physically relevant vacuum states as , for instance , fundamental states for the energy operator , since the very concept of energy as a global quantity is in general not defined .
however , an important aspect of the energy concept which keeps its full value is the notion of energy `` relative to an observer whose world - line is an orbit of a one - parameter group of isometries of the spacetime '' : for such observers , the usual quantum notion of energy , represented by the generator of time - translations , is in fact applicable with respect to the proper - time parameter . therefore the properties of energy - positivity and temperature ( i.e. the notions of ground state and kms - state ) are meaningful _ relatively to such observers _ and technically characterized by relevant analyticity properties of the correlation functions of the fields in the corresponding complexified orbits .
of course , this approach of the energy concept remains particularly simple because we have to deal with curved spacetimes of holomorphic type , such as the de sitter and ads quadrics . in a previous work dealing with qft on de sitter spacetime ( @xcite )
, we had shown that the absence of global energy operators could _ in that case _ be successfully replaced by appropriate global analyticity properties of the @xmath0-point functions of the fields in `` natural '' tuboidal domains : the latter played the same role as the tube domains resulting from energy - positivity in the case of minkowskian fields . moreover a thermal interpretation of these qft for all geodesic ( as well as uniformly accelerated ) observers could then be proved as a byproduct of these global analyticity properties , thus providing an extension of the bisognano - wichmann analyticity property obtained in the minkowskian case @xcite , or in physical terms of the `` unruh effect '' .
surprisingly , in the ads case , the situation concerning energy operators turns out to be more fortunate than in the de sitter case , and in fact quite favourable !
this is because there exist _ two classes _ of time - like orbits of _ `` generic '' one - parameter isometry groups _ , namely the class of elliptic orbits and the class of hyperbolic orbits , also supplemented by a `` boundary - type '' class of parabolic orbits .
the hyperbolic orbits of ads represent uniformly accelerated motions similar to _ all _ the geodesic or uniformly accelerated motions of de sitter spacetime : the corresponding complexified orbits , which are complex hyperbolae , are thus also `` plagued '' by a natural periodicity in the _ imaginary part _ of their proper - time parameter , which forbids the positivity of the corresponding energy operator ( but can support at best a kms - condition whose temperature is related to the radius of the hyperbola and thereby to the acceleration ) . on the contrary , the ( complexified ) elliptic orbits ,
whose class contains geodesic as well as uniformly accelerated motions , only present the peculiarity of periodicity in the _ real part _ of their proper - time parameter : this pathology already mentioned above ( namely the `` time loops '' ) is cured by considering qft on the _ universal covering _ of ads ; in fact in all the following , this covering space will appear as much more natural than ads itself for the setting of interacting fields . at any rate , since there is no geometrical periodicity in the imaginary part of the proper - time of the elliptic orbits and since the corresponding one - parameter groups have no orbits of other type , their generators can be considered as genuine global energy generators : nothing forbids one to postulate the positivity of the corresponding energy operators , expressed technically by relevant analyticity properties in half - planes of the proper - time parameter .
so , as in the wightman axiomatics of minkowskian qft , it is still natural here to postulate the spectral condition for all the generators of time - like elliptic orbits , the latter playing the same role as the time - like straight - lines ( or uniform motions ) in the minkowskian spacetime .
we note that such a condition has been proposed in fronsdal s group - theoretical study @xcite of klein - gordon ads qft . in this connection
one must also quote @xcite whose authors point out the discrepancy between the elliptic and hyperbolic ads trajectories and give arguments for attributing to them respectively a zero - temperature and a finite temperature specified in terms of the acceleration .
another postulate which will play a crucial role in our approach is an adaptation of the property of local commutativity ( or microcausality ) .
it still appears as natural for the covering of ads space - time , in spite of its lack of global - hyperbolicity , while it is harder to justify on the `` pure '' ads space - time itself , because of the time - loop phenomenon .
however we may regard a field theory on the pure ads space - time as just a special case of one defined on its covering .
another justification is also provided by @xcite . after having set the relevant geometrical notions in sections [ pre ] and [ tub ]
, we shall propose in section [ hyp ] a plausible set of hypotheses for an interacting ads qft , among which the positivity of the spectrum of the above mentioned energy operators , ads - covariance and an adaptation of microcausality .
the spectral hypothesis readily implies that the @xmath0-point correlation functions admit analytic continuations in tuboidal domains in the cartesian product of @xmath0 copies of ( the covering of ) the complexified ads manifold .
these analyticity properties of the correlation functions parallel as closely as possible what happens in wightman qft @xcite , where the analyticity of the correlation functions in similar tubular domains is analogously obtained from the positivity of the spectrum of the energy .
these domains are described in section [ tub ] .
unfortunately the @xmath0-point tuboids are rather complicated geometrical objects and , at present , can be given a simple description only in the case @xmath1 , namely for the two - point functions .
note however that for general @xmath0 they contain `` flat domains '' corresponding to all points moving on orbits of _ the same one - parameter isometry group _
, whose description is identical with those of the corresponding sets in complex minkowskian spacetime ; in @xcite , these flat domains have played a useful role in the construction of the `` asymptotic forms '' of ads qft , recognized as lscher - mack - type theories @xcite on the asymptotic cone of ads , in correspondence with conformal qft in minkowskian spacetime of one dimension less . at the end of our section [ tub ] ( subsection [ comptub ] ) ,
we have tried to give a summary of some analogies and discrepancies between the @xmath0-point tuboids of complex ads and the corresponding ones of complex minkowski space .
one of the interesting points of the ads geometry is that there exist families of submanifolds that can be identified with minkowski space - times in one dimension less ( branes ) : as a matter of fact , these submanifolds contain all the two - plane sections of parabolic - type of the ads - quadric mentioned above as the third class of timelike orbits .
this very fact has raised recently a large interest @xcite .
our construction guarantees the possibility of considering restrictions of ads quantum fields to these `` poincar branes '' and obtaining this way completely well - defined minkowskian qft s .
it is perhaps worthwhile to stress that this result , described in section [ parsec ] , is not as obvious as the well - known restrictibility of minkowskian theories to lower dimensionality space - times because of the more complicated geometry . from a geometrical viewpoint
, the conformal theories obtained in @xcite under asymptotic scaling assumptions then also appear as limits of the previous minkowskian qft s when the corresponding parabolic sections tend to infinity . in any approach of qft
, the case of two - point functions deserves a special study .
this is why we give in section [ 2pt ] and in appendix [ ap2pt ] a complete characterization of the ads two - point functions ; the latter are actually maximally analytic , exactly as in the minkowski @xcite and de sitter @xcite cases . as usual
this permits the constructions of generalized free fields and their wick powers which fulfill all the hypotheses .
this study also provides the opportunity of displaying some strange implications of the postulate of microcausality in the `` pure ads '' case .
in fact , the discrepancy between the pure ads spacetime and the covering of ads appears in a characteristic way in the classification of the two - point functions ; it is revealed by the property of uniformity ( or nonuniformity ) of these functions in their analyticity domain @xmath2 $ ] in the complex plane of the cosine of the ads invariant distance .
these phenomena introduce the more general problem of characterizing the interacting qft s on the pure ads spacetime with respect to those on the covering , which will be briefly discussed in our outlook ( section [ o ] ) . in section [ bwa ]
we derive from our postulates the property of bisognano - wichmann analyticity in all the hyperbolic orbits corresponding to the class of uniformly accelerated motions mentioned earlier , or in other words the `` ads - unruh effect '' : an accelerated observer of the ads world whose world - line belongs to that class will perceive a thermal bath of particles with inverse temperature equal to @xmath3 times the radius of the corresponding hyperbolic orbit . we note that this result , first described in a special free - field theory in @xcite , has been also justified in the general framework of algebraic qft in @xcite , by taking the principle of passivity of @xcite as a starting point .
we also note that , in spite of the peculiarities of the global geometry of ads , this result is similar to the one proved in @xcite for minkowskian qft and in @xcite for its extension to de sitterian qft .
section [ wr ] is devoted to the ads version of the `` euclidean '' field theory and to the corresponding osterwalder - schrader reconstruction on the covering of ads .
section [ o ] contains some final remarks among which a brief discussion of the relationship of @xcite and the present work .
we start with some notations and some well - known facts .
let @xmath4 ( resp @xmath5 ) denote @xmath6 ( resp .
@xmath7 ) equipped with the scalar product @xmath8 where @xmath9 denotes @xmath10 .
a vector @xmath11 in @xmath4 is called timelike , spacelike or lightlike according to whether @xmath12 is positive , negative or equal to zero .
we also use the notation @xmath13 for @xmath14 or @xmath15 and we introduce the corresponding orthonormal basis of vectors @xmath16 in @xmath4 ( @xmath17 ) . if @xmath18 is a linear operator in @xmath4 or @xmath5 we put @xmath19 .
we denote @xmath20 ( resp .
@xmath21 ) the group of real ( resp .
complex ) `` ads transformations '' , i.e. the set of real ( resp .
complex ) linear transformations of @xmath4 ( resp @xmath5 ) which preserve the scalar product ( [ 2.1 ] ) , @xmath22 and @xmath23 the connected components of the identity in these groups , @xmath24 and @xmath25 the correponding covering groups . an element of @xmath22 ( resp .
@xmath23 ) will be called a proper ads transformation ( resp . a proper complex ads transformation ) . 0.2 cm the real ( resp .
complex ) _ `` pure '' anti - de - sitter space - time _ @xmath26 ( resp .
@xmath27 ) of radius @xmath28 is defined as the submanifold of @xmath4 ( resp @xmath5 ) consisting of the points @xmath11 such that @xmath29 . except for the thermal considerations in section [ bwa ]
, we shall always take for simplicity @xmath30 throughout this paper .
the group @xmath22 ( resp .
@xmath23 ) acts transitively on @xmath26 ( resp .
@xmath27 ) . by changing @xmath31 to @xmath32 for @xmath33 ,
the complex quadric @xmath27 becomes the complex unit sphere in @xmath7 , which has the same homotopy type as the real unit sphere @xmath34 .
in particular @xmath35 , @xmath36 ( i.e. @xmath27 is simply connected ) for @xmath37 .
it follows that for @xmath38 the _ covering space _ of @xmath27 is @xmath39 itself .
however , as seen below , @xmath26 admits a nontrivial _ covering space _ @xmath40 whose `` physical '' role is that it suppresses the time - loops of pure ads ; its construction will also imply the existence of nontrivial coverings of important domains of @xmath39 ( although the full space @xmath39 itself has a trivial covering ) .
0.2 cm it is possible to introduce in @xmath27 an analog of the so - called euclidean spacetime in complex minkowski space ( where space is real and time purely imaginary ) : we choose it to be the connected real submanifold @xmath41 of @xmath27 defined by putting @xmath42 real , @xmath43 .
this sheet ( @xmath43 ) of the two - sheeted hyperboloid with equation @xmath44 , equipped with the riemannian metric induced by the ambient quadratic form ( [ 2.1 ] ) , will be called _
`` euclidean '' ads spacetime_. this choice singles out the `` base point '' @xmath45 as the analog of the origin in minkowski spacetime . 0.2 cm a concrete way of representing the `` euclidean '' spacetime @xmath41 together with @xmath26 _ and its covering _
@xmath40 is to introduce the diffeomorphism @xmath46 of @xmath47 onto @xmath26 given by @xmath48 ( where @xmath49 is identified to @xmath50 ) .
@xmath41 is obtained by changing @xmath51 into @xmath52 in this representation , namely in the extension @xmath53 of @xmath46 to @xmath54 .
this yields the following parametrization of @xmath41 : @xmath55 the diffeomorphism @xmath56 , defined by lifting @xmath46 on the covering @xmath57 of @xmath47 provides a global coordinate system on @xmath40 .
there also exists an extension @xmath58 of @xmath56 to @xmath59 , whose image is a partial complexification of the covering @xmath40 of ads ; this complexified covering contains the same `` euclidean '' spacetime @xmath41 as @xmath27 .
it is clear that since @xmath22 acts transitively on @xmath26 the diffeomorphism @xmath46 can be transported by any transformation of the group @xmath22 ; it follows that @xmath24 also acts transitively on @xmath40 .
0.2 cm the schwartz space @xmath60 of test - functions on @xmath61 will be defined as the space of functions on @xmath61 which admit extensions in @xmath62 .
a @xmath63 function @xmath64 on @xmath65 belongs to @xmath66 if every derivative of @xmath64 with respect to the ambient coordinates decreases faster than any power of the geodesic distance in the riemannian geometry induced by @xmath67 .
equivalently @xmath68 .
the restriction to @xmath26 of the pseudo - riemannian metric @xmath69 is locally lorentzian with signature @xmath70 .
an elementary description follows from the fact that @xmath22 acts transitively on @xmath26 : it is sufficient to look at the situation in the tangent hyperplane to the base point @xmath71 ( i.e. @xmath72 ) whose intersection with @xmath26 is the light - cone @xmath73 . at the base point
@xmath74 , the future ( resp .
past ) cone is then defined by @xmath75 ( resp .
@xmath76 ) . at any point @xmath77 ,
the tangent hyperplane to @xmath26 is @xmath78 .
its intersection with @xmath26 is the light - cone with apex at @xmath11 , @xmath79 .
we say that a tangent vector @xmath80 at any point @xmath11 is time - like , light - like or space - like according to whether @xmath81 , @xmath82 or @xmath83 , which is consistent with the situation at the base point . defining the local future ( resp .
past ) cone at each point @xmath11 can also be done by using the transitive action of @xmath22 ( or simply by continuity ) starting from @xmath74 , but a more explicit characterization can be given by using the lie algebra of @xmath20 ( see below ) .
it can be easily seen that the circular sections of @xmath26 by the planes parallel to @xmath84 , parametrized by @xmath51 in the representation ( [ 2.11 ] ) have time - like tangents whose future is in the direction of increasing t , for all points @xmath85 ; this clearly exhibits the phenomenon of closed time - loops announced earlier and its suppression by going to the covering of ads .
0.2 cm from a global viewpoint , two events @xmath86 , @xmath87 of @xmath26 are space - like separated if @xmath88 , i.e. @xmath89 . the acausal set of the base point @xmath74 , i.e. the set of points @xmath11 which are space - like with respect to @xmath74 , is then given by @xmath90 let us introduce the causal set of @xmath74 as the complement in @xmath26 of the closure of @xmath91 : @xmath92 @xmath93 can be decomposed into three sets @xmath94 : @xmath95 the regions @xmath96 and @xmath97 can be conventionally called future and past of @xmath74 .
similar regions @xmath98 , @xmath99 , @xmath100 , @xmath101 can be associated with any point @xmath11 of @xmath26 ( again by the transitive action of @xmath22 ) .
the geodesics of @xmath26 are conic sections by 2-planes containing the origin of the ambient space ; if @xmath86 and @xmath87 are two points on a connected branch of geodesic @xmath102 the distance @xmath103 is defined by @xmath104 , where @xmath105 is the angle under which the arc @xmath106 of @xmath102 is seen from the origin ; otherwise stated , it is the parameter of the isometry which transforms @xmath86 into @xmath87 in the minimal subgroup of @xmath20 admitting @xmath102 as an orbit .
@xmath105 is an ordinary angle if @xmath102 is an ellipse , and a hyperbolic angle if @xmath102 is a branch of hyperbola .
the former case is interpreted as `` normal '' time - like separation if @xmath107 ( future and past being distinguished by the sign of @xmath105 ) and one has @xmath108 ; @xmath103 is interpreted as the interval of proper time elapsed between @xmath86 and @xmath87 for the geodesic observer sitting on @xmath102 .
this case is typically realized by the geodesic with equations @xmath109 ( i.e. the curve @xmath110 in ( [ 2.11 ] ) ) .
the second case corresponds to space - like separation ( @xmath111 ) and one has @xmath112 .
it is typically realized by taking the section of @xmath26 by the plane @xmath113 ( or @xmath114 ) .
the ads space - time is not geodesically convex : indeed for any point @xmath11 , all the temporal geodesics which contain @xmath11 also contain the antipodal point @xmath115 , and the proper time interval beetwen @xmath11 and @xmath115 is @xmath116 on each of these geodesics . furthermore
, the temporal geodesics emerging from an event @xmath11 do not cover the full causal region @xmath117 but only @xmath118 . in order to go from @xmath11 to a point @xmath80 in the region @xmath119
, one needs to follow at least two arcs of temporal geodesics , which implies a boost ( i.e. some `` interaction '' ) at the junction of these two arcs .
similar remarks apply as regards causality and geodesics on the covering @xmath40 .
two - planes of @xmath4 containing the origin can always be spanned by a pair of linearly independent orthogonal vectors @xmath120 and classified by considering the possibility for each vector @xmath121 and @xmath122 to be timelike , spacelike or lightlike , which gives six different cases .
to each pair @xmath120 there corresponds an element @xmath123 of the lie algebra of @xmath20 , an isometry group @xmath124 which is a one - parameter subgroup of @xmath22 and a family of parallel two - planes whose sections by ads are conics ( possibly degenerated into straight lines ) invariant under this subgroup ; it follows that each connected component of these sections is either a timelike or a spacelike curve ( or a lightlike straightline in the degenerate case ) .
being interested by the classification of the _ timelike _ curves ( or trajectories ) , we retain three possible cases which reduce to simple models in terms of the basis @xmath125 , by using again the fact that @xmath22 acts transitively on @xmath26 .
0.2 cm i ) _ the elliptic trajectories _ : this is the case when @xmath126 and @xmath127 . since this entails that ( for all @xmath128)@xmath129 , all the corresponding parallel sections of ads ( in a family specified by @xmath120 ) are timelike .
their models are the circular sections by all two - planes parallel to @xmath130 , described above in ( [ 2.11 ] ) : the isometry subgroup associated with them is the rotation group with parameter @xmath51 .
there is one and only one geodesic in each such family ( in the unique plane of the family containing the origin ) . 0.2 cm ii)_the hyperbolic trajectories
_ : this is the case when @xmath126 and @xmath131 . in each family of parallel sections of ads specified by @xmath120 ,
there are two subfamilies .
one of them is composed of spacelike branches of hyperbolae and contains the unique geodesic of the family .
the other one is composed of timelike branches of hyperbolae , interpreted as uniformly accelerated motions : there is no timelike geodesic of that type .
the isometry subgroup associated with such a family is a group of pure lorentz transformation . a model for this family
is given by the sections parallel to the @xmath132plane .
since it will be used repeatedly ( in particular in section [ bwa ] ) we introduce the following notations . for each @xmath133 , we denote @xmath134 $ ] the special lorentz transformation such that @xmath135x)^0 = { \lambda + \lambda^{-1 } \over 2 } x^0 + { \lambda - \lambda^{-1 } \over 2 } x^1,\ \ \ ( [ \lambda]x)^1 = { \lambda - \lambda^{-1 } \over 2 } x^0 + { \lambda + \lambda^{-1 } \over 2 } x^1\ , \label{2.2}\ ] ] the other components of @xmath11 remaining unchanged . in other words @xmath136x)^0 = x^0\ch s + x^1\sh s,\ \ \ \ ( [ e^s]x)^1 = x^0\sh s + x^1\ch s\ . \label{2.3}\ ] ] the corresponding subfamily of timelike orbits on ads is characterized by the following condition to be satisfied by the other components : @xmath137 .
the corresponding orbits are then branches of `` hyperbolae with radius @xmath138 '' , namely with equations @xmath139 0.2 cm iii)_the parabolic trajectories _ : this is the case when @xmath126 and @xmath140 : since this entails that ( for all @xmath128 ) @xmath141 , all the corresponding parallel sections of ads ( in a family specified by @xmath120 ) are timelike or exceptionally lightlike .
their models are the parabolic sections by all two - planes parallel to @xmath142 , admitting the following representation in terms of a translation group parameter @xmath51 : @xmath143 where @xmath144 .
the complete description of the corresponding subgroup of @xmath22 admitting these parabolic orbits will be given and used in sec [ parsec ] ( see eq .
( [ 6.4 ] ) ) .
these timelike sections admit as a limiting case ( for @xmath145 tending to @xmath146 ) the lightlike section by the plane with equations @xmath147 .
this lightlike section gives the only geodesics in the family .
0.2 cm we note that only the elliptic trajectories have nontrivial liftings into the covering of @xmath26 .
the hyperbolic and parabolic trajectories belong to `` a single sheet of @xmath40 '' .
this is connected with the following difference between these families of trajectories : while each family of elliptic trajectories ( resp .
their liftings ) covers the _ whole space _
@xmath26 ( resp .
@xmath40 ) , each family of the two other types is decomposed into two subfamilies which cover disjoint domains of @xmath26 , namely _ wedge - shaped regions _ of the form @xmath148 for the hyperbolic case and _ halves of ads _ of the form @xmath149 for the parabolic case .
it is also interesting to note that by putting @xmath150 in the representations of the _ three families of trajectories _ ( [ 2.11]),([hyperbol]),([parabol ] ) , one obtains curves inside the `` euclidean '' ads spacetime @xmath151 , namely respectively , branches of hyperbolae , circles and parabolae .
the orbits of the second case , which are associated with purely imaginary lorentz transfomations and exhibit periodicity in the corresponding imaginary time parameter are to be compared with those occurring as well in the euclidean space of complex minkowski space as in the ( `` euclidean '' ) hypersphere of the complex de sitter spacetime ( see e.g. ( @xcite ) ) : as shown below in sec [ bwa ] , they correspond to the existence of an unruh effect as in these other two spacetimes . on the other hand ,
the other two classes of orbits are rather to be compared with those of the time - translation groups of minkowskian space ( although not corresponding to geodesical motions on @xmath26 generically ) , as far as they allow a topologically equivalent `` wick - rotation '' procedure to be performed ( see sec [ wr ] and [ parsec ] ) .
0.3 cm _ interpretation of the planar trajectories as uniformly accelerated motions : _ 0.2 cm in this paragraph
we reintroduce the radius @xmath28 of the ads spacetime . if @xmath152 denotes an arbitrary ads trajectory parametrized by its _ proper time _
@xmath51 , the corresponding velocity - vector @xmath153 at the point @xmath154 satisfies the following relations ( in terms of scalar products in the ambient space @xmath155 ) : @xmath156 , @xmath157 .
it follows that the _ ambient acceleration - vector _ @xmath158 satisfies the relation @xmath159 .
the latter together with the ads equation @xmath160 then imply that the _ ads acceleration - vector _ @xmath161 , defined as the @xmath162orthogonal projection of @xmath163 onto the tangent hyperplane to ads at the point @xmath154 , is given by the following formula : @xmath164 . in view of the previous relations , this entails that @xmath165 . _ motions with constant acceleration
_ are those for which @xmath166 or equivalently @xmath167 is independent of @xmath51 .
consider any planar trajectory of ads ; if it is either elliptic or hyperbolic , we can represent it by the equation @xmath168 , where @xmath169 varies in a two - plane @xmath170 and the vector @xmath171 can always be chosen orthogonal to @xmath170 . if @xmath172 , the trajectory is elliptic ( this is the case described by ( [ 2.11 ] ) , with @xmath173 ) . if @xmath174 , the trajectory is hyperbolic ( this is the case described by ( [ hyperbol ] ) , with @xmath175 , @xmath176 ) .
since @xmath163 as well as @xmath177 remain in @xmath170 , and since @xmath178 , there holds the colinearity condition @xmath179 and the condition @xmath180 then yields @xmath181 .
it then follows that @xmath182 .
is independent of @xmath51 .
this motion is therefore uniformly accelerated with the following value of the ads acceleration : @xmath183 .
this leads us to state all the planar trajectories of the ads spacetime correspond to uniformly accelerated motions , with the following specifications : \i ) if the trajectory is elliptic , the corresponding acceleration @xmath184 takes any value between @xmath185 ( i.e.the geodesic case obtained for @xmath186 ) and @xmath187 ( i.e. the case @xmath188 corresponding to ellipses in far - away two - planes ) .
\ii ) if the trajectory is hyperbolic , the corresponding acceleration @xmath189 takes any value between @xmath187 ( i.e. the case @xmath188 corresponding to hyperbolae in far - away two - planes ) and @xmath190 ( i.e. the case @xmath191 corresponding to degenerate hyperbolae ( or `` bifurcate horizons '' ) ) .
\iii ) if the trajectory is parabolic , the corresponding acceleration is @xmath192 [ accel ] to complete the proof , we just have to treat the case of parabolic trajectories , whose prototype is described by eqs ( [ parabol ] ) . introducing the proper time parameter and the radius @xmath28 of ads ,
the latter can be rewritten : @xmath193 , with @xmath194 .
the ambient acceleration - vector is @xmath195 , which is such that @xmath196 and therefore the ads acceleration - vector @xmath164 is such that @xmath197 .
the lie algebra @xmath198 of @xmath20 can be identified with the real vector space of the linear operators @xmath18 on @xmath4 such that @xmath199 with @xmath200 .
hence there is a canonical linear bijection @xmath201 of the space of antisymmetric 2-tensors over @xmath4 onto @xmath198 .
a basis of @xmath198 is provided by @xmath202 , with @xmath203 ; using the standard notation @xmath204 , we write : @xmath205 . in particular @xmath206 and the diffeomorphism @xmath46 ( see eq .
( [ 2.11 ] ) ) can be rewritten as follows @xmath207 with @xmath208 .
the same formula can also be used for describing the lifting @xmath56 of @xmath46 on the covering spaces , provided @xmath51 is allowed to vary in @xmath209 , and @xmath210 is understood as the exponential map of @xmath198 into @xmath24 , and @xmath211 is identified with a point of one of the fundamental domains of @xmath212 .
the same remark applies to the formulae @xmath213 for simplicity , we keep the same notation , and let the context decide on its interpretation .
it is also easy to verify that , with the notations of ( [ 2.3 ] ) , @xmath214 = \exp sm_{10}$ ] . the scalar product ( [ 2.1 ] ) naturally induces a @xmath20-invariant scalar product in the space of the contravariant tensors of order @xmath215 , namely @xmath216 .
this satisfies @xmath217 in particular , given @xmath218 , @xmath219 @xmath220 the square of this 2-dimensional determinant is a 2-dimensional gramian : @xmath221\left[(b^0)^2 + ( b^d)^2\right ] - ( a^0 b^0 + a^d b^d)^2 \nonumber\\ & = & ( ( a , a ) + \vec{a}^2)((b , b ) + \vec{b}^2 ) - ( ( a , b ) + \vec{a}\cdot \vec{b})^2\nonumber\\ & = & ( a , a)(b , b)-(a , b)^2 + \vec{a}^2 \vec{b}^2 - ( \vec{a}\cdot \vec{b})^2 \nonumber\\ & & + \ ( a , a)\vec{b}^2 + ( b , b)\vec{a}^2 - 2(a , b)\vec{a}\cdot \vec{b}\ . \label{2.5.1}\end{aligned}\ ] ] supposing @xmath222 , @xmath223 , and @xmath224 , we find @xmath225 .
if @xmath121 and @xmath122 are continuously varied while @xmath226 , @xmath227 , @xmath120 are kept constant with the above inequalities satisfied , the sign of @xmath228 remains constant . under the same conditions
, @xmath121 can be brought to the form @xmath229 , @xmath230 , by a transformation in @xmath22 .
then , denoting @xmath231 , we have @xmath232 , i.e. @xmath233 is time - like in the minkowski space @xmath234 .
it can be brought to the form @xmath235 by a lorentz transformation acting in the same space , after which @xmath236 , @xmath237 , @xmath238 .
in particular if @xmath121 and @xmath122 are orthonormal , the necessary and sufficient condition for @xmath239 , @xmath240 , @xmath241 , is that the above scalar product ( [ 2.5 ] ) be positive .
suppose @xmath121 and @xmath122 are as above and @xmath242 .
then the two dimensional real vector subspace spanned by @xmath121 and @xmath122 , and any parallel 2-plane , equipped with the metric induced by ( [ 2.1 ] ) , is a euclidean space ( with positive metric ) .
conversely given a 2-plane with strictly positive induced metric , the parallel two dimensional real vector subspace has an orthonormal base @xmath243 and the scalar product @xmath244 can be made positive by changing @xmath122 to @xmath245 if necessary .
we call such a 2-plane time - like , and always regard it as oriented by the 2-form @xmath246 .
the one - parameter subgroup of @xmath22 defined by @xmath247 leaves this 2-plane invariant , and every orbit of this subgroup is contained in a parallel 2-plane .
this subgroup is conjugated in @xmath22 to @xmath248 .
it follows that @xmath249 we denote @xmath250 the subset of @xmath198 consisting of all elements of the form @xmath251 with @xmath252 , @xmath253 , and @xmath254 .
equivalently , @xmath255 we denote @xmath256 the cone generated in @xmath198 by @xmath250 , i.e. @xmath257 .
we note that for all the elements @xmath123 of @xmath250 the corresponding group elements @xmath258 can be considered as belonging either to @xmath22 or to @xmath24 according to whether @xmath259 is regarded as the exponential map of one or the other group .
as mentioned at the beginning of this section , we always identify @xmath22 ( resp .
@xmath23 ) with a subgroup of @xmath260 ( resp .
@xmath261 ) , and its exponential map with the matrix exponential
. 0.2 cm while in the real minkowski space the isotropic subspaces are one - dimensional , the maximal isotropic subspaces in @xmath4 are two - dimensional when @xmath262 .
let @xmath263 be linearly independent and satisfy @xmath264 .
then @xmath265 and there is a @xmath241 such that @xmath266 and @xmath267 , where @xmath268 .
any @xmath14 such that @xmath269 is a linear combination of @xmath121 and @xmath122 .
_ if @xmath270 , @xmath4 has a metric opposite to a minkowskian metric and the isotropic subspaces are one - dimensional .
we assume that @xmath271 .
the vector @xmath121 must have a non - vanishing projection in the 2-plane spanned by @xmath272 and @xmath74 , and can be brought by some @xmath273 to have @xmath274 , @xmath275 .
in the minkowski space @xmath276 , there is a lorentz transformation in the connected component of the identity which brings @xmath121 to be equal to @xmath277 .
after these transformations @xmath122 satisfies @xmath278 .
hence the projection of @xmath122 into the minkowski space orthogonal to @xmath272 and @xmath279 is light - like and can not vanish , since then @xmath122 would be colinear to @xmath121 . in particular @xmath280 .
assuming e.g. @xmath281 , there exists , in the minkowski space generated by @xmath282 , a lorentz transformation in the connected component of the identity which brings @xmath233 to the form @xmath283 ( @xmath268 ) without affecting the subspace generated by @xmath272 and @xmath279 .
at this point , @xmath284 , @xmath285 .
we now apply the transformation @xmath286 given , in the space generated by @xmath287 , by @xmath288 this brings @xmath121 and @xmath122 to the forms @xmath277 and @xmath283 , respectively .
suppose now that @xmath14 satisfies @xmath289 .
then @xmath290 , @xmath291 and the projection of @xmath11 into the subspace orthogonal to @xmath287 has to vanish .
thus @xmath292 .
we finally give an explicit characterization of the local future cone @xmath293 at any point @xmath11 of @xmath26 .
we define @xmath293 as the connected component of @xmath294 which contains the timelike vector @xmath295 . in the case
when @xmath71 , @xmath296 and any @xmath297 can be written as @xmath298 , where @xmath241 belongs to the stabilizer of @xmath74 : @xmath299 and @xmath300 . hence @xmath301 and therefore , for any @xmath77 @xmath302
we denote @xmath303 for @xmath304 , we define @xmath305 [ deftub1 ] the future tuboid @xmath306 and the past tuboid @xmath307 of @xmath27 are defined by @xmath308 ( @xmath250 being defined by eq . ( [ 2.8 ] ) ) .
the tuboids @xmath309 are defined as the universal coverings of @xmath310 .
they are given by the same formula where @xmath311 is now understood as an element of @xmath25 , @xmath312 varies in @xmath313 and @xmath314 as an arbitrary element of @xmath212 .
[ pseuclid ] the `` euclidean '' ads spacetime @xmath41 is contained in @xmath315 _ proof .
_ since @xmath41 is represented by ( [ difflie ] ) where @xmath51 is changed into @xmath316 , all the complex points of this manifold are seen to be obtained as points of @xmath317 by taking @xmath318 , @xmath319 and @xmath320 in ( [ 3.1.1 ] ) .
[ tub1 ] @xmath306 has the following properties : @xmath321 _ proof . _
denote @xmath322 the rhs of ( [ 3.2 ] ) , and @xmath323 the rhs of ( [ 3.2.1 ] ) .
it is clear that @xmath306 , @xmath322 , @xmath323 are invariant under @xmath22 , and that @xmath324 .
we first prove that @xmath325 .
suppose that @xmath326 with @xmath327 , @xmath328 , and @xmath329 .
let @xmath330 be such that @xmath331 .
then @xmath332 , where @xmath333 , @xmath334 , @xmath335 .
thus @xmath336 with @xmath337 and @xmath338 , i.e. @xmath339 , @xmath340 , @xmath341 , hence @xmath342 .
it follows that @xmath343 , and therefore also @xmath344 , belong to @xmath322.we now show that @xmath345 .
suppose that a point @xmath346 belongs to @xmath322 .
there exists a @xmath241 and a real @xmath347 such that @xmath348 and @xmath349 , i.e. @xmath350 .
this can be rewritten as @xmath351 to prove _
, we suppose that @xmath352 .
then @xmath353 implies that both @xmath11 and @xmath80 belong to a `` usual '' minkowski space , hence they can not be both time - like as well as orthogonal .
the same argument excludes @xmath354 .
finally @xmath355 . the form ( [ 3.2 ] ) makes it obvious that @xmath306 and therefore @xmath307 are open subsets of @xmath356 while the definition makes it obvious that they are connected .
@xmath306 and @xmath307 are disjoint since every point @xmath344 of @xmath307 satisfies @xmath357 .
[ z2 ] let @xmath358 .
then for every @xmath328 and @xmath359 , the point @xmath360 is in @xmath361 .
_ by the invariance of @xmath306 under @xmath22 , it suffices to prove the statement in the case when @xmath362 and @xmath319 with real @xmath363 .
we suppose that @xmath364 , @xmath365 , @xmath366 , and @xmath367 .
this implies @xmath368 .
let @xmath369 .
a simple calculation shows that @xmath370\sh^2 t \nonumber\\ & + & 2(b^0 a^d - b^d a^0)\sh t\ , \ch t > 0,\end{aligned}\ ] ] @xmath371\sh t\ , \ch t\nonumber\\ & + & ( b^0 a^d - b^d a^0)(\ch^2 t + \sh^2 t ) > 0.\end{aligned}\ ] ] [ cutplane ] \(i ) the image of the domain @xmath306 ( or @xmath372 under the coordinate map @xmath373 is the cut - plane @xmath374 $ ] .
\(ii ) the image of the domain @xmath375 ( or @xmath376 ) by the ( scalar product ) mapping @xmath377 is the cut - plane @xmath378 .
\(iii ) the map @xmath379 of @xmath375 onto @xmath378 can be lifted to a map of @xmath380 onto the covering @xmath381 of the cut - plane @xmath378 .
_ if @xmath382 , then , by lemma [ tub1 ] , it can be written as @xmath383 , with @xmath241 and @xmath363 .
hence @xmath384 where @xmath385 can be written as @xmath386 hence @xmath387 .
conversely any @xmath388 can be written as @xmath389 , @xmath390 , i.e. @xmath391 is in the image of @xmath306 . _
( ii ) _ let @xmath392 , @xmath393 .
then @xmath394 , @xmath395 , with @xmath396 , @xmath397 , @xmath398 , @xmath399 .
hence @xmath400 where @xmath401 ( for some @xmath241 ) belongs to @xmath306 by lemma [ z2 ]
. therefore @xmath402 by _ ( i)_. conversely any @xmath388 can be written as @xmath389 , @xmath390 , i.e. @xmath403 is in the image of @xmath404 . _
( iii ) _ it is easy to see that elements of @xmath375 such that @xmath405 , @xmath406 , and @xmath407 , @xmath408 , have a scalar product @xmath409 which runs on an elliptic path with foci @xmath410 in the infinite - sheeted domain @xmath411 when @xmath412 and @xmath413 vary in @xmath414 at fixed @xmath415 .
these elliptic paths remain in @xmath411 , which they fully recover , when @xmath416 and @xmath417 vary in their ranges .
similar elliptic paths homotopic to the latter would be generated by starting from more general elements of @xmath375 .
see also appendix [ ap2pt ] for an alternative argument based on a more global representation of the tuboids .
[ natural1 ] the domain @xmath306 is a domain of holomorphy . _
_ as proved in appendix [ ap2pt ] , there exists a biholomorphic map of @xmath306 onto the domain @xmath418 in the usual @xmath419-dimensional complex minkowski space .
this also exhibits the topoplogy of @xmath306 .
we define the minimal @xmath0-point future tuboid @xmath420 as the subset of @xmath421 consisting of all points @xmath422 such that @xmath423 where , for @xmath424 , @xmath397 , @xmath396 , and @xmath398 .
equivalently @xmath425 indeed if @xmath344 is of the form ( [ 3.3 ] ) with @xmath426 then @xmath427 we define the minimal @xmath0-point past tuboid as @xmath428 .
we denote @xmath429 the universal coverings of these sets . however , we shall write simply @xmath430 when no confusion arises .
[ open ] the set @xmath420 is open in @xmath421 . _ proof .
_ we prove , by induction on @xmath0 , the following more detailed statement ( st@xmath431 ) : _ let @xmath432 be such that , for @xmath433 , @xmath434 for every neighborhood @xmath435 of @xmath436 in @xmath437 there is a neighborhood @xmath438 of @xmath344 in @xmath421 such that for every @xmath439 there exists a @xmath440 such that @xmath441 for every @xmath442 . _ we start with the case @xmath443 , for which we give a detailed proof .
the cases @xmath444 will then be more sketchily treated .
since the statement st@xmath445 is invariant under @xmath22 , it suffices to consider the case when @xmath446 with @xmath447 and @xmath448 , @xmath449 .
thus @xmath450 let @xmath451 . if @xmath452 is of the form @xmath453 , with @xmath454 , then @xmath455 where @xmath456 and @xmath457 are continuous functions of @xmath458 .
hence given @xmath459 , there is a @xmath460 such that if @xmath458 is of the form @xmath453 , and @xmath461 , then ( [ 3.8.1 ] ) holds with @xmath462 and @xmath463 , and @xmath464 .
there is a @xmath465 such that , for every @xmath466 such that @xmath467 , there exists a @xmath241 , satisfying @xmath468 , @xmath469 , and such that @xmath470 satisfies @xmath461 and @xmath471 , @xmath472 .
then ( [ 3.8.1 ] ) holds , and @xmath473 , with @xmath474 , @xmath475 .
since @xmath476 , @xmath477 , this implies @xmath478 and @xmath479 .
we now assume that the statement st@xmath480 has been proved for all @xmath481 .
let @xmath482 satisfy ( [ 3.6 ] ) and let @xmath483 be sufficiently close to @xmath344 . by st@xmath445 , @xmath484 with @xmath485 , @xmath486 , @xmath487 respectively close to @xmath412 , @xmath488 , @xmath314 .
the point @xmath489 is close , in @xmath490 , to the point @xmath491 by st@xmath492 the point ( [ 3.9 ] ) can be rewritten as @xmath493 where , for @xmath494 , @xmath495 , @xmath496 , @xmath497 are respectively close to @xmath498 , @xmath499 , @xmath500 .
this proves st@xmath431 .
[ bigs ] we denote @xmath501 ( resp .
@xmath502 ) the subset of @xmath23 ( resp .
@xmath25 ) consisting of all elements of the form @xmath503 , where @xmath504 , @xmath397 , @xmath505 for all @xmath506 .
we denote @xmath507 .
this is also the complex conjugate of @xmath501 .
we define the @xmath0-point future and past tuboids @xmath508 and @xmath509 by @xmath510 note that if @xmath511 , then @xmath512 . if @xmath241 then @xmath513 .
similar properties hold for @xmath502 .
[ g+open](i ) @xmath501 is open in @xmath23.(ii ) @xmath514 .
this lemma is proved in appendix [ g+pf ] .
similar properties hold for @xmath502 .
the tuboids @xmath515 are invariant under @xmath22 , i.e. if @xmath241 and @xmath516 , then @xmath517 .
obviously @xmath518 .
as a consequence of lemma [ z2 ] , @xmath519 . [ bopen ] for any @xmath520
, @xmath508 is open .
_ as in the proof of lemma [ open ] , we prove , by induction on @xmath0,(p@xmath431 ) : _ let @xmath432 be such that , for @xmath433 , @xmath521 then , for any @xmath522 sufficiently close to @xmath344 , there exist @xmath523 and @xmath524 , respectively close to @xmath525 and @xmath526 , such that @xmath527 .
_ we start with @xmath443 and suppose , without loss of generality in view of @xmath22-invariance , that @xmath528 with @xmath529 , and @xmath530 , @xmath531 for all @xmath532 .
assume e.g. @xmath533 and let @xmath534 be sufficiently close to @xmath344 . then @xmath535 is close to @xmath536 and it can , by the proof of lemma [ open ] , be written as @xmath537 , where @xmath538 , @xmath539 and @xmath333 are respectively close to @xmath540 , @xmath541 , and @xmath171 .
thus @xmath542 with @xmath543 . the inductive proof of p@xmath431 for all @xmath544 follows the same line as in the proof of lemma [ open ] .
note that @xmath545 .
another proof of lemma [ bopen ] can be based on lemma [ g+open ] .
the tuboids @xmath546 are defined as the universal covering spaces of @xmath515 .
the transformation @xmath547 = e^{i\pi m_{10}}$ ] belongs to @xmath20 and @xmath23 but not to @xmath22
. indeed @xmath547 m_{0d } [ -1]\linebreak[0 ] = -m_{0d}$ ] .
if @xmath548 , and @xmath549 l [ -1 ] = \ell(u ' \wedge v')$ ] we find @xmath550 , i.e. @xmath551 hence @xmath552 . as a consequence if @xmath553 ( resp .
@xmath420 ) then @xmath554z_1,\ldots,\ [ -1]z_n ) \in \tt_{n+}^*$ ] ( resp .
@xmath555 ) . for @xmath556 and for any permutation @xmath557 of @xmath558 ,
the permuted tuboid @xmath559 is defined by @xmath560 an analogous definition is used for @xmath561 .
two permuted tubes @xmath559 and @xmath562 are called adjacent if @xmath563 is the transposition of two consecutive indices .
an interesting peculiarity of the ads space - time is that ( in contrast to the minkowskian situation ) adjacent permuted tuboids are not disjoint .
the simplest example is [ inters ] let @xmath314 and @xmath564 be real and @xmath565 .
then,(i ) for each @xmath566 , @xmath567.(ii ) for each @xmath566 , @xmath568 . here
@xmath569 .
_ it is obvious that ( ii ) follows from ( i ) , since the hypotheses are invariant under the exchange of @xmath314 and @xmath564 . to prove ( i ) , we may assume that , in the coordinates @xmath185 , @xmath570 , @xmath419 ( all others kept equal to 0 ) , @xmath571 let @xmath572 be real and satisfy @xmath573 and @xmath574 .
then @xmath575 @xmath576 hence @xmath577 .
let @xmath566 and @xmath578 , @xmath579 .
for sufficiently small @xmath580 , since @xmath501 is open , we have @xmath581 , and @xmath582 so that @xmath583 .
similarly @xmath584 . [ disj ]
denote @xmath585 then the sets @xmath586 and @xmath587 are disjoint . if @xmath588 , each of them is connected .
lemma [ inters ] can be generalized to @xmath0-point tubes as follows : [ intersn ] let @xmath589 be such that @xmath590 . then for any choice of @xmath591 , @xmath592 or @xmath593 , and of @xmath566 ,
the point @xmath344 such that @xmath594 belongs to @xmath508 .
0.25 cm therefore points of the form ( [ t2.27 ] ) belong to the intersection of @xmath508 with the permuted tuboid @xmath595 obtained by exchanging the indices @xmath596 and @xmath597 .
let @xmath598 .
this is an open subset of @xmath61 which has two connected components if @xmath599 , only one otherwise . as a result of lemma [ intersn ] ,
the intersection @xmath600 is an open tuboid which has connected components bordered by those of @xmath601 .
0.25 cm _ proof of lemma [ intersn ] .
_ we may again assume that @xmath602 for sufficiently small @xmath603 , and @xmath604 , let @xmath51 and @xmath605 be given by ( [ t2.25 ] ) , and let @xmath606 , @xmath607 .
then @xmath608 furthermore @xmath609 on the other hand `` opposite '' tuboids such as @xmath508 and @xmath509 do not intersect since they are respectively contained in @xmath610 and @xmath611 .
the family of one - parameter subgroups @xmath612 of @xmath24 whose generator @xmath123 belongs to the set @xmath250 of @xmath613 can be considered as the analog of the family of timelike translation groups @xmath614 acting on minkowskian spacetime , whose generator @xmath121 belongs to the unit hyperboloid shell @xmath615 , but of course there is a major difference : while in the latter case , this family of one - parameter groups form a commutative subgroup @xmath616 of the group of spacetime translations , the groups of the former family are all mutually noncommutative .
nevertheless , the analogy between the two families has provided us with the basic idea for defining the @xmath0-point tuboids in the ads case .
in fact , if in the definitions of @xmath420 ( resp .
@xmath508 ) one replaces the lie elements @xmath396 by @xmath617 ( resp .
@xmath618 by @xmath619 ) and the points @xmath500 on ads by points in minkowskian spacetime , one exactly reobtains the usual @xmath0-point tubes of complex minkowski space as they are defined in @xcite
. the most obvious ( and unpleasant ) effect of noncommutativity in the ads case is that for @xmath556 the description of the tuboids remains very implicit , since the defining conditions involve group elements in a heavy way instead of being characterized directly by equations on the ads manifold ; in particular it is not clear whether the `` minimal tuboids '' @xmath420 are really smaller than the ( `` complete '' ) @xmath0-point tuboids @xmath508 , which will be used in a natural way for expressing the spectral condition of ads quantum fields ( see below in sec .
[ hyp ] ) .
we wish however to emphasize a simple result which displays a reassuring analogy between the ads tuboids and the minkowskian tubes , namely the inclusion of regions of the corresponding `` euclidean '' spacetimes in these domains .
in fact , one has the following property , in which @xmath620 denotes the extension to @xmath621 of the diffeomorphism @xmath622 ( see subsection [ pureetc ] after formulae ( [ 2.11 ] ) and ( [ parameucl ] ) ) .
[ tubeucl ] for each n , the tuboid @xmath623 contains the image by @xmath624 of the following `` flat tube '' : @xmath625 in particular @xmath623 contains the following open subset of @xmath626 : @xmath627 the proof is readily obtained by putting @xmath628 and @xmath629 \sqrt { 1 + \vec x_j^2})$ ] , ( @xmath630 ) in ( [ 3.3 ] ) and changing the sequence @xmath631 into @xmath632 .
the subset ( [ final3 ] ) of @xmath623 which is then exhibited is clearly ( in view of ( [ parameucl ] ) ) a subset of @xmath626 .
this result will entail the validity of a `` wick - rotation procedure '' for the _ universal covering _ of ads spacetime ( see sec .
in fact , we note that the flat tubes of lemma [ tubeucl ] represent domains in the complexification ( in the time variable ) of minkowski space as well as @xmath40 : these domains are isomorphic . in the pure ads
spacetime , this isomorphism does not exist since one has to consider quotients of the previous flat tubes corresponding to the geometric periodicity conditions under the transformations @xmath633 as described in ( [ 3.3 ] ) .
0.2 cm another close analogy between the tuboids @xmath430 and the corresponding minkowskian @xmath0-point tubes will be displayed in subsection [ dbw ] : it will be proved that there exists a special subset of real points in @xmath634 enjoying the same property as the jost points of minkowski space , namely all the complex points obtained by the action of appropriate _
one - parameter complex subgroups _ of @xmath23 on these `` jost points '' are contained in @xmath635 and this is the starting point of analytic completions ( of the type of glaser - streater s theorem ) which are crucial for proving the bisognano - wichmann property ( see section [ bwa ] ) .
0.3 cm we shall now mention two interesting discrepancies between the tuboids @xmath420 and the corresponding @xmath0-point tubes of complex minkowski space .
the first one , which is developed in section [ parsec ] , concerns the parabolic trajectories already introduced in section [ pre ] ; the corresponding complexified curves turn out to exhibit sections of the tuboids @xmath420 which lie in the complexified `` poincar sections of ads '' .
these peculiarities are at the origin of the fact that qft s in these poincar sections can be generated as restrictions of qft s on ads .
no analogs of such families of complexified trajectories in isotropic ( or `` lightlike '' ) hyperplanes exist in complex minkowski space .
0.2 cm finally , we have exhibited above in subsection [ permtub ] a property of pairs of permuted tuboids in @xmath26 which is definitely new with respect to the corresponding pairs in complex minkowski space . in the latter case , such tubes always have an empty intersection and the property of a common analytic continuation for pairs of functions analytic in `` adjacent permuted tubes '' necessitates the application of the edge - of - the - wedge theorem to the boundary values from these two domains through an appropriate `` coincidence region '' . in the present case ,
the situation is somewhat simpler , since according to lemma [ inters ] it is a general fact that _ adjacent pairs of permuted tuboids in @xmath27 have nonempty intersections_.
as usual , it is possible to formulate the main assumptions in terms of distributions ( the test - functions having then compact supports ) or tempered distributions .
we denote @xmath636 the space of test - functions on @xmath61 or @xmath65 .
this may be either @xmath637 ( resp .
@xmath638 ) or @xmath60 ( resp .
@xmath66 ) .
the borchers algebra @xmath639 on @xmath26 ( resp .
@xmath40 ) is the complex vector space of terminating sequences of test - functions @xmath640\ \ldots)$ ] , where @xmath641 and @xmath642 for all @xmath643 , the product and @xmath644 operations being given by @xmath645 the action of @xmath241 ( resp .
@xmath24 ) on @xmath639 is defined by @xmath646 , where @xmath647 it will also be useful to denote @xmath648 where @xmath642 and @xmath525 belong to @xmath22 or @xmath24 . if @xmath557 is a permutation of @xmath649 , and @xmath642 we define the function @xmath650 by @xmath651 the theory of a single scalar quantum field theory on @xmath26 ( resp .
@xmath40 ) is specified by a continuous linear functional @xmath652 on @xmath639 , i.e. by a sequence @xmath653 ( resp .
@xmath653 ) , called wightman functions , with the following properties : \1 .
* covariance : * each @xmath654 is invariant under the group @xmath22 ( resp .
@xmath24 ) , i.e. for all @xmath655 for all @xmath241 ( resp .
@xmath24 ) . \2 .
* locality : * @xmath656 if @xmath657 and @xmath658 are space - like separated .
* positive definiteness : * for each @xmath659 , @xmath660 .
explicitly , given @xmath661 @xmath662 , then @xmath663 if these conditions are satisfied the gns construction ( see @xcite ) provides a hilbert space @xmath664 , a continuous unitary representation @xmath665 of @xmath22 ( resp .
@xmath666 ) and a representation @xmath667 ( by unbounded operators ) as well as a unit vector @xmath668 , invariant under @xmath669 , such that @xmath670 for all @xmath659 . as a special case the field operator @xmath671 is the operator valued distribution over @xmath26 ( resp .
@xmath40 ) such that @xmath672 where @xmath673 .
in addition the construction provides vector valued distributions @xmath674 such that @xmath675 where @xmath676 . as usual , for every @xmath241 ( resp .
@xmath24 ) , @xmath677 the details of this construction are completely analogous to those of the minkowskian case . to every element @xmath123 of the lie algebra @xmath198 we can associate the one - parameter subgroup @xmath678 of @xmath22 ( resp .
@xmath24 ) and a self - adjoint operator @xmath679 acting in @xmath664 such that @xmath680 for all @xmath681 . with these notations ,
we postulate the following \4 .
* strong spectral condition : * for every @xmath682 , every @xmath683 , and every @xmath63 function @xmath684 with compact support contained in @xmath685 , @xmath686 equivalently @xmath679 has its spectrum contained in @xmath687 . using the spectral decomposition of @xmath679 ,
this implies that , for every @xmath683 , @xmath688 extends to a function , again denoted @xmath689 , continuous on @xmath690 and holomorphic in @xmath313 , and bounded in norm by @xmath691 .
for any finite sequence @xmath692 of elements of @xmath256 , the function @xmath693 is therefore continuous and bounded in norm by @xmath691 on the `` flattened tube '' @xmath694 it is holomorphic in @xmath695 in @xmath313 when the other @xmath696 are kept real . by the flattened tube ( malgrange - zerner ) theorem
, this function extends to a continuous function on @xmath697 , holomorphic in @xmath698 , and bounded in norm by @xmath691 .
thus the function @xmath699 extends to a bounded holomorphic function on @xmath501 ( resp .
@xmath502 ) with continuous boundary value on @xmath22 ( resp .
@xmath24 ) .
we now consider @xmath700 where , for @xmath424 , @xmath701 , @xmath505 , and @xmath702 .
suppose that @xmath703 where @xmath704 .
then ( [ 4.11 ] ) extends to a function of @xmath705 which is @xmath63 on the flattened tube @xmath706 and holomorphic in @xmath707 in @xmath313 when the other @xmath708 are kept real . for every @xmath709 , the restriction of this function to @xmath710
is bounded in modulus by @xmath711 where @xmath712 is one of the seminorms defining the topology of @xmath713 .
the envelope of holomorphy of the set ( [ 4.13 ] ) contains the topological product @xmath714 therefore the function ( [ 4.11 ] ) extends to a @xmath63 function on @xmath715 , holomorphic in @xmath716 and bounded in modulus by ( [ 4.14 ] ) on the set @xmath717 for every @xmath709 .
for every @xmath718 , the value of ( [ 4.11 ] ) defines a continuous @xmath0-linear form on @xmath719 , hence , by the nuclear theorem , a unique continuous linear functional on @xmath636 .
we conclude that , for a general @xmath642 , the function ( [ 4.11 ] ) extends to a @xmath63 function on @xmath715 , holomorphic in @xmath716 . by standard arguments
it follows that there exists a function @xmath720 , holomorphic in @xmath420 , having @xmath721 as its boundary value in the sense of distributions .
the same proof ( but with a more cumbersome notation ) shows that @xmath720 is holomorphic in @xmath508 .
in the remainder of this paper we will require the wightman functions @xmath721 to be tempered distributions on @xmath61 ( resp .
@xmath65 ) , i.e. we will take @xmath722 ( resp .
@xmath66 ) , and we will assume , instead of the `` strong spectral condition '' , that the following holds : \5 .
* tempered spectral condition : * for each pair of integers @xmath723 and @xmath724,@xmath725 is the boundary value , in the sense of tempered distributions , of a function @xmath726 of @xmath727 holomorphic and of tempered growth in @xmath728 .
( in particular @xmath729 is the boundary value of a function @xmath720 holomorphic and of tempered growth in @xmath508 . )
moreover for any @xmath642 and every choice of @xmath730 the function defined by ( [ 4.11 ] ) is @xmath63 and at most of polynomial growth in @xmath731 .
if the positive definiteness condition holds , the tempered spectral condition implies the strong spectral condition .
however the tempered spectral condition makes sense even if the positive definiteness condition does not hold .
the following lemma follows from arguments given in @xcite ( sect .
5 ) , using as the main tool a theorem of v. glaser @xcite .
[ locpos ] let @xmath652 be a wightman functional satisfying the conditions of covariance and locality and the tempered spectral condition .
suppose in addition that there is a real open set @xmath732 such that @xmath660 for all @xmath659 with support in @xmath438 ( i.e. such that @xmath733 has support in @xmath734 for all @xmath735 ) .
then there exists , for each @xmath520 , a vector valued function @xmath736 , holomorphic in @xmath508 , and with tempered growth at infinity and near the boundaries , having @xmath674 as a boundary value in the sense of tempered distributions .
in particular @xmath652 satisfies the unrestricted positive definiteness condition , and the reeh - schliedertheorem holds . the permuted wightman functions @xmath737 defined , as usual , by @xmath738 \langle \ww_n\,,\linebreak[0]\ _
\pi f_n\rangle$ ] , are boundary values of functions @xmath739 holomorphic in the permuted tuboids @xmath740 . owing to local commutativity , they are branches of a single holomorphic function .
[ extloc]if a set of wightman functions satisfies the tempered spectral condition but only a part of the locality condition , i.e. if it is assumed that @xmath721 and @xmath741 coincide in an open subset of @xmath601 ( necessarily symmetric under the exchange of @xmath596 and @xmath597 ) , then it follows from lemma [ intersn ] that they coincide in the whole of @xmath601 .
similar extension theorems are well - known in the minkowskian case ( see e.g. @xcite ) due to phenomena of analytic completion .
it is remarkable that no completion is needed in the ads case .
if the conditions 1 - 5 hold , one may want to assume \6 .
* uniqueness of the vacuum : * the invariant subspace of @xmath664 , @xmath742 is one - dimensional , i.e. it is equal to @xmath743 . in the minkowskian case this condition is equivalent to a clustering property of the wightman functions , namely the truncated wightman functions tend to zero when a proper subset of their arguments tend to space - like infinity while the others remain bounded . ( the truncated wightman functions have the same inductive definition as in the minkowskian case ( @xcite p. 66 ) .
they have the same linear properties as the wightman functions . )
a similar equivalence holds in the anti - de sitter case . in sect .
[ parsec ] , we discuss some properties of this type , which are equivalent to the uniqueness of the vacuum in the presence of positivity .
a convenient chart of a part of @xmath26 ( resp .
@xmath356 ) is provided by the parabolic coordinates @xmath744 given by @xmath745 in this equation @xmath746 , @xmath747 are the coordinates of an arbitrary event in a real ( resp .
complex ) @xmath748-dimensional minkowski space - time with metric @xmath749 , @xmath750 and @xmath751 ( resp .
@xmath752 ) .
this explains why the coordinates @xmath753 of the parametrization ( [ poinc ] ) are also called poincar coordinates . as @xmath754 vary in @xmath755 , the image of this map is @xmath756 . the scalar product and the ads metric
can then be rewritten as follows : @xmath757 eq .
( [ 7 ] ) implies that @xmath758 for a given real @xmath145 we denote @xmath759 the parabolic section @xmath760 the subgroup @xmath761 ( resp .
@xmath762 ) of @xmath22 ( resp .
@xmath23 ) which fixes @xmath763 , and therefore leaves @xmath759 globally invariant , is isomorphic to the real ( resp .
complex ) poincar group operating on the @xmath748-minkowski space . in particular , for @xmath764 , the transformation @xmath765 operates in the minkowski leaf as @xmath766 .
if @xmath767 , this transformation leaves the coordinates @xmath768 unchanged and , in the 3-dimensional space of the coordinates @xmath769 , is given by @xmath770 here @xmath771 similarly @xmath772 where @xmath773 .
with these notations , @xmath774 we denote @xmath775 if @xmath171 is a real point of @xmath759 , i.e. @xmath776 for some real @xmath777 and @xmath145 , and @xmath778 then @xmath779 is a point of the future tube in the complexified @xmath759 considered as a minkowski space .
conversely any point of the future tube is of this form .
0.25 cm more generally , we define the @xmath0-point forward tuboid @xmath780 as @xmath781 the intersection of this set with @xmath782 is the set obtained by restricting the @xmath500 to lie in @xmath759 in the above definition .
this intersection is just the @xmath0-point forward tube in the minkowskian variables @xmath783 .
the main point of this section is [ partube ] for all @xmath0 , @xmath784 .
the proof of this lemma consists of lemmas [ 1tube ] and [ ntube ] below .
we begin with the following remark .
[ ulim]let @xmath785 then @xmath786 for @xmath773 , and @xmath787 therefore @xmath788 and , for @xmath773 , @xmath789 @xmath790 @xmath791 [ 1tube ] let @xmath792 and @xmath793 , and let @xmath794 .
then(i ) there is an @xmath682 , and a @xmath333 , arbitrarily close to @xmath795 and @xmath171 , respectively , such that @xmath796.(ii ) for every neighborhood @xmath435 of @xmath797 in @xmath798 , there is a neighborhood @xmath438 of @xmath799 in @xmath356 such that every @xmath800 can be written as @xmath801 , with @xmath802 and @xmath803 .
_ ( i ) it suffices to prove the statement in case @xmath804 for some @xmath805 , and @xmath806 . in the coordinates @xmath807 , for any @xmath808 ,
@xmath792 , @xmath809 as expected the two last components add up to @xmath810 .
we now set @xmath806 and @xmath811 ( @xmath805 ) in ( [ 6.7 ] ) .
we look for @xmath123 in the form @xmath812 with @xmath813 as in ( [ 6.6.1 ] ) and @xmath814 large .
then @xmath815 the set of all vectors @xmath816 with pure imaginary 0-component and all other components real is mapped into itself by @xmath817 for all real @xmath818 and @xmath51 .
as @xmath818 tends to @xmath190 , @xmath819 tends to @xmath820 provided @xmath821 , and @xmath817 tends to @xmath822 provided @xmath823 . for fixed real @xmath824 , satisfying @xmath825 , and real @xmath826 , we wish to find a real @xmath827 and @xmath363 such that @xmath828 , i.e. @xmath829 . the condition that @xmath830 gives @xmath831 the other components of @xmath832 are then real and tend to those of @xmath171 as @xmath833 .
this proves ( i ) .
\(ii ) using ( i ) , @xmath834 , where @xmath682 and @xmath333 can be chosen arbitrarily close to @xmath795 and @xmath171 respectively .
if @xmath452 is sufficiently close to @xmath344 , we may apply st@xmath445 of the proof of lemma [ open ] , which yields the conclusion of ( ii ) .
the condition that @xmath835 in lemma [ 1tube ] can be replaced by the condition @xmath836 , and in fact by @xmath837 since the case @xmath838 can be dealt with by changing @xmath171 to @xmath839 .
we note also that @xmath840 .
we define @xmath841 ( the same calculations as in ( [ 2.5.1 ] ) show that if @xmath842 , @xmath843 , then @xmath844 ) .
the following lemma extends lemma [ 1tube ] to the case of @xmath0 points
. it may be useful to spell it out in some detail .
[ ntube ] let @xmath432 be such that , for @xmath424 , @xmath845 for each @xmath846 , there exists a @xmath847 such that , for any @xmath848 satisfying @xmath849 for all @xmath850 , there exist @xmath851m_n \in \cc_+$ ] and @xmath524 such that @xmath852 , @xmath853 , and @xmath854 for all @xmath850 . _ proof .
_ let st@xmath855 denote the statement of the lemma .
st@xmath856 follows from lemma [ 1tube ] in view of the preceding remarks .
assume that st@xmath857 has been proved for all @xmath858 and let @xmath859 be as in ( [ 6.13 ] ) .
let @xmath459 , @xmath860 .
denote @xmath861 , @xmath862 , @xmath863 , and @xmath864 by st@xmath856 , there exists a @xmath460 such that for any @xmath865 with @xmath866 , there exist @xmath867 and @xmath868 such that @xmath869 , @xmath870 , @xmath871 , and @xmath872 .
let @xmath873 satisfy @xmath874 , and let @xmath488 , @xmath487 be as above .
then , for @xmath875 , @xmath876 by st@xmath877 , there exist @xmath878 and @xmath879 such that , for @xmath880 , @xmath881 , @xmath882 , and @xmath883 .
this proves st@xmath855 .
we digress at this point and take advantage of the above calculations to recall the proof of the following lemma ( which appears in @xcite ) .
[ bobu ] let @xmath669 be a continuous unitary representation of @xmath22 in a hilbert space @xmath664 .
let @xmath683 be such that @xmath884 = \psi$ ] for all real @xmath818 .
then @xmath885 for all @xmath241 .
obviously @xmath886 can be replaced , in the statement of lemma [ bobu ] , by any of its conjugates under @xmath22 , e.g. @xmath887 etc .
0.25 cm _ proof of lemma [ bobu ] .
_ by remark [ ulim ] @xmath888 here @xmath889 is any real .
suppose that @xmath683 is such that @xmath890 for all real @xmath818 .
then for all real @xmath818 and @xmath51 , @xmath773 , @xmath891 we set @xmath892 for some fixed real @xmath889 .
as @xmath893 , @xmath51 tends to 0 , so that the rhs of this equation tends to 0 . by the continuity of @xmath669 and ( [ g.4 ] ) , the lhs tends to @xmath894 hence this quantity is equal to 0 . by the continuity of @xmath669 , the subgroup @xmath895 of @xmath22 is closed ,
hence it is a lie subgroup ( see e.g. @xcite , pp . 228 ff ) .
its lie algebra contains @xmath886 and @xmath896 for all @xmath897 , hence all @xmath898 for @xmath899 .
these elements generate the whole of @xmath198 since @xmath900 = m_{\nu\mu}$ ] for @xmath901 .
hence @xmath902 .
lemma [ partube ] follows , as announced , from lemmas [ 1tube ] and [ ntube ] . if @xmath903 is a sequence of wightman functions satisfying the conditions of sect .
[ hyp ] , but not necessarily the positive definiteness condition , the tempered distribution @xmath721 can be restricted to @xmath782 , and more generally to @xmath904 .
the distributions @xmath905 z({{\mathbf{z}}}_n,\ v_n))$ ] have the linear properties of the @xmath0-point wightman functions for a set of minkowskian fields on @xmath906 , @xmath907 , labelled by a real parameter @xmath145 and depending in a @xmath63 manner on @xmath145 . the usual minkowskian covariance and analyticity properties are satisfied by virtue of lemma [ partube ] .
the local commutativity is inherited from that postulated for the @xmath721 in view of the formula ( [ lll ] ) . if the @xmath908 satisfy the positivity condition for all @xmath145 in a non - empty open interval @xmath243 , then by lemma [ locpos ] ( sect .
[ hyp ] ) , the original @xmath721 satisfy the positivity condition on the whole of @xmath26 . in case
the positive definiteness condition holds , we also have [ limspec ] let @xmath909 and let @xmath910 denote the self - adjoint operator on @xmath664 such that @xmath911 for all @xmath681 .
then the spectrum of @xmath910 is contained in @xmath912 .
more precisely , for every @xmath683 , and every @xmath63 function @xmath684 on @xmath209 with compact support contained in @xmath685 , @xmath913 _ proof .
_ let @xmath914 .
we may assume that @xmath915 and @xmath916 .
given @xmath917 , let @xmath918 be such that @xmath919 .
let @xmath438 be a neighborhood of the identity in @xmath22 ( or @xmath24 ) such that @xmath920 for all @xmath921 .
it is possible to choose @xmath682 such that @xmath922 for all @xmath923 $ ] .
then @xmath924 since the first term in the rhs is zero . under the same hypotheses , we note that if @xmath683 is invariant under @xmath925 then it is in particular invariant under @xmath926 , so that , by lemma [ bobu ] , it is invariant under @xmath927 . let again @xmath907 , which we consider as an operator valued distribution on the minkowski space @xmath906 depending smoothly on the real parameter @xmath145 .
by the reeh - schlieder theorem for the field @xmath671 , the vacuum is cyclic for the set of fields @xmath928 , where @xmath243 is any non - empty open interval in @xmath209 . as a consequence ,
the uniqueness of the vacuum for the original theory is equivalent to the uniqueness of the vacuum for the fields @xmath18 , hence to the cluster property for their wightman functions , which is also a certain cluster property for the wightman functions of @xmath671 .
we note that the borchers - buchholz lemma also provides the following characterization of the uniqueness of the vacuum in terms of the wightman functions .
assume that the conditions 1 - 5 of sect .
[ hyp ] hold ( this includes the positivity condition )
. then the condition of uniqueness of the vacuum is equivalent to @xmath929 _
proof_. if all the conditions 1 - 5 of sect . [ hyp ] hold , ( [ u.1 ] ) is equivalent to @xmath930 by the mean ergodic theorem , the limit in the lhs is equal to @xmath931 , where @xmath932 is the projector on the subspace of all vectors invariant under @xmath933 . by lemma [ bobu ]
this is the subspace of all vectors invariant under @xmath927 .
therefore ( by the density of @xmath934 ) the condition ( [ u.1 ] ) is equivalent to the fact that the subspace of all vectors invariant under @xmath927 is @xmath743 .
we now consider the two - point function of a scalar field theory on @xmath212 which satisfies the general requirements described in the previous sections : @xmath935 by the tempered spectral condition , this is the boundary value on @xmath936 , in the sense of tempered distributions , of a function @xmath937 , holomorphic and of tempered growth in @xmath938 .
the permuted wightman function @xmath939 is the boundary value of @xmath940 , holomorphic in @xmath941 .
the two permuted wightman functions are invariant under @xmath24 and coincide in the real open subset of @xmath936 in which @xmath86 and @xmath87 are space - like separated
. therefore @xmath942 are branches of a single holomorphic function @xmath943 .
extensions of standard arguments ( see appendix [ ap2pt ] ) show that there exists a function @xmath944 , holomorphic on the universal covering @xmath945 of @xmath946 $ ] , such that @xmath947 when @xmath948 and @xmath949 belong to @xmath950 and @xmath951 , respectively .
theories on the _ covering _ of anti - de sitter spacetime are thus in this respect closely similar to minkowski @xcite and de sitter @xcite field theories . in the case of a field theory on @xmath212 , the commutator function can be written ( non uniquely ) as the difference of a retarded and an advanced `` function '' with supports in @xmath952 , respectively ( see ( [ 2.11 ] ) ) . in the case of a field theory on
the _ pure _ ads spacetime @xmath26 , @xmath944 is actually a function holomorphic on @xmath378 , and , in particular , @xmath953 and @xmath954 coincide not only on @xmath955 , but also on the `` exotic region '' @xmath956 . in this case
the support of the commutator function @xmath957 is contained in @xmath958 .
this is the union of the two closed sets @xmath959 and @xmath960 , and the commutator function can be written as the difference of an advanced and retarded function with supports in these two sets ( respectively ) . this splitting is , as usual , not unique .
any two - point function with the above mentioned properties is the two - point function of a generalized free field on @xmath212 or @xmath26 , which satisfies the positive definiteness condition if and only if , for any @xmath63 function @xmath961 with compact support in @xmath212 ( resp .
@xmath26 ) , @xmath962 wick powers of such a field are well - defined : their wightman functions can be obtained by the standard formulae , using @xmath435 , as boundary values of holomorphic functions in the relevant tuboids .
they satisfy all the requirements of sect .
[ hyp ] with the possible exception of positive definiteness , which holds if and only if ( [ 2pt.2 ] ) holds .
note that two wick powers of generalized free fields on @xmath26 share the property of commuting when their arguments are in @xmath963 .
let us consider fields satisfying the ads klein - gordon equation on @xmath964 : @xmath965 there is a preferred choice of solutions of that equation that display the simplest example of the previous analytic structure .
the corresponding two - point functions are expressed in terms of generalized legendre functions @xcite @xmath966 by the following formula : @xmath967 where the parameter @xmath968 is related to the mass by the formula @xmath969 the normalization can be obtained by imposing the local hadamard condition , that gives @xmath970 it can be checked directly that the functions @xmath971 defined by eqs ( [ kgtp1 ] ) , ( [ glf ] ) , ( [ hd+1 ] ) are holomorphic in @xmath378 or @xmath972 according to whether @xmath968 is an integer or not .
0.2 cm it is useful to display also an equivalent expression for the wightman functions ( [ kgtp1 ] ) in terms of the usual associate legendre s function of the second - kind ) with the upper index in parentheses .
this is the way these wightman functions were first written in @xcite for the four dimensional case ( @xmath973 ) .
their identification with second kind legendre functions is worth being emphasized again , in place of their less specific ( although exact ) introduction under the general label of hypergeometric functions , used in recent papers .
in fact legendre functions are basically linked to the geometry of the ds and ads quadrics from both group theoretical and complex analysis viewpoints @xcite ] @xcite : @xmath974 where we introduced the parameter @xmath975 such that @xmath976 theories with @xmath977 are acceptable in the sense that they satisfy all the axioms of section [ hyp ] including the positive - definiteness property , which we shall prove below .
there are in fact two regimes @xcite : \i ) for @xmath978 the axioms uniquely select one field theory for each given mass ; \ii ) for @xmath979 there are two acceptable theories for each given mass .
the case @xmath980 is a limiting case .
( [ kgtp ] ) shows that the difference between the theories parametrized by opposite values of @xmath981 is in their large distance behavior .
more precisely , in view of eq .
( 3.3.1.4 ) of @xcite , we can write : @xmath982 the last term in this relation is _
regular on the cut _
@xmath983 $ ] and
therefore does not contribute to the commutator . by consequence
the two theories represent the same algebra of local observables at short distances .
but since the last term in the latter relation grows the faster the larger is @xmath984 ( see @xcite eqs .
( 3.9.2 ) ) , the two theories drastically differ by their long range behaviors .
let us discuss for now the positive definiteness of the wightman function ( [ kgtp ] ) . to this end , following the remarks in section [ parsec ] we can consider the restriction @xmath985 of the two - point function @xmath986 to @xmath987 defined for @xmath988 real and @xmath989 , @xmath990 such that @xmath991 , @xmath992 , @xmath993 and @xmath994 ( see eq .
( [ 7a ] ) ) .
the results of section [ parsec ] imply that this restriction defines a local and ( poincar ) covariant two - point function which satisfies the condition of positivity of the energy spectrum @xcite .
let us consider now the following hankel - type transform of @xmath995 w.r.t .
the brane coordinate @xmath996 : @xmath997 with @xmath998 inversion gives @xmath999 where @xmath1000 ( see @xcite , p. 64 , formula ( 12 ) , and @xcite for more details )
. 0.2 cm eq .
( [ aaa ] ) together with lemma [ locpos ] show that the wightman functions [ kgtp ] are positive - definite .
we postpone to a further paper the proof of klln - lehmann - type representations which will be based on an appropriate laplace - type transformation , in a spirit similar to that given in @xcite for the case of two - point functions on de sitter spacetime . 0.2 cm for two - point functions of general fields on @xmath40 : @xmath1001 for two - point functions of general fields on @xmath26 : @xmath1002 in these equations , @xmath1003 ( resp .
@xmath1004 ) represents a positive measure ( resp .
sequence ) , as a consequence of the positive - definiteness property .
in this section we prove that our assumptions of locality and tempered spectral condition imply a kms - type analyticity property of the wightman functions in the complexified orbits of one - parameter lorentz subgroups of @xmath22 : the physical interpretation of this property can be called an `` ads - unruh effect '' , since the relevant hyperbolic - type orbits we are considering are trajectories of uniformly accelerated motions . in fact
, this property is similar to the analyticity property in the complexified orbits of the lorentz boosts proved by bisognano and wichmann @xcite in minkowskian theory .
however , in contrast with the method used in @xcite , our method does not rely at all on the positivity properties of the theory but only makes use of an analytic completion procedure of geometrical type which has been presented in @xcite for the case of de sitter spacetime ( noting there that it applied similarly to the minkowskian case ) and which can be adapted in a straightforward way to the ads case as we show below . throughout this section
, the distinction between the cases of pure ads and covering of ads will be irrelevant because all the regions and corresponding analyticity properties considered will always take place in a _ single sheet _ of @xmath40 ; so we can speak here of ads without caring about that distinction ; in particular the notation @xmath430 will denote here as well the corresponding tuboids @xmath429 if the spacetime considered is @xmath40 .
0.2 cm to be specific , let us consider once for all the lorentz subgroup of @xmath22 whose orbits are parallel to the two - plane of coordinates @xmath1005 , with the notations of eqs ( [ 2.2]),([2.3 ] ) ( all the other lorentz subgroups of @xmath22 acting on ads being conjugate of the latter by the action of @xmath22 itself )
. the following `` wedge - shaped '' region of ads , which is invariant under that subgroup is then distinguished : @xmath1006 ( note that this region of ads admits the base point @xmath74 as an exposed boundary point ) . considering @xmath0-point test - functions @xmath642 and lorentz transformations @xmath134 = [ e^s ] \in { \bf r}\setminus \{0\}$ ] we abbreviate @xmath1007\}}$ ] into @xmath1008 , i.e. @xmath1009x_1,\ldots,\ [ \lambda^{-1}]x_n ) .
\label{5.9}\ ] ] then the main result of this section is the following [ bw ] if a set of wightman functions satisfies the locality and tempered spectral conditions , then for all @xmath1010 , @xmath1011 and @xmath1012 , there is a function @xmath1013 holomorphic on @xmath1014 with continuous boundary values @xmath1015 on @xmath1016 from the upper and lower half - planes such that , for all @xmath1017 , @xmath1018 the thermal interpretation of this theorem could be presented with all the details of theorem 3 of @xcite , but here we shall only dwell on the following remarks , specific to the geometry of ads .
consider an ads spacetime with radius @xmath28 ( i.e. with equation @xmath1019 and base point @xmath1020 ) and assume that all the test - functions @xmath1021 considered in the statement of theorem [ bw ] have their supports contained respectively in sets of the form @xmath1022 , @xmath1023 , where @xmath1024 is a neighbourhood in @xmath26 of a certain point @xmath1025 in the wedge @xmath1026 which we take of the following form : @xmath1027 @xmath121 being a positive number .
the lorentzian orbits followed by the points in the supports of @xmath1028 and @xmath1029 are in a neighbourhood of the orbit of @xmath1025 which is the branch of hyperbola with equations @xmath1030x(a ) = ( ra\,\sh { t\over ra},\ ra\,\ch { t\over ra},\ 0\ldots,0,\ r \sqrt { a^2 + 1})\ .
\label{groupx(a)}\ ] ] in the latter , the normalization of the group parameter @xmath1031 has been chosen such that @xmath51 is the _ proper time _ along the trajectory of @xmath1025 with respect to the metric of the ads spacetime of radius @xmath28 .
expressed in this parameter @xmath51 such that @xmath1032 , the analyticity property stated in the theorem corresponds to the analyticity in a kms - type domain , namely a @xmath1033-periodic cut - plane generated by the strip @xmath1034 , with the period @xmath1035 ( accompanied by the relevant conditions of kms - type for the boundary values at @xmath51 real and @xmath1036 real in terms of the products of field - observables ) .
thus an observer living on the trajectory ( [ groupx(a ) ] ) , and whose measurements are therefore supposed to be performed in terms of field observables whose support at @xmath1037 lies in the neighbourhood @xmath1024 of @xmath1025 , will perceive a thermal bath of particles at the temperature @xmath1038 . in view of lemma [ accel ] , the motion of such an observer is uniformly accelerated and its acceleration @xmath1039 is related to the `` radius '' @xmath1040 of the hyperbolic trajectory by the following formula ( noting that the parameter @xmath1041 of lemma [ accel ] is @xmath1042 ) : @xmath1043 . therefore one has : [ temperature ] the temperature @xmath1044 perceived by an ads - unruh observer living on the trajectory described by ( [ groupx(a ) ] ) is given in terms of his ( or her ) acceleration @xmath1039 by @xmath1045 0.2 cm the proof of theorem [ bw ] contains two steps ( as the similar theorem 2 of @xcite ) . in subsection [ jostpt ]
we introduce special sets of points in @xmath634 which play the role of the jost points of the minkowskian case , because all the complex lorentz transformations @xmath134 $ ] transport them into @xmath1046 .
this is the starting point of a standard analytic completion procedure which generates the analyticity of the wightman functions in domains obtained by the action of all complex lorentz transformations on the tuboids @xmath420 ( or @xmath1047 ) as in lemma 2 of @xcite . here also , there is ( for each pair @xmath1048 considered in the statement of theorem [ bw ] ) a quartet of wightman functions analytic respectively in the tuboids @xmath430 and in two other tuboids @xmath1049 which are involved in that procedure .
this is fully explained in subsection [ dbw ] .
let @xmath385 be given by @xmath1050 with the notations of ( [ 2.2 ] ) and ( [ 2.3 ] ) , and real @xmath1051 , @xmath1052\,a = e^{ism_{10}}\,a = ( i\sh u \sin s,\ \sh u \cos s,\ \vec{0},\ \ch u)\ . \label{5.2}\ ] ] this can be reexpressed as @xmath1053 with @xmath1054 , i.e. @xmath1055\,a & = & ( ir\sh t,\ \sh u \cos s,\ \vec{0},\ r\ch t),\nonumber\\ \th t & = & \th u \sin s,\nonumber\\ r & = & \sqrt{(\ch u)^2 -(\sh u)^2 ( \sin s)^2 } \label{5.3}\end{aligned}\ ] ] let @xmath1056 be the set of points @xmath1057 such that @xmath1058 let @xmath1057 be of the form ( [ 5.4 ] ) .
then for @xmath1059 , @xmath1060\ a_j = \exp(it_j m_{0d})b_j,\nonumber\\ & & \th t_j = \th u_j \sin s,\ \ \ j = 1,\ldots , n,\nonumber\\ & & 0 < t_1 < \cdots < t_n\ , \label{5.5}\end{aligned}\ ] ] with @xmath1061 .
since @xmath498 can be rewritten as @xmath1062 with @xmath1063 , the point @xmath1064a$ ] is in @xmath420 . by the invariance of @xmath420 under @xmath22 ( resp .
@xmath24 ) it follows that @xmath134 a \in \zz_{n+}$ ] for all @xmath1065 . obviously @xmath134 a \in \zz_{n-}$ ] for all @xmath1066 , since @xmath428 .
this subsection closely follows the treatment given in @xcite .
we refer the reader to that reference for more details .
let @xmath1067 if @xmath1068 and @xmath43 , we denote ( consistently with the notation @xmath1026 introduced in ( [ wedgeshape ] ) ) @xmath1069 let @xmath1070 @xmath1071 is defined by reflecting @xmath1072 across the hyperplane @xmath1073 or equivalently @xmath1074\kk_{n+}$ ] with @xmath547 = [ e^{i\pi}]$ ] as in ( [ 2.3 ] ) .
if @xmath1075 , then @xmath1076 and @xmath1077 are space - like separated whenever @xmath1078 .
note that @xmath1079 . if @xmath1080 belongs to @xmath61 or @xmath421 ,
we denote @xmath1081 .
besides @xmath420 and @xmath1047 we shall also use the other two tuboids @xmath1082 we fix @xmath723 , @xmath643 and a function @xmath1083 with support in @xmath1084 .
there exist two functions @xmath1085 and @xmath1086 , respectively holomorphic in @xmath420 and @xmath428 , having boundary values @xmath1087 on @xmath61 ( resp .
@xmath65 ) in the sense of distributions , such that for every @xmath1088 with compact support , @xmath1089 @xmath1090 ( with @xmath40 instead of @xmath26 when needed ) . the functions @xmath1091 and @xmath1092 are respectively holomorphic in @xmath1093 and @xmath1094 .
their boundary values at real points , in the sense of distributions , are @xmath1095 and @xmath1096 . in the sense of distributions , @xmath1097 and @xmath1098
coincide for @xmath1099 by virtue of local commutativity .
hence , by the edge - of - the - wedge theorem , @xmath1100 and @xmath1101 have a common holomorphic extension @xmath1102 in @xmath1103 , where @xmath1104 is a complex neighborhood of @xmath1071 such that @xmath134\vv_n = \vv_n$ ] for all @xmath1105 .
let @xmath1106 . as we have noted @xmath1064a \in \zz_{n+}$ ] for all @xmath1051 .
moreover @xmath1107a$ ] is in @xmath1104 .
hence there is an @xmath917 ( depending on @xmath121 ) such that @xmath1064a$ ] is in @xmath1108 for all @xmath1109 .
this also means that , denoting @xmath1110a \in \zz_{n+}$ ] , all points of the compact set @xmath1111a'\ : \ 0 \le s \le \pi\}$ ] belong to @xmath1108 .
since the latter is open , there exists a @xmath1112 such that the open ball @xmath1113 is contained in @xmath420 , and @xmath1064b_{a'}(\rho_a ) \subset \delta_n$ ] for all @xmath1114 $ ] hence @xmath134 b_{a'}(\rho_a ) \subset \delta_n$ ] for all @xmath1115 .
the function @xmath1116z)$ ] is holomorphic in @xmath1117 where @xmath1118 for all @xmath1119 .
moreover @xmath20 is also holomorphic in @xmath1120 . by lemma 3 of @xcite ( appendix a )
, @xmath20 extends to a function holomorphic in @xmath1121 .
however we wish to prove that the boundary values of @xmath20 as @xmath344 tends to the reals ( in the sense of distributions ) are also analytic in @xmath968 in @xmath313 . as a consequence of the temperedness assumption , for real @xmath968
, @xmath1122 defines a holomorphic function of tempered behavior in @xmath420 with values in the functions of @xmath968 bounded by some power of @xmath1123 . if @xmath1124 , @xmath1125 extends to a function analytic in @xmath313 and also bounded there by some power of @xmath1123 .
let @xmath914 , where @xmath684 is a @xmath63 function with compact support contained in @xmath685 .
for @xmath1124 , and sufficiently large @xmath504 , @xmath1126 the l.h.s of this equation is holomorphic and of tempered behavior in @xmath420 , hence it vanishes together with its boundary values .
therefore the boundary value @xmath1127 extends to a function of @xmath968 holomorphic in @xmath313 .
we note that @xmath1128\ [ \lambda]x)$ ] for @xmath1129 and @xmath1130x)$ ] for@xmath1131 . in the same way
@xmath1132 and @xmath1133 have a common extension @xmath1134 holomorphic in @xmath1135 with @xmath1136 .
note that the domain @xmath1137 is equal to @xmath1138 .
hence @xmath1139z)$ ] extends to a function holomorphic in @xmath1140 , and its boundary value @xmath1141 has properties that mirror those of @xmath1127 .
finally suppose that @xmath1142 .
then , by local commutativity , for @xmath1143 , @xmath1144x)$ ] coincides with @xmath1145x_\leftarrow ) = { f'}^{(b)}_+(f_m;\ [ \lambda]x)$ ] . hence if @xmath1142 , @xmath1146 and @xmath1147 have a common analytic continuation in @xmath1148
this ends the proof of theorem [ bw ] .
[ bwrem]the tuboids @xmath430 and @xmath1049 can be replaced in the above discussion by @xmath515 and @xmath1149 ( similarly defined ) . in the proof of the preceding subsection ,
if we assume that @xmath1028 has support in the left , instead of the right wedge , we find that @xmath1150 extends to a function holomorphic in @xmath1151 instead of @xmath1152 .
thus in case @xmath1153 it is easy to obtain the following special case of the glaser - streater theorem : [ bwm=0 ] if a set of wightman functions satisfies the locality and tempered spectral conditions , then for all integer @xmath643 , there exists a function @xmath1154 , holomorphic and of tempered growth in @xmath1155 , such that @xmath1156z_1,\ldots,\ [ \lambda]z_n ) \ \ { \rm for}\ \lambda > 0 , \nonumber\\ g_n(z_1,\ldots,\ z_n,\ \lambda ) & = & \ww_n([\lambda]z_n,\ldots,\ [ \lambda]z_1 ) \ \ { \rm for}\ \lambda < 0 . \label{5.16}\end{aligned}\ ] ] if we now assume the covariance condition ( [ 4.3 ] ) holds , then @xmath1157 is actually independent of @xmath968 and we obtain [ ctp ] if a set of wightman functions satisfies the locality , covariance , and tempered spectral conditions , then for all integer @xmath643 , and all @xmath1158 , @xmath1159z_n,\ldots,\ [ -1]z_1 ) . \label{5.17}\ ] ] if positivity holds this implies , as usual , the existence of an anti - unitary operator @xmath105 such that @xmath1160x)^*$ ] . in @xcite ,
the existence of this operator is a non - trivial step in the derivation of commutativity for opposite wedges .
note that the above proof of lemma [ ctp ] , also valid in minkowski space , does not require the bargmann - hall - wightman lemma .
in this section , we consider qft s on the _ covering _ of ads which satisfy the _ tempered spectral condition _ and we wish to show that such theories can be formulated equivalently in terms of theories on the `` euclidean '' ads spacetime @xmath41 in a way which is reminiscent of the wick rotation in complex minkowski space .
in fact , the simple geometrical fact which allows the latter to hold is the property of the tuboids @xmath623 described in lemma [ tubeucl ] and obtained by making use of the complexification in @xmath312 of the map @xmath1161 of @xmath1162 into @xmath40 . as a matter of fact , making @xmath312 and @xmath9 complex yield charts for a certain complexification denoted @xmath1163^{(c)}$ ] of @xmath40 : it is defined as the image of the extension of @xmath56 to the set @xmath1164 .
( of course this set @xmath1163^{(c)}$ ] is the covering of a complex region which is not the full space @xmath27 , since @xmath1165 ) . in the sequel
we omit the symbol @xmath1166 and identify a point @xmath1167 with its image @xmath1168 . according to lemma [ tubeucl ] , all points @xmath1169^{(c)}\right)^n$ ]
such that @xmath1170 is real and @xmath1171 , are in @xmath1172 .
therefore ( using the global invariance under @xmath24 ) @xmath1173 extends to a function of @xmath1174 , holomorphic in the tube @xmath1175 , of tempered growth at infinity and at the boundaries , with values in the tempered distributions ( in fact the polynomially bounded @xmath63 functions ) in @xmath1176 .
( this function depends only on the differences of the @xmath707 variables , but this does not , of course , hold for the @xmath1170 . ) together with the other permuted wightman functions , this defines a function of @xmath1174 , holomorphic in @xmath1177 .
we denote @xmath1178 for real @xmath1179 such that @xmath1180 , for all permutations @xmath557 of @xmath649 .
standard analytic completions , using the edge - of - the - wedge theorem and bremermann s continuity theorem , show that @xmath1181 is actually analytic at all non - coinciding points of @xmath1182 .
geometrically , in view of the representation ( [ parameucl ] ) of @xmath1183 , this means that _ the functions @xmath1184 appear as @xmath0-point schwinger functions defined at all noncoinciding points of the `` euclidean '' ads spacetime @xmath1183 _ , namely @xmath1185 by construction , @xmath1186 for every permutation @xmath557 . in case the positive definiteness condition holds , the sequence @xmath1187 has the osterwalder - schrader positivity property : if @xmath1188 and , for @xmath1189 , @xmath1190 has its support in @xmath1191 , then , with the convention @xmath1192 , @xmath1193 this follows from the existence of the vector - valued holomorphic functions @xmath736 ( see sec .
[ hyp ] ) .
note that the analytic completion mentioned above applies to these vector - valued functions .
0.2 cm conversely , let @xmath1187 be a sequence of functions defined at all noncoinciding points of @xmath1183 , namely on the images of the sets @xmath1194 , symmetric , invariant under a common translation of the variables @xmath1195 , and satisfying the osterwalder - schrader positivity property ( [ 7.5 ] ) when the supports of the @xmath1196 are as above .
then it is possible , by the same method as in the `` flat '' case ( @xcite ) , to construct a hilbert space @xmath664 , @xmath668 , and a sequence of functions @xmath1197 with values in @xmath664 on @xmath1198 , holomorphic in the @xmath707 and @xmath63 in the @xmath1170 , such that @xmath1199 whenever @xmath1200 and @xmath1201 .
the temperedness of the @xmath736 at the real points , hence the existence of boundary values , can be obtained if , as we shall assume , the sequence @xmath1187 satisfies suitable growth conditions , as in @xcite .
note that the boundary values of the scalar products of the @xmath736 , namely the distributions @xmath721 thus obtained are automatically defined on the covering space @xmath65 of @xmath634 , since the analytic continuation in the complex variables @xmath1202 corresponds to travelling in the covering space @xmath1203^{(c)}\right)^n$ ] ( no periodicity condition in the variables @xmath707 being produced in general ) .
if the functions @xmath1204 are invariant under the transformations of @xmath25 which preserve @xmath1205 , then differential operators representing @xmath198 can be used as in @xcite to show that the @xmath721 are invariant under @xmath24 .
finally the analyticity of the @xmath721 shows that the operator @xmath1206 representing @xmath1207 in @xmath664 has positive spectrum and so does the representative @xmath679 of any @xmath328 by invariance under @xmath24 .
first of all , we wish to emphasize the peculiar thermal aspects of `` generic '' field theories on ads or its covering , which are strongly related to the geometry of the ads quadric , namely to the existence of uniformly accelerated motions on three types of trajectories : while the elliptic and parabolic observers perceive a world at zero temperature ( satisfying a condition of energy - positivity ) in spite of the fact that the acceleration can take all values between zero and @xmath187 , the hyperbolic observers perceive an unruh - type temperature effect growing from zero to infinity when the acceleration grows from @xmath187 to infinity . our subsequent remarks concern some important peculiarities of qft s on the _ pure _ ads spacetime with respect to those on its covering , which appeared in the study of general two - point functions ( see section [ 2pt ] ) , namely : \i ) the vanishing of the commutator vacuum expectation value in the region of spacelike separation @xmath1208 implies its vanishing in the region of time - like separation @xmath1209 , which we called the `` region of exotic causal separation '' , \ii ) the periodicity condition on time - like geodesics implies that the two - point function is a holomorphic function of @xmath1210 in the domain @xmath1211 $ ] itself instead of its covering . in the free - field case , this selects the legendre functions @xmath1212 for which @xmath1213 integer , which results in a quantization of the mass parameter @xmath1214 .
more generally all the fields on the pure ads spacetime have two - point functions which are ( positive ) superpositions of the free - field functions @xmath1215 with integer indices @xmath201 .
these peculiarities suggest respectively the following questions which may be linked together and deserve to be studied in further works .
\a ) does local commutativity , formulated as the vanishing of the commutator in the region of spacelike separation implies its vanishing in the region of exotic causal separation ( or `` exotic causality '' ) ?
\b ) do the axioms of section [ hyp ] applied to the pure ads spacetime exclude the existence of non - trivial interacting qft ? it has in fact been noted in @xcite that the region of spacelike separation does not really deserve that name in the pure ads case , since pairs of points in that region can also be connected by classes of _ non - geodesic _ timelike paths ; as it is stated , the condition of local commutativity therefore appears as a strong constraint which would force the interactions to respect the time - periodicity of the geometry .
one can then also say that the condition of exotic locality would represent a constraint of similar nature acting at odd number of half - periods of time , and that is why one could possibly expect it to be a consequence of local commutativity .
we know that , besides generalized free fields , their wick polynomials are allowed to exist , since their @xmath0-point functions , which are combinations of products of two - point functions , do satisfy the required ( periodic ) analyticity properties in the relevant tubes @xmath1216 of the pure ads spacetime . however , if we think of nontrivial interacting fields in terms of perturbation theory and consider feynman - type convolution products of free two - point functions involving not only products but integrals on the ads spacetime , one can show that the constraints of periodic analyticity can not be maintained in general : in particular the action of retarded propagators can be seen to propagate the interaction in the covering tubes @xmath1217 .
so one could formulate as a conjecture , that there might be no other qft s on the pure ads spacetime than the ones previously mentioned , which would then also entail ( as a byproduct ) a positive answer to our question a ) . however , there is at the moment no genuine proof relying on the axioms of section [ hyp ] that such a conjecture is valid .
we end these remarks with a brief comparison of the formalism of this paper with that proposed by @xcite .
a difficulty is of course the usual gap between a formulation based on fields and one based on local algebras .
this difficulty is compounded in the ads case by the lack of a translation group . in @xcite the pure ads , and not its covering space , is considered . from the principle of passivity of the vacuum , the authors succeed in deriving the positivity of the energy , the existence of a ctp operator ( both in the same sense as here ) , and the commutativity of operators localized in opposite wedges . if we suppose that , in terms of fields , these results correspond to the tempered spectral , covariance , and positivity conditions , and a portion of local commutativity , then , by remark [ extloc ] ( section [ hyp ] ) , this implies full local commutativity , and thus all our axioms .
we begin by considering a complex function @xmath1218 holomorphic and of polynomial growth on @xmath1219 .
the function @xmath1220 defined by @xmath1221 is holomorphic in @xmath1222 .
we denote @xmath1223 the boundary value of @xmath1224 on @xmath26 in the sense of tempered distributions .
we suppose that @xmath1225 and @xmath1226 coincide on the real open subset @xmath1227 of @xmath1228 defined by @xmath1229 since we are ultimately interested in the case when @xmath1230 for all @xmath241 , we make the simplifying assumption that @xmath1231 has a boundary value which is @xmath63 in @xmath86 and of tempered growth in @xmath1232 when @xmath1233 tends to 0 while @xmath1234 . actually the general case can be reduced to the simplified one by considering @xmath1235 for suitable test - functions @xmath961 .
the functions @xmath1236 defined by @xmath1237 are respectively holomorphic and of tempered growth in @xmath1238 , and have boundary values @xmath1239 on @xmath26 in the sense of tempered distributions .
these boundary values coincide in the real open subset @xmath1240 of @xmath26 defined by @xmath1241 we also define @xmath1242 let @xmath1243 ( resp .
@xmath1244 ) denote the subgroup of @xmath22 ( resp .
@xmath23 ) which leaves @xmath74 unchanged .
this is just the connected real ( resp .
complex ) lorentz group for the @xmath419-dimensional minkowski space @xmath1245 ( resp .
@xmath1246 ) . the properties mentioned above for @xmath1236 are invariant under @xmath1243 .
let @xmath1247 the last equality is due to the fact that @xmath1248 , where @xmath1249 , @xmath1250 , @xmath1251 for @xmath1252 , and that @xmath1253 is a bijection of @xmath1254 onto @xmath1255
. obviously @xmath1256 .
we also denote : @xmath1257 @xmath1258 @xmath1259 the intersection of @xmath26 with the light - cone with apex at @xmath1260 is contained in the hyperplane @xmath1261 , in fact @xmath1262 note also that @xmath1263 .
indeed we know that the image of @xmath1238 under the map @xmath1264 is @xmath1265 $ ] . as a consequence @xmath1266 .
indeed the image of @xmath1267 under @xmath373 is the same as that of @xmath1238 since @xmath1268 by definition .
we will prove : [ gs ] let @xmath1236 be functions respectively holomorphic and of tempered growth in @xmath1238 , @xmath1269 their boundary values , in the sense of tempered distributions , on @xmath26 .
we suppose that @xmath1270 and @xmath1271 coincide in @xmath1240 .
then:(i ) there exists a function @xmath64 holomorphic on @xmath1267 whose restriction to @xmath1238 is @xmath1236.(ii ) @xmath1272$]}. in particular @xmath1267 contains @xmath1240 and @xmath1273.(iii ) if @xmath1274 is invariant under @xmath1243 , i.e. if @xmath1275 in the sense of distributions for all @xmath1276 , then there exists a function @xmath1277 holomorphic on @xmath378 such that @xmath1278 for all @xmath1279 .
= by -12.000by 2 = by -12.000by 2 _
proof_(i ) we will reduce this statement to a simple case of the glaser - streater theorem @xcite by using the inversion ( stereographic projection ) @xmath1280
@xmath1281 @xmath1282 is a holomorphic involution of @xmath1283 onto itself and commutes with every element of @xmath1244 .
if @xmath1284 , and @xmath1285 , we find @xmath1286 which imply @xmath1287 .
conversely if @xmath1288 , it follows from ( [ a2pt.4 ] ) that @xmath1289 .
therefore @xmath1282 restricted to @xmath1290 is a holomorphic bijection onto @xmath1291 . moreover (
if @xmath1292 ) @xmath1293 @xmath1294 @xmath1295 assume that @xmath1296 .
then @xmath1297 , hence @xmath1298 .
moreover @xmath1299 , @xmath1300 , @xmath1301 , and @xmath1302 hence @xmath1303 .
thus @xmath1304 belongs to @xmath1305 , the future tube in the complex minkowski space @xmath1306 , and in fact @xmath1307 .
similarly @xmath1282 maps @xmath1255 into the past tube of @xmath1306 .
conversely suppose that @xmath1308 .
then @xmath1309 where @xmath1310 is given by the last equality in ( [ a2pt.4 ] ) .
it satisfies @xmath81 by ( [ a2pt.8 ] ) , and @xmath1311 since otherwise we would have @xmath1312 by ( [ a2pt.9 ] ) . for real @xmath1313 and @xmath1314
, the preceding formulae specialize to @xmath1315 as a consequence @xmath1316 the extended tube @xmath1317 in @xmath1306 is defined as usual as @xmath1318 .
obviously @xmath1319 ( recall that @xmath1320 ) .
as it is well - known @xmath1321 hence @xmath1322 , and also @xmath1323 .
furthermore @xmath1324 let @xmath1325 .
then @xmath1326 is holomorphic and of tempered growth in @xmath1327 , where @xmath1328 , @xmath1329 being given by ( [ a2pt.7 ] ) .
the boundary values if @xmath1326 coincide in @xmath1330 so that , by the edge - of - the - wedge theorem , @xmath1326 have a common holomorphic extension in @xmath1331 , where @xmath1332 is a complex open neighborhood of @xmath1333 invariant under @xmath1243 . by following the steps of the proof of the glaser - streater theorem
as presented e.g. in @xcite it is immediate to see that all the arguments apply to the present case , owing to the fact that @xmath1334 is invariant under @xmath1244 .
we therefore conclude that @xmath1326 have a common extension @xmath1335 holomorphic in @xmath1336 .
note in particular that @xmath1335 is holomorphic in a neighborhood of @xmath1337 , hence that the boundary values of @xmath1326 coincide there .
since @xmath1282 is a bijection of @xmath1267 onto @xmath1338 , setting @xmath1339 proves part(i ) .
\(ii ) follows from ( [ a2pt.15 ] ) and from the fact that for @xmath1292 , @xmath1309 , @xmath1340 this implies that @xmath1341 , and similarly @xmath1342 .
\(iii ) in addition to the preceding hypotheses we now assume that @xmath1236 , and therefore @xmath1326 , are invariant under the real lorentz group @xmath1243 . applying the bargmann - hall - wightman theorem @xcite , we find that there exists a function @xmath1343 , holomorphic in @xmath1344 such that @xmath1345 for all @xmath1346 .
setting @xmath1347 we obtain part ( iii ) .
0.5 truecm [ rema1]another proof of part ( i ) can be given by using the temperedness of @xmath1236 .
indeed there is an integer @xmath1348 such that the functions @xmath1349 are holomorphic in @xmath1350 ( respectively ) and coincide on @xmath1330 . by the `` double - cone theorem '' @xcite
, their region of coincidence extends to @xmath1351 and the standard glaser - streater theorem can be applied
. 1 truecm we return to the functions @xmath1224 , which we now assume invariant under @xmath22 , and apply lemma [ gs ] to @xmath1237 : these functions have a common holomorphic extension to @xmath1267 , and there exists a function @xmath1277 , holomorphic on @xmath378 , such that @xmath1352 . for @xmath1353 and @xmath1354 , let @xmath1355 . for real @xmath1356 and @xmath1354
, we have @xmath1357 .
indeed we can write @xmath1358 , @xmath1359 for some @xmath241 and @xmath1360 , so that @xmath1361 . therefore @xmath1362 coincides with @xmath1363 i.e. with @xmath1364 for all @xmath1353 and @xmath1354 .
similarly @xmath1365 coincides with @xmath1364 on @xmath1366 .
we have proved : [ ads2pt ] with the preceding assumptions on @xmath1224 , and assuming in addition that these functions are invariant under @xmath22 , there exists a function @xmath1277 , holomorphic on @xmath378 , such that @xmath1224 coincide with @xmath1367 in their domains of definition .
1 truecm [ rema2]this implies that @xmath1368 coincide not only on @xmath1227 , but also on the `` exotic region '' @xmath1369 : @xmath1370 our proof shows that this also holds without assuming that @xmath1224 are invariant under @xmath22 .
the topology of @xmath1238 ( and hence of @xmath1371 ) is made clear by the holomorphic bijection @xmath1282 , which maps @xmath1238 onto @xmath1327 in the complex minkowski space @xmath1306 . to make things even clearer
, one can use the map @xmath1372 defined in @xmath1306 by @xmath1373 this is a holomorphic bijection of @xmath1374 onto itself , which maps @xmath1305 onto itself ( i.e. in particular @xmath1375 for @xmath1376 ) .
it maps @xmath1377 onto @xmath1378 and therefore it is a holomorphic bijection of @xmath1379 onto @xmath1380 ( and similarly for @xmath1381 )
. we shall however continue to work with @xmath1327 which has the advantage of being invariant under @xmath1243 .
we denote @xmath1382 this is an analytic hypersurface containing @xmath1334 .
[ connec ] the open set @xmath1383 is connected and simply connected .
_ the set @xmath1384 is star - shaped with respect to 0 , i.e. @xmath1385 for every @xmath1386 ( but @xmath1387 ) .
for every @xmath1388 , @xmath1389 . hence if @xmath1390 denotes the open ball @xmath1391 , we have @xmath1392 = @xmath1393 and this intersection is convex .
we can define a map @xmath1394 of @xmath1395\times a$ ] into @xmath18 such that @xmath1396 and @xmath1397 sends @xmath18 into @xmath1392 .
hence any two points in @xmath18 can be connected by a continuous arc , and every closed curve in @xmath18 is homotopic to 0 .
we now suppose given a pair of functions @xmath1326 respectively holomorphic and of tempered growth in the covering of @xmath1327 .
this is equivalent to giving a pair of sequences @xmath1398 , with the following properties:(1 ) for each @xmath1399 , @xmath1400 is holomorphic in @xmath1401;(2 ) every @xmath1402 has an open complex neighborhood @xmath1403 such that @xmath1404 and @xmath1405 have a common holomorphic extension in @xmath1403;(3 ) similarly for @xmath1406 in @xmath1407.in addition we suppose that(4 ) the boundary values of @xmath1408 coincide in @xmath1330 .
the proof of the glaser - streater theorem again applies to show that @xmath1408 have a common holomorphic extension @xmath1409 in @xmath1410 , in particular that , the map @xmath1411 of @xmath1412 into @xmath752 extends to a holomorphic function on @xmath1413 . from this
it follows that @xmath1414 also extends to a holomorphic function on @xmath1413 .
the bargmann - hall - wightman lemma again shows that each of the functions @xmath1400 extends to a function holomorphic in @xmath1410 .
moreover these two functions coincide , and we denote @xmath1415 .
this can be seen , for @xmath724 , by induction on @xmath0 . supposing it to hold up to @xmath1416 , we consider points of the form @xmath1417 for @xmath1418 and @xmath1419 , which belong to @xmath1420 . at such a point @xmath1421 note
however that if in addition the `` hermiticity condition '' @xmath1422 holds , this extends to @xmath1423 for all @xmath1399 .
there is no general reason for @xmath1424 and @xmath1425 to coincide for @xmath1426 .
if we suppose that @xmath1408 are invariant under @xmath1243 , then for each @xmath1399 the functions @xmath1029 are locally lorentz invariant and , again by the bargmann - hall - wightman lemma , there is a function @xmath1427 , holomorphic in @xmath1428 , such that @xmath1429 .
moreover @xmath1430 for all @xmath1431 and all @xmath1399 .
therefore there exists a function @xmath1432 , holomorphic on @xmath1433 such that , for each @xmath1399 , @xmath1434 in @xmath1435 let now @xmath1224 be functions holomorphic and of tempered growth on @xmath1436 and @xmath1437 , respectively . we suppose that @xmath1438 for all @xmath1439 and all @xmath344 in the relevant domain , and that the boundary values of @xmath1224 on @xmath212 coincide at space - like separated arguments . as in the case of @xmath26
, we assume that the boundary values can be extended to @xmath63 functions of a real first argument , holomorphic and of tempered growth in the second argument in @xmath1440 ( respectively ) and define @xmath1237 , @xmath1441 .
then @xmath1326 satisfy the conditions ( 1)-(4 ) mentioned above , @xmath1400 are @xmath1243 invariant , and the preceding remarks provide the functions @xmath1029 , @xmath1427 and @xmath1432 .
we can transport back these properties to the functions @xmath1224 .
[ rema3]it is worth noting the following formula , where @xmath1442 and @xmath1443 , @xmath1444 : @xmath1445
we start with _ proof . _
let @xmath1448 be sufficiently small .
we denote @xmath1449 for all @xmath1450 . for @xmath1451 , we define @xmath1452 as follows ( @xmath596 and @xmath1453 are integers in @xmath1454 $ ] ) : @xmath1455 this gives @xmath1456 we adjust @xmath1457 and @xmath1458 so that only the first term survives in the rhs of this identity , i.e
@xmath1459 with these values , @xmath1460 and @xmath1461 can be recovered from @xmath1462 and @xmath1463 , respectively , then @xmath1464 can be recovered from @xmath1465 .
thus the @xmath1466 form a basis of @xmath1467 .
we obtain a basis @xmath1468 of @xmath198 by setting @xmath1469 .
clearly , given a neigborhood @xmath438 of @xmath1470 , @xmath1471 can be chosen so small that @xmath1472 for all @xmath1473 and @xmath981 . _
_ we first show that , if @xmath359 has sufficiently small modulus , @xmath1476 is an interior point of @xmath501 . denoting @xmath1477 , let @xmath1478h_{d-1,d})$ ] , be the basis of @xmath198 constructed in the proof of lemma [ basis ] .
in particular @xmath1479 .
we consider the two holomorphic maps of @xmath1480 into @xmath23 @xmath1481 these maps are tangent at 0 , and , for sufficiently small @xmath917 , both are biholomorphic maps of the polycylinder @xmath1482 into @xmath23 . in particular the subset @xmath1483 of @xmath23 is open and contained in @xmath501 . for sufficiently small @xmath1484 the curve @xmath1485 exists and is tangent to @xmath1486 .
therefore there exists an @xmath1487 such that @xmath1488 and @xmath1489 imply @xmath1490 , i.e. @xmath1491 .
we now consider a point @xmath1492 of the form @xmath1493 , where @xmath1494 and @xmath1495 .
this point can be rewritten as @xmath1496 where @xmath105 is arbitrary in @xmath1497 and @xmath1498 . for sufficiently small @xmath105 , @xmath1499 hence @xmath1500 , an open set contained in @xmath501 . since @xmath501 is invariant under conjugations from @xmath22 , this proves the corollary .
_ let @xmath1492 and @xmath682 .
for any real @xmath572 and @xmath51 , @xmath1502 since @xmath501 is open , for sufficiently small @xmath363 , @xmath1503 and @xmath1504 hence @xmath1505 .
any @xmath1506 can be written as a finite product @xmath1507\exp(s_n m_n)$ ] where @xmath1508 and @xmath505 , hence @xmath1509 for any @xmath1492 .
thus @xmath1510 , and since @xmath1511 for all @xmath1506 , this implies @xmath1512 .
bertola , m. , bros , j. , moschella , u. , schaeffer , r. : a general construction of conformal field theories from scalar anti - de sitter quantum field theories .
. phys .
b * 587 * , 619644 ( 2000 ) [ hep - th/9908140 ] birrell , n.d . , davies , p.c.w . : _ quantum fields in curved space_. cambridge : cambridge university press , 1982 bros , j. , viano , g.a .
: connection between the harmonic analysis on the sphere and the harmonic analysis on the one - sheeted hyperboloid : an analytic continuation viewpoint .
forum mathematicum * 8 * , 621 - 658 ; 659 - 722 ( 1996 ) and * 9 * , 165 - 191 ( 1997 ) randall , l. and sundrum , s. : a large mass hierarchy from a small extra dimension . phys . rev .
lett . * 83 * , 33703373(1999 ) [ hep - ph/9905221 ] .
an alternative to compactification .
rev.lett .
* 83 * , 46904693 ( 1999 ) [ hep - th/9906064 ] . | we propose a general framework for studying quantum field theory on the anti - de - sitter space - time , based on the assumption of positivity of the spectrum of the possible energy operators . in this framework
we show that the @xmath0-point functions are analytic in suitable domains of the complex ads manifold , that it is possible to wick rotate to the euclidean manifold and come back , and that it is meaningful to restrict ads quantum fields to poincar branes .
we give also a complete characterization of two - point functions which are the simplest example of our theory .
finally we prove the existence of the ads - unruh effect for uniformly accelerated observers on trajectories crossing the boundary of ads at infinity , while that effect does not exist for all the other uniformly accelerated trajectories . |
if we could map , at fixed time , the instantaneous locations in three - dimensional space of all kuiper belt objects ( kbos ) , on what two - dimensional surface would the density of kbos be greatest ?
we call this surface the plane of the kuiper belt ( kbp ) , though by `` plane '' we do not mean to imply that the kbp is flat ( we shall find that it is not ) .
the kbp depends on the mass distribution of the solar system principally , the orbits and masses of the giant planets .
there are as many different kbps as there are dynamical classes of kbo , since each class of object feels a distinct time - averaged force . here
we study the kbp defined by classical kbos : objects whose fairly circular , low inclination orbits are not in any strong mean - motion resonance with neptune ( see elliot et al .
2005 for a classification scheme ) . in principle
, theoretical determination of the kbp would help observers to discover new kbos .
conversely , by measuring differences between the theoretical kbp and the actual kbp , we might hope to infer the presence of solar system bodies as yet undetected ( `` planet x '' ; see gaudi & bloom 2005 for a summary of current limits ) .
there is disagreement regarding the location of the kbp .
brown & pan ( 2004 ) analyzed the instantaneous proper motion vectors of hundreds of kbos irrespective of dynamical class and concluded , with greater than @xmath0 confidence , that the kbp did not coincide with the invariable plane ( ip , the plane perpendicular to the total angular momentum vector of the solar system ) .
they argued that the observed kbp was consistent instead with the forced plane given by linear secular perturbation theory .
we will refer to this plane as the bvwp , after brouwer & van woerkom ( 1950 ) , who developed a linear secular theory for the motions of all eight of the major planets .
their theory , in turn , has its origin in the laplace - lagrange equations ( see , e.g. , murray & dermott 1999 ) .
the bvwp is a warped and time - variable surface whose properties we review in [ review ] .
by contrast , elliot et al . ( 2005 ) found that the plane determined by classical kbos that were observed over multiple epochs was more consistent with the ip ( @xmath1 difference ) than with the bvwp ( @xmath23@xmath3 difference ) .
they listed some arguments , none conclusive , for why the ip might be preferred over the bvwp .
the low order of the bvw theory , and its inability to account for time variations in semi - major axes , are causes for concern .
we seek to resolve this disagreement using numerical orbit integrations . in
[ review ] , we review the linear theory and how it equates the kbp with the bvwp . in [ answer ] , numerical integrations lasting the age of the solar system are used to reveal the theoretical location of the kbp . in [ obs ] , we compare theory against current observations of the kbp . a summary is given in
[ conc ] , including a brief comment on the prospects for detecting an unseen , outer solar system planet using the kbp .
we study the secular evolution of classical kbos by solving the laplace - lagrange equations of motion , which neglect all terms higher than second order in orbital eccentricity ( @xmath4 ) and orbital inclination ( @xmath5 ) .
the solution is detailed in the textbook by murray & dermott ( 1999 ) ; we provide a summary here .
we start by describing the motions of the planets .
define an inclination vector @xmath6 , where @xmath5 and @xmath7 equal the inclination and longitude of ascending node .
lagrange s equations governing the inclination vector for the @xmath8th planet are @xmath9 @xmath10 where @xmath11 , @xmath12 , and @xmath13 are , respectively , the mean motion , semi - major axis , and disturbing function .
we consider only the four giant planets so that @xmath14 represents jupiter , saturn , uranus , and neptune , respectively .
the disturbing function , keeping only the leading terms relevant to the inclination evolution of the @xmath8th planet , reads @xmath15 , \ ] ] where @xmath16 is a laplace coefficient , and allow the laplace coefficient to take @xmath17 as an argument , then separating the case @xmath18 from @xmath19 is not necessary .
we stick here with the textbook convention , however . ]
@xmath20 is the mass of the @xmath8th planet , @xmath21 and @xmath22 eqns .
( 1)(2 ) yield two coupled systems of first - order differential equations : @xmath23 and @xmath24 , where @xmath25 is a @xmath26 matrix of constant coefficients depending on the masses and semi - major axes of the planets .
these equations describe coupled harmonic oscillators ; their solution is @xmath27 @xmath28 where the frequencies @xmath29 are the eigenvalues of @xmath25 .
the elements @xmath30 of the eigenvectors of @xmath25 , and the phases @xmath31 , are fitted to the initial inclinations and ascending nodes of the planets .
we take initial conditions and planetary data from the nasa jpl horizons database for jd @xmath32 ( jan 1 2000 ; @xmath33 in equations ( 3 ) and ( 4 ) ) .
table [ table_first ] lists the resultant values for @xmath29 , @xmath30 , and @xmath31 .
we refer to this solution as the bvw solution , even though brouwer and van woerkom ( 1950 ) included all eight planets .
rrrrr @xmath34 & 1.5792 & 0.36081 & 0.055195 & -0.066653 + 2 & 1.5792 & -0.89904 & 0.044853 & -0.064203 + 3 & 1.5792 & 0.040670 & -1.01200 & 0.062630 + 4 & 1.5792 & 0.0045122 & 0.11941 & 0.67290 + @xmath35 & 0 & -123.13 & -14.102 & -3.2984 + @xmath36 & 1.8788 & -0.91730 & 2.3699 & -2.7418 [ table_first ] we now turn to our main concern , the motion of a kbo of negligible mass with semi - major axis @xmath37 .
its disturbing function is @xmath38 \,,\ ] ] where unsubscripted variables refer to the kbo .
because all masses and semi - major axes are fixed for this secular problem , lagrange s equations of motion are of the form @xmath39 @xmath40 where the @xmath41 s are constants . substituting each equation into the time derivative of the other ,
we find @xmath42 \label{ddotq}\ ] ] @xmath43 \ , .
\label{ddotp}\ ] ] eqns .
( [ ddotq])([ddotp ] ) describe harmonic oscillators of natural frequency @xmath44 , forced by the planetary terms in the sums .
the motion is composed of a forced oscillation and a free oscillation : @xmath45 @xmath46 where @xmath47 and @xmath48 are constants determined by initial conditions , and the free precession frequency @xmath49 the functions @xmath50 and @xmath51 depend only on planetary parameters and the kbo semi - major axis @xmath52 : @xmath53 @xmath54 where @xmath55 the inclination vector of a kbo is the vector sum of a forced inclination vector @xmath56 and a free inclination vector @xmath57 . throughout this paper , we refer to a forced inclination @xmath58 , a forced node @xmath59 , a free inclination @xmath60 , and a free node @xmath61 thus free and forced inclinations refer to magnitudes ( not vectors ) .
the free inclination is constant ; it is the undamped amplitude of the free oscillation .
figure [ fig_diameter ] shows the evolution of @xmath62 at various semi - major axes in the kuiper belt .
the open diamonds , located at the endpoints of @xmath62 , mark what we call the forced poles . at fixed time ,
the forced poles all lie along a line intersecting the invariable pole , denoted * i*. to the extent that the @xmath63 mode , driven mainly by neptune , dominates , the forced poles rotate clockwise ( regress ) about the invariable pole at a single frequency ( @xmath64 ) , maintaining constant distance to * i*. note that the forced poles at @xmath65au lie diametrically opposite to those at @xmath66au , and that as @xmath52 approaches 40.5au from either above or below , the separation between the forced pole and * i * increases .
these latter two properties reflect the @xmath67 secular resonance at @xmath68au , where the denominators @xmath69 of equations ( [ eq_qforced])([eq_pforced ] ) vanish : the forced response becomes infinite in magnitude at resonance , and changes phase by @xmath70 across resonance ( the sign of @xmath69 switches across @xmath71au ) .
figure [ fig_phasemix ] illustrates how free nodes @xmath72 phase mix and how such phase mixing helps determine the kbp within the bvw theory . at @xmath33 , one hundred kbos having semi - major axes within 0.5 au of 43 au
are set down with orbit normals approximately aligned about an arbitrary direction ( the orbit poles are actually distributed in a small box in @xmath73-@xmath74 space ) . over tens of myrs ,
the orbit poles of the particles drift away from one another : the small dispersion in semi - major axis produces a small dispersion in the free precession frequency @xmath75 ( the rate at which the free inclination vector rotates about the forced pole ) . and @xmath29 do not depend on eccentricity or inclination , but in higher order theories they do . ]
while the free nodes distribute themselves over all phases , the free inclinations @xmath47 remain fixed .
thus , at late times , the collection of free inclination vectors @xmath76 are distributed axisymmetrically about the mean forced pole at @xmath77 au .
the axisymmetry arises practically independently of how the particles orbit normals are initially distributed ; the only requirements are that many particles share the same @xmath47 ( so that such particles , when phase mixed , trace a full circle ) and that there exists a small but non - zero dispersion in semi - major axis ( so that there exists a non - zero spread in free precession frequencies , enabling phase mixing ) .
note in figure [ fig_phasemix ] how the invariable pole does not coincide with the center of the circular distribution .
we conclude that according to the bvw theory , the mean orbit normal of a phase - mixed group of particles having approximately the same semi - major axis is given by the forced pole corresponding to that semi - major axis . that local forced pole varies with time ( figure [ fig_diameter ] ) , but the particles always encircle it ( figure [ fig_phasemix ] ) .
the kbp is warped because the forced pole changes direction with kbo semi - major axis ( figure [ fig_diameter ] ) .
note finally that the ability of test particles to keep their free inclinations constant relative to a time - variable forced pole should not be confused with adiabatic invariance .
generally the frequencies @xmath75 and @xmath29 are not cleanly separated .
the constant @xmath47 simply reflects the undamped amplitude of free oscillation , and is set by initial conditions . in the next section
, we use numerical simulations to test the constancy of free inclination .
we calculate the evolution of the kbp at three semi - major axes : @xmath78 , 43 , and 44 au . these are chosen to lie away from strong mean - motion resonances ( e.g. , the 3:2 resonance resides at 39.5 au ) and outside the @xmath7942 au region of instability carved by the @xmath67 , @xmath80 and @xmath81 secular resonances ( see , e.g. , chiang et al .
2007 ) . at each @xmath52
we lay down @xmath82 test particles whose initial free inclination vectors are distributed axisymmetrically about the local forced pole .
that is , each particle s initial @xmath83 and initial @xmath84 , where @xmath85 takes 1 of @xmath86 values ( 0.01 , 0.03 , 0.1 , 0.3 rad ) and @xmath87 takes 1 of @xmath88 values distributed uniformly between 0 and @xmath89 .
this set - up permits us to directly test the bvw theory , which predicts that all @xmath90 particles corresponding to a given @xmath52 and given initial @xmath47 should keep the same @xmath47 for all time : a circle of points in @xmath73-@xmath74 space should continue to trace the same - sized circle . coordinates @xmath51 and @xmath50 for the initial forced poles are computed using the bvw solution of [ review ] , for @xmath91 .
initial osculating eccentricities are zero and initial mean anomalies are chosen randomly between 0 and @xmath89 .
the four giant planets are included in the integration , with initial conditions taken from the jpl horizons database for jd = 2451544.5 ( @xmath33 in the bvw theory ) .
the integration is performed with the swift_rmvs3 code , written by levison & duncan ( 1994 ) and based on the algorithm developed by wisdom & holman ( 1991 ) .
the duration of the integration is 4 gyr and the timestep is 400 days ( about 1/11 the orbital period of jupiter ) .
we work in a heliocentric coordinate system , the better to compare with the linear secular theory which uses heliocentric elements .
according to the bvw theory , each set of @xmath88 particles having the same initial @xmath52 and initial @xmath47 should trace the perimeter of a single circle in @xmath73-@xmath74 space at any given time , with the center of the circle yielding the local normal to the kbp .
figure [ 43_0.10 ] tests this prediction ; the panels display the @xmath73-@xmath74 positions of particles initially having @xmath92 au and @xmath93 rad , sampled at four different times .
most particles at a given time do lie approximately on one circle , although there are outliers ( see crosses in panels b , c , and d ) .
the outliers represent particles whose inclinations and eccentricities grow to large values .
many of these particles suffer close encounters with neptune , whereupon they are removed from the integration .
we did not identify the cause of the instability , but probably the various high - order mean - motion resonances in the vicinity ( nesvorn & roig 2001 ) are to blame . of the 20 particles shown at @xmath33 in panel
a , only 12 survive to @xmath94 gyr in panel d. those that survive have semi - major axes that remain constant to within @xmath950.5 au . at each @xmath96
, we fit a circle ( in a least squares sense ) to those particles known to survive the entire length of the integration .
the fit is substantially improved by also discarding particles whose instantaneous eccentricities exceed some value @xmath97 , as we find that particles on eccentric orbits tend also to be outliers in @xmath73-@xmath74 space .
the value of @xmath97 is reduced from one until the fit parameters cease to change significantly .
note that a particle that is discarded from the fit by the @xmath97-criterion at one time can be restored to the fit at a later time ( if its eccentricity falls below @xmath97 at that later time ) .
table [ stability ] lists the values of @xmath97 chosen for the various combinations of initial @xmath52 and initial @xmath47 .
the circles so fitted are overlaid in figure [ 43_0.10 ] .
a typical fit is excellent and the local normal of the kbp ( center of the fitted circle ; solid diamond ) is confidently identified .
figures [ 43_0.30 ] , [ 38_0.10 ] , and [ 38_0.30 ] report analogous results for other choices of initial @xmath52 and initial @xmath47 .
all 20 particles having @xmath92 au and @xmath98 rad are stable for 4 gyr ( figure [ 43_0.30 ] ) .
by contrast , at @xmath99 au , where various high - order mean - motion resonances are known to cause instability ( nesvorn & roig 2001 ) , only 2 out of 20 particles having @xmath98 rad remain at the end of 4 gyr ( figure [ 38_0.30 ] ) . fitting a unique circle to 2 points
is impossible . and
as is clear from figure [ 38_0.30 ] , even if we were to try fitting circles at earlier times when more particles are present , such fits would be poor .
only certain combinations of @xmath52-@xmath47 permit determination of the kbp , as documented in our table [ stability ] .
panels d of figures [ fig_fourplot_a ] and [ fig_fourplot_b ] plot the radii of the fitted circles versus time , for two choices of initial @xmath52-@xmath47 .
the bvw linear theory predicts that these radii should be constant .
in fact they are nearly so , varying by at most one part out of six .
how can we further test the bvw theory when we know that it fails to predict the orbits of the planets on gyr timescales ?
we compute instead a semi - analytic , bvw - based solution as follows . at each time in the numerical integration , we output the inclinations and ascending nodes of the giant planets and use these to recompute the eigenvectors and thus the forced poles of the linear theory . thus we obtain a prediction of where the kbp should reside according to the linear theory at a given instant , using the simulation results for the planets at that instant to supply the integration constants .
this `` continuously updated bvw solution '' for the local normal ( forced inclination vector ) is shown as an open diamond in figures [ 43_0.10][38_0.30 ] .
its location tracks that of the numerically fitted pole ( solid diamond ) well much better than does the invariable pole ( upright cross ) ; see especially figures [ 43_0.10 ] and [ 38_0.10 ] .
figures [ fig_fourplot_a ] and [ fig_fourplot_b ] also demonstrate that the continuously updated bvwp hews closely to the numerically fitted kbp .
cccccc 38 & 0.01 & 17 & 0.05 & none applied & yes + & 0.03 & 16 & 0.05 & none applied & yes + & 0.10 & 8 & 0.07 & none applied & yes + & 0.30 & 2 & 0.02 & none applied & no + 43 & 0.01 & 19 & 0.12 & 0.04 & yes + & 0.03 & 20 & 0.04 & none applied & yes + & 0.10 & 12 & 0.12 & 0.08 & yes + & 0.30 & 20 & 0.02 & none applied & yes + 44 & 0.01 & 20 & 0.11 & 0.03 & yes + & 0.03 & 20 & 0.11 & 0.04 & yes + & 0.10 & 20 & 0.15 & 0.06 & yes + & 0.30 & 20 & 0.29 & 0.08 & yes + [ stability ] to summarize the results of our numerical simulations : linear secular theories like bvw correctly predict how the warped kbp evolves with time , provided the parameters of those analytic theories are continuously updated using either observations or numerical simulations of the giant planets orbits .
the kbp today is accurately predicted by the updated bvw solution .
the kbp does not , in general , coincide with the invariable plane , except at infinite distance from the planets .
we have shown by numerical simulations in [ answer ] that the classical kbp is given by the bvwp , i.e. , by linear secular theory . here
we assess whether observations of kbos bear out this result , by locating the actual poles of the kbp near 38 au and 43 au .
these semi - major axes lie to either side of the @xmath67 resonance ; theory predicts that the corresponding poles should lie to either side of the invariable pole , with all three orbit normals lying in one plane ( see figure [ fig_diameter ] ) .
our dataset consists of kbos listed on the minor planet center website on jan 22 2008 whose ( a ) astrometric arcs extend longer than 50 days ( many objects in our sample have much longer arcs ) , ( b ) eccentricities are less than 0.1 , ( c ) inclinations are less than 10 degrees , and ( d ) are classified by the deep ecliptic survey ( des ) as `` classical '' in at least two out of their three orbital integrations ( see elliot et al .
2005 for a description of the classification scheme ) . in the vast majority of cases , all three des integrations yield a classification of `` classical .
'' moreover , the typical @xmath0 uncertainties in semi - major axes are smaller than @xmath1000.1 au .
we assemble two samples , one for which @xmath101 ( the `` 38 au '' sample , containing 10 objects ) and another for which @xmath102 ( the `` 43 au '' sample , containing 80 objects ) .
object designations are given in table [ table_desig ] .
cl `` 38 au '' & 1998 wv@xmath103 , 1999 oj@xmath104 , 2000 yb@xmath105 , 82157 , 2003 fd@xmath106 , 2003 qa@xmath107 , 2003 yl@xmath108 , + & 2003 qq@xmath109 , 144897 , 119951 + `` 43 au '' & 19255 , 1994 ev@xmath110 , 1996 tk@xmath111 , 33001 , 1998 wy@xmath103 , 1998 wx@xmath103 , 1999 cn@xmath112 , + & 1999 rt@xmath113 , 1999 on@xmath104 , 1999 xy@xmath114 , 1999 rw@xmath113 , 1999 ch@xmath115 , 1999 ru@xmath116 , + & 1999 hv@xmath117 , 1999 da , 1999 hj@xmath118 , 1999 cw@xmath119 , 2000 pu@xmath120 , 2000 px@xmath120 , + & 134860 , 2000 cl@xmath121 , 2000 on@xmath122 , 2000 pc@xmath123 , 2000 fs@xmath124 , 2000 wv@xmath118 , + & 2000 wl@xmath125 , 2000 ou@xmath126 , 2000 yu@xmath127 , 88268 , 2001 qb@xmath128 , 2001 qd@xmath128 , 2001 xr@xmath129 , + & 2001 qo@xmath130 , 2001 hz@xmath131 , 2001 rw@xmath114 , 2001 ok@xmath132 , 2001 db@xmath133 , 88267 , + & 2001 xu@xmath129 , 2001 fk@xmath134 , 2001 oz@xmath132 , 2002 cd@xmath135 , 2002 px@xmath136 , 2002 pv@xmath136 , + & 2002 fw@xmath137 , 2002 wl@xmath138 , 160256 , 2002 vb@xmath119 , 2002 py@xmath136 , 2002 cs@xmath115 , 2002 pd@xmath139 , + & 2003 sn@xmath140 , 2003 ut@xmath141 , 2003 fk@xmath142 , 2003 qg@xmath109 , 2003 fa@xmath143 , 2003 hy@xmath144 , + & 2003 qy@xmath145 , 2003 tk@xmath131 , 2003 qf@xmath109 , 2003 qe@xmath109 , 2003 qz@xmath146 , 2003 ql@xmath109 , 2003 qe@xmath147 , + & 2003 yr@xmath108 , 2003 qy@xmath146 , 2003 qd@xmath109 , 2003 tl@xmath131 , 2003 qu@xmath145 , 2003 yt@xmath108 , + & 2003 yx@xmath108 , 2003 ys@xmath108 , 2003 yj@xmath108 , 2004 ud@xmath148 , 2004 dm@xmath149 , 2005 jz@xmath150 , + & 2005 gd@xmath151 , 2005 jp@xmath108 , 2005 xu@xmath152 , 2006 ha@xmath153 + [ table_desig ] figure [ fig_real ] plots the observed ( @xmath74,@xmath73 ) positions for the two samples , and compares with theory .
theory predicts that the average @xmath154 measured for each sample should equal @xmath155 calculated for the average semi - major axis of the sample .
the good news is that the data are consistent with this prediction ; the differences between the observed poles and the forced poles are less than 2@xmath3 for both samples ( the error bars in figure [ fig_real ] are @xmath156 , where @xmath3 is the standard deviation of the mean ) .
the bad news is that the observed poles are also consistent with the invariable pole , at a similar confidence level . while it is encouraging for the theory that the observed pole at 38 au indeed lies to the left ( toward smaller @xmath74 ) of the invariable pole , and that the observed pole at 43 au lies to the right , there are not enough observations to make more precise statements and to rule out the hypothesis that the kbp equals the ip with greater than @xmath0 confidence . note that by selecting our sample to have orbital inclinations less than 10 degrees with respect to the ecliptic plane , we bias our measurement of the average pole position towards the ecliptic pole
this systematic error is probably still smaller , however , than our random error .
for example , the observed pole at 38 au actually lies further from the ecliptic pole than does the theoretically expected forced pole .
see elliot et al .
( 2005 ) for ways of reducing this systematic bias .
we also performed a two - dimensional kolmogorov - smirnov test to see whether the @xmath157 distributions for the two samples differ ( they should ) .
the probability that they do not is 4.9%small enough to be suggestive of a real difference , but in our judgement too large to be conclusive .
classical kuiper belt objects trace a sheet that warps and precesses in response to the planets . the current location and shape of this sheet the `` plane '' of the classical kuiper belt can be computed using linear secular theory , with the observed masses and current orbits of the planets as input parameters ( brown & pan 2004 ) .
the local normal to the plane is given by the theory s forced inclination vector , which is specific to every semi - major axis . at infinite distance from the planets
, the plane coincides with the invariable plane .
the deviations of the kuiper belt plane away from the invariable plane , while generally non - zero , are typically small : less than 3 degrees outside the secularly unstable gap at @xmath158 au ( inside the gap no kbos have been observed , as expected ) . as the semi - major axis varies from @xmath159 au to @xmath160 au , the ascending node of the kuiper belt plane on the invariable plane rotates by very nearly 180 degrees , a result of the sign change in the forced response across the @xmath67 resonance .
these conclusions are supported by our numerical integrations of giant planet and test particle orbits lasting 4 gyr .
these integrations show that a kuiper belt object maintains a nearly fixed orbital inclination with respect to the time - variable kuiper belt plane .
this accords with linear theory , in which the free inclination represents the undamped , constant amplitude of a test particle s free oscillation . it may seem surprising that the linear theory is vindicated in this regard while it can not accurately predict planetary motions and hence the future location of the kuiper belt plane .
but the inaccuracies accrue only slowly figure [ fig_ad ] shows that it takes several precession periods for the linear theory to diverge from numerical simulation in predicting the location of the plane , and the analytic and numerical solutions are always qualitatively similar . referring back to the equations of motion ( [ ddotq])([ddotp ] ) , we see that if we take the planetary forcing terms @xmath161 and @xmath162 to be given by the more realistic numerical integration instead of the inaccurate but qualitatively similar analytic solution ( [ q_j])([p_j])then the free component of the motion ( the homogeneous solution of the differential equation ) would still be given by ( [ qfree])([pfree ] ) , irrespective of the forced component ( the inhomogenous solution ) .
thus an object can keep its free inclination with respect to the forced plane fixed , even though the location of the forced plane itself can not be forecast analytically .
currently the data on actual kbos are consistent with , but do not conclusively verify , our theoretical finding that the kuiper belt plane warps by a few degrees to either side of the secularly unstable gap .
our analysis of the observations , like that of elliot et al .
( 2005 ) , suffers from large random errors .
the study by brown & pan ( 2004 ) does not , but at the expense of including objects of all dynamical classes and not following variations in the pole position with semi - major axis .
quadrupling the sample size of low-@xmath5 , low-@xmath4 objects at 38 au , where there are currently ten usable objects , can increase our confidence in the reality of the warp to greater than 3@xmath3 .
detecting `` planet x '' via its influence on the kuiper belt plane will be substantially more challenging .
we calculate that a 100-@xmath163 planet with a semi - major axis of 300 au and an orbital inclination with respect to the ecliptic of 10 degrees would shift the forced poles in the 4350 au region by 0.1 degree .
we thank ruth murray - clay and mike brown for discussions , and chris culter for suggesting this undergraduate research collaboration .
an anonymous referee provided numerous helpful comments .
this work was supported by nsf grant ast-0507805 . | by numerically integrating the orbits of the giant planets and of test particles over a period of four billion years , we follow the evolution of the location of the midplane of the kuiper belt .
the classical kuiper belt conforms to a warped sheet that precesses with a 1.9 myr period .
the present - day location of the kuiper belt plane can be computed using linear secular perturbation theory : the local normal to the plane is given by the theory s forced inclination vector , which is specific to every semi - major axis .
the kuiper belt plane does not coincide with the invariable plane , but deviates from it by up to a few degrees in stable zones .
for example , at a semimajor axis of 38 au , the local kuiper belt plane has an inclination of 1.9 degrees and a longitude of ascending node of 149.9 degrees when referred to the mean ecliptic and equinox of j2000 . at a semimajor axis of 43
au , the local plane has an inclination of 1.9 degrees and a nodal longitude of 78.3 degrees . only at infinite semimajor axis
does the kuiper belt plane merge with the invariable plane , whose inclination is 1.6 degrees and nodal longitude is 107.7 degrees .
a classical kuiper belt object keeps its inclination relative to the kuiper belt plane nearly constant , even while the latter plane departs from the trajectory predicted by linear theory .
the constancy of relative inclination reflects the undamped amplitude of free oscillation ; that is , the homogeneous solution to the forced harmonic oscillator equation retains constant amplitude , even while the inhomogeneous solution can not be written down accurately because the planetary forcing terms are chaotic .
current observations of classical kuiper belt objects are consistent with the plane being warped by the giant planets alone , but the sample size will need to increase by a few times before confirmation exceeds @xmath0 in confidence . in principle , differences between the theoretically expected plane and the observed plane could be used to infer as yet unseen masses orbiting the sun , but carrying out such a program would be challenging . |
a good knowledge of the pion - pion scattering is important in studies of different processes in nuclear and particle physics @xcite .
construction of phenomenological @xmath0 amplitudes requires not only experimental input but also use of theoretical constraints such as unitarity , analyticity and crossing symmetry .
direct study of the @xmath0 collisions is beyond present experimental possibilities .
phenomenological phase shifts are obtained through partial wave analyses of final states in which pions are produced .
these analyses are often model dependent and can sometimes lead to ambiguous results . in 1997
a study of the @xmath1 reaction on a polarized target has been performed in the @xmath2 effective mass between 600 and 1600 mev giving four solutions for the @xmath0 scalar - isoscalar phase shifts below 1 gev @xcite . using the unitarity constraint two steep " solutions were rejected and the two remaining ones , called down - flat " and up - flat " , passed this test @xcite .
elimination of this remaining ambiguity is necessary for extension of studies done near the @xmath0 threshold and above using chiral perturbation models based on qcd ( see for example @xcite ) . a steady increase of the @xmath0 scalar - isoscalar phase shifts below 1 gev can be interpreted as due to the presence of a broad @xmath3-meson @xcite .
previous elimination of steep " solutions excludes a narrow @xmath3 with a width comparable to that of the @xmath4-meson @xcite . in this channel , near the @xmath5 threshold and above it , other resonances exist which could be mixed with glueball states predicted by qcd . in order to eliminate the above mentioned up - down " ambiguity
one can check if the corresponding amplitudes satisfy crossing symmetry .
roy s equations @xcite can serve as a tool to perform this check .
they also correlate different experimental phase shifts determined near the @xmath0 threshold and at higher energy , in particular for the scalar - isoscalar , @xmath6 , the scalar - isotensor , @xmath7 , and the vector - isovector , @xmath8 , @xmath0 partial waves .
these equations have been already applied by pennington and protopopescu @xcite to study different @xmath0 phase shift solutions obtained in the early seventies .
they were able to resolve a
steep - flat " type ambiguity @xcite present in the phenomenological @xmath0 amplitudes considered at that time .
later two high - statistics experiments were performed @xcite and , as previously said , the analyses of these experiments lead to two plausible solutions @xcite . recently
a comprehensive analysis of roy s equations for the @xmath0 interaction has appeared @xcite .
there , a special emphasis is put on the @xmath2 range from the @xmath0 threshold to 0.8 gev . in the present paper
we pay a particular attention to the range between 0.8 and 1 gev .
in this range the largest differences between the up - flat " and down - flat " solutions occur , reaching values up to 45@xmath9 ( see fig . 4 in @xcite and fig . 2 here )
assuming analyticity one can write twice subtracted fixed - t dispersion relations @xcite for the unitary @xmath0 amplitudes , @xmath10 \\ & + & \displaystyle { { 1}\over { \pi } } \displaystyle \int\limits_{4\mu^2}^{\infty } \displaystyle \frac{ds'}{s^{\prime 2 } } \left ( \displaystyle \frac{s^2}{s'-s}\ { 1}^{ii^\prime } + \displaystyle \frac{u^2}{s'-u}c_{su}^{ii^\prime } \right ) \mbox{im } t^{i^\prime}(s',t , u ' ) .
\end{array } \label{eq:1}\ ] ] in this equation @xmath11 and @xmath12 are the usual mandelstam variables satisfying @xmath13 , where @xmath14 is the pion mass ; @xmath15 and @xmath16 are the isospin matrix elements and @xmath17 is the unit matrix .
the subtraction functions @xmath18 and @xmath19 can be related to the isospin 0 and 2 @xmath20-wave scattering lengths @xmath21 and @xmath22 , respectively .
projection of ( [ eq:1 ] ) onto partial waves leads to roy s equations , @xmath23 here the functions @xmath24 are the kernels and @xmath25 are the so - called driving terms .
detailed expressions of these functions can be found for instance in @xcite .
the driving terms contain the low - energy , @xmath26 , contributions from the partial waves @xmath27 and the high - energy , @xmath28 , contributions from all the partial waves .
if one assumes mandelstam analyticity , the range of validity of roy s equations ( [ eq:2 ] ) extends to @xmath29 ( 1.15 gev)@xmath30 @xcite .
the partial waves amplitudes @xmath31 are related to the @xmath32 phase shifts @xmath33 and inelasticities @xmath34 : @xmath35 each set of experimentally determined phase shifts and inelasticities can serve as input to calculate the real and imaginary parts of the partial wave amplitudes . after a suitable parameterization
, the imaginary parts can be inserted into roy s equations ( [ eq:2 ] ) from which one obtains the real parts as _ output _ to be compared with the corresponding real part _
input_. the quantitative agreement between output and input will be used to verify how well a given set of phase shifts satisfies crossing symmetry in a particular range of @xmath2 where roy s equations can be applied .
the down - flat " and up - flat " data @xcite together with the results on the @xmath5 phase shifts @xcite were analysed by us using a separable potential model of three coupled scalar - isoscalar channels ( @xmath36 and an effective @xmath37 system ) @xcite . this model yields particularly good fits to the down - flat " solution from 600 to 1600 mev .
the up - flat " data are reasonably well described only above 970 mev . from 800 to 970 mev
the model values are too low in comparison with the up - flat " data ( see fig . 1b of @xcite ) . in our calculations
we need a faithful representation of both up - flat " and down - flat " solutions and therefore below 970 mev we use the following pad approximation for the @xmath38 phase shifts : @xmath39 where @xmath40 is the pion momentum and @xmath41 are constant parameters .
this choice is dictated by the analytical properties of the jost functions .
these constants will be obtained from the best fits to data and to roy s equations for the up - flat " and down - flat " separately . in these fits
we also use the near threshold phase - shift differences , @xmath42 , recently extracted from the high statistics @xmath43 decay experiment @xcite . using the power expansion of the @xmath20-wave amplitudes near the @xmath0 threshold ,
@xmath44 we can relate the parameters @xmath45 and @xmath46 to the constants @xmath47 .
one has : @xmath48 and @xmath49 .
the parameters @xmath50 and @xmath51 are chosen to match the values of @xmath38 following from the three - channel fits at 969 and 970 mev @xcite . above 970 mev
we use the fit a to describe the down - flat " solution and the fit c for the up - flat " one .
these model amplitudes are used up to the @xmath52 value even if it exceeds 1600 mev ( see ( [ eq:2 ] ) ) .
the isotensor wave is parameterized within the separable potential model of @xcite .
we use a rank - two potential : @xmath53 where @xmath54 are form factors with range parameters @xmath55 , @xmath56 and @xmath57 are the pion center of mass momenta in the initial and final states , respectively .
the threshold parameters @xmath58 and @xmath59 can be related to the strength parameters @xmath60 .
the @xmath61 @xmath32 phase shifts from @xcite , obtained with their method a , serve as input in our fitting procedure . here
we assume the isotensor wave to be elastic ( @xmath62 ) from the @xmath0 threshold to @xmath52 . for the @xmath63-wave , from threshold to 970 mev
, we use an extended schenk parameterization as defined in @xcite @xmath64 the parameter @xmath65 is equal to the @xmath63-wave scattering length @xmath66 and the parameter b is related to the slope parameter @xmath67 .
the parameter @xmath68 is equal to the @xmath4-mass squared . above 970 mev
the @xmath63-wave amplitude is represented by the k - matrix parameterization of hyams _ et al . _
the parameters @xmath69 and @xmath70 of ( [ eq : schenk ] ) are chosen to match both parameterizations at 969 and .
we have checked that the near threshold scalar - isoscalar phase shifts determined from the differences @xmath71 in @xcite are insensitive to different @xmath63-wave parameterizations .
differences between parameterizations of @xcite and @xcite are smaller than the errors of @xmath71 .
the driving terms @xmath72 in ( [ eq:2 ] ) are calculated including the contributions of @xmath73 and @xmath74 and the regge contributions from the pomeron , @xmath4- and .
we use the breit - wigner parameterization of @xmath73 and @xmath74 as described in @xcite with masses , widths and @xmath0 branching ratios taken from @xcite .
the range parameters are chosen to be 5.3 gev@xmath75 and 6.4 gev@xmath75 for the @xmath76 and @xmath77 resonances , respectively ( see table 4 of @xcite ) .
the regge contributions are parameterized as in @xcite without inclusion of the small @xmath12-crossed terms .
the @xmath52 limit is set to ( 2 gev)@xmath30 and our results are fairly close to those of @xcite .
the most important contribution in the @xmath78 channel comes from the @xmath73 resonance .
it roughly agrees with the result of basdevant , froggatt and petersen @xcite where only the @xmath76 contribution was considered and @xmath52 was set equal to ( 1.46 gev)@xmath30 .
the @xmath79 results are much smaller than those of @xcite due to the difference in @xmath52 and to the lack of the @xmath74 contribution in @xcite . in the @xmath63-wave
there is a strong cancellation between the contributions of @xmath73 and @xmath74 leading to very small values , again much smaller than in @xcite .
we have considered the different parameterizations of @xmath76 and @xmath77 resonances used by @xcite .
the changes from the breit - wigner form , that we used , do not affect the phase shifts in all three partial waves by more than one degree below 970 mev . in the @xmath6 case the regge contributions are of the order of a few percent of the resonance contributions .
for the isospin 1 and 2 they are of the same order as the resonance contributions but the corresponding overall driving terms are by an order of magnitude smaller than the isospin 0 term .
alternative regge contributions were considered following papers @xcite , @xcite and @xcite .
these do not change significantly the driving terms , the changes being much smaller than one degree in the effective mass range up to 1 gev .
our thirteen free parameters , six for the isoscalar @xmath20-wave , four for the isotensor one and three for the @xmath63-wave are determined through a least square fit to the data together with the minimization of the squares of the differences between the input ( _ `` in '' _ ) and output ( _ `` out '' _ ) of roy s equations for the three waves .
we define @xmath80 , \label{eq : chi2total}\ ] ] where @xmath81 ^ 2 .
\label{eq : chi2exp}\ ] ] in ( [ eq : chi2exp ] ) @xmath82 and @xmath83 represent the experimental phase shifts and their errors , respectively .
the @xmath84 of the fit to roy s equations is defined as @xmath85 ^ 2 , \label{eq : chi2roy}\ ] ] where @xmath86\mu^2 $ ] for @xmath87 and @xmath88 .
we take a @xmath89 value of @xmath90 to obtain reasonable values of @xmath91 corresponding to an accuracy of one - half percent .
simultaneously we require that the @xmath92 should not be larger than about 18 for the fit to the 18 data points between 600 and 970 mev .
we use the cern minuit program which provides errors of the fitted parameters . for the isospin 0
wave the number of data points @xmath93 is 24 in ( [ eq : chi2exp ] ) .
it consists of the 18 down - flat " or 18 up - flat " data points @xcite between 600 and 970 mev and of the 6 points below 400 mev from the @xmath43 decay experiment @xcite . in the isotensor wave we use the 12 data points @xcite from 350 to 1450 mev . for the isospin 1 we generate 8 pseudo - data points using the @xmath94-matrix parameterization between 600 and 970 mev @xcite and we choose @xmath95 .
the pad approximants for the isoscalar @xmath20-wave amplitude supplemented by the model amplitudes @xcite for @xmath96 mev together with the @xmath63-wave parameterization and that of the isotensor @xmath20-wave as described above , constitute our input to roy s equations .
a good global fit to data and roy s equations can be achieved only for the down - flat " data .
the resulting @xmath97 phase shifts , plotted as solid lines in fig . 1 are compared to the data .
the @xmath84 values are summarized in table 1 .
for the 18 points between 600 and 960 mev the @xmath84 is 16.6 and it is 5.7 for the 6 near threshold points .
the @xmath84 per degree of freedom on the isoscalar phase shifts is 1.2 .
roy s equations are very well fulfilled , the differences between the real parts _ in " _ and _ out " _ being smaller than @xmath98 for the isoscalar wave and smaller than @xmath99 for the isotensor amplitude . for the @xmath63-wave this difference does not exceed @xmath100 .
the corresponding parameters are given in table 2 .
the low - energy parameters @xmath101 , @xmath102 , @xmath58 , @xmath59 and @xmath66 , compare well within errors with those of @xcite but @xmath103 is smaller by about @xmath104 . relative errors of those parameters vary from about @xmath105 for @xmath101 , @xmath102 and @xmath66 to @xmath106 for @xmath103 . for the other parameters , they are less than @xmath107 with the exception of @xmath108 . this parameter is negative ,
so its influence on the values of @xmath109 is small as can be seen from ( 4 ) .
it furthermore does not appear in the two first coefficients of the low energy expansion of tan @xmath109 .
one then expects the error of @xmath108 to be large . in the up - flat " case
a good global fit can not be obtained , as the @xmath92 on the 18 data points of @xcite between 600 and 960 mev is as large as 46.4 . including the @xmath84 of 6.6 from the 6 points of the @xmath43 experiment , one obtains the total value equal to 53.0 ( see table [ tab : chi2 ] ) .
the @xmath84 per degree of freedom on the isoscalar data is as large as 2.9 which shows that this fit for the up - flat " case , plotted as the solid line in fig .
1b , is not acceptable .
the agreement with roy s equations is , however , almost as good as in the down - flat " case . in fig
. 2 we plot , together with the data , the solid curves of figs .
1a and 1b representing roy
s fits to the down - flat " and up - flat " data .
these two curves define a band of isoscalar phase shifts fitting well threshold data and roy s equations for two sets of data above 600 mev .
the down"curve reproduces well the down - flat " data while the up " curve _ does not reproduce _ the up - flat " data .
.@xmath84 values for the different fits ; @xmath110 being the number of data and @xmath111 that of free parameters , @xmath112 , @xmath113 , @xmath114 , @xmath115 , @xmath116 , @xmath117 .
[ cols="^,>,>,>,>,>,>,^,^ , > , > " , ] [ allparameters ] [ padebands ] + fig . 1 : scalar - isoscalar phase shifts for * a ) * down - flat " data ( full circles ) and * b ) * up - flat " data ( open circles ) @xcite
. diamonds denote the k@xmath118 data @xcite .
solid lines represent fits to roy s equations and to data .
dashed lines represent bands of the pad fits to the data of @xcite .
the dotted line in * b ) * shows the special pad fit .
[ oneband ] + fig .
2 : solid lines fits to roy s equations and to the down - flat " data ( full circles ) and to the up - flat " data ( open circles ) of @xcite as well as to the @xmath43 data of @xcite represented by diamonds .
the dashed curve is roy s fit to the @xmath43 data and to the up - flat " data in the restricted range of @xmath2 up to 740 mev .
the main differences between the down - flat " and up - flat " solutions are for energies between 800 and 970 mev .
below 740 mev phase shifts of both solutions are compatible within their error bars and the down " curve reproduces well the up - flat " data ( see fig . 2 ) .
we have checked that there is a possibility to find a good fit to roy s equations and to the up - flat " data in this limited range of effective mass .
this is shown as the dashed line in fig . 2 .
differences of phase shifts between this line and the down " one are smaller than one degree up to 840 mev . therefore , based on the down - flat " fit , we build up a special pad fit to the up - flat " data in order to reproduce as well as possible their @xmath2 dependence .
we proceed in the following way .
below 600 mev this fit is constrained to approximate the previously obtained down - flat " isoscalar amplitude and in particular to reproduce its scattering length @xmath119 and slope parameter @xmath120 as well as its two values at 500 and 550 mev .
the corresponding @xmath121 of 19.0 is good but the @xmath122 of @xmath123 is very large .
the differences between the _ in " _ and _ out " _ real parts are as large as 0.25 around 900 mev .
so , if one tries to improve the fit to the up - flat " data then one spoils the fit to roy s equations .
[ banddown ] + fig . 3 : real parts of the @xmath0 amplitudes ( multiplied by @xmath124 ) corresponding to the down - flat " data @xcite ( full circles ) .
in * a ) * input and output bands of the scalar - isoscalar real parts are shown .
diamonds denote the @xmath43 data @xcite .
solid and dashed lines in * b ) * and * c ) * represent the output bands . in * b ) * triangles denote the isotensor data of @xcite
. crosses in * c ) * are the isovector pseudo - data calculated from the @xmath94-matrix fit of @xcite .
[ bandup ] + fig . 4 : same as in fig . 3 , but for the up - flat " data of @xcite we have studied the influence of the experimental errors on the @xmath0 input amplitudes by calculating roy s equations for two extreme isoscalar amplitudes fitted to the data points shifted upwards ( higher - in " ) or downwards ( lower - in " ) by their errors .
below 600 mev these fits were constrained in the same way as in the special pad fit to the up - flat " data just described above . the corresponding long- and short - dashed lines in figs .
1a and 1b show typical bands delimiting the possible values of the experimental data within their errors .
the resulting output curves , lower out " and higher out " , of the numerical integrations of roy s equations for the three partial - wave real parts , multiplied by @xmath124 , are displayed in figs . 3 and 4 . in figs .
3a and 4a we also show the input band for the real part of the isoscalar wave , limited by two dashed lines , called lower in " and higher in " .
these dashed lines correspond to the dashed lines shown in figs .
1a and 1b .
we do not show the _ in " _ lines in figs .
3b , 3c , 4b and 4c since they are almost indistinguishable from the _ out " _ lines .
below 600 mev roy s equations are well satisfied for all waves . at higher energies a good agreement
is found for the @xmath63-wave and @xmath61 @xmath20-wave . as seen in fig . 3 , in the down - flat " case both _ out " _
curves lie inside the band limited by the _ in " _ curves up to 937 mev . above this energy
the lower out " curve starts to lie below the higher in " and at about 950 mev the higher out " starts to be above the lower in " curve .
the _ in " _ band is then inside the _ out " _ one .
it means that there is still a possibility to find an _ out " _ solution within the _ in " _ band .
we can then conclude that within the error bars the down - flat " solution satisfies roy s equations and consequently crossing symmetry . on the contrary , in fig . 4 , for the scalar - isoscalar up - flat " solution ,
the output band lies outside the input one from 840 to 970 mev .
this eliminates the up - flat " solution as it does not satisfy crossing symmetry in that energy range .
we have studied the scalar - isoscalar @xmath0 amplitudes constructed from two sets of phenomenological phase shifts called up - flat " and down - flat " @xcite .
it has been shown that only the down - flat " solution satisfies well crossing symmetry .
the conclusion about violation of roy s equations for the up - flat " data is insensitive to the variation of the different parameters in our input : reasonable changes of the @xmath8 and @xmath7 amplitudes , shift of the upper integration limit @xmath52 from ( 2 gev)@xmath30 to ( 1.46 gev)@xmath30 , modifications of the @xmath76 and @xmath77 resonance parameters and of the regge amplitudes in the driving terms .
all these changes do not lead to complete overlap of the input and output bands for the up - flat " solution .
recently a joint analysis of the @xmath20-wave @xmath125 data @xcite and of the new @xmath126 data @xcite , obtained by the e852 collaboration at 18.3 gev / c , has been performed @xcite .
it has been shown that , using the one - pion and @xmath127-exchange model developed in @xcite , the calculated @xmath20-wave intensity of the @xmath126 system agrees with the measured @xmath126 intensity only for the @xmath0 amplitudes obtained from the down - flat " phase shifts .
so , in @xcite the up - flat " solution has also been eliminated and a new down - flat " set of phase shifts , compatible with the @xmath125 and @xmath126 data , has been found .
we have verified that these new down - flat " phase shifts , parameterized as described above , satisfy well roy s equations . to conclude , using the theoretical constraints of crossing symmetry , unitarity and analyticity ,
the four - fold ambiguities in the isoscalar @xmath20-wave @xmath0 phase shifts , in the @xmath0 invariant mass range from 800 to 1000 mev , have been eliminated in favour of the down - flat " solution .
we acknowledge very useful correspondence on the calculation of the driving terms with h. leutwyler and b. ananthanarayan .
we thank b. nicolescu and p. enczykowski for fruitful discussions on regge amplitudes .
we are grateful to k. rybicki for helpful collaboration on the @xmath32 phase shifts and to g. colangelo , j. gasser , b. moussallam and j. stern for enlightening comments .
this work has been performed in the framework of the in2p3-polish laboratories convention ( project number 99 - 97 ) .
99 proc . of panic99 int .
conf . on particles and nuclei , uppsala ,
sweden , 1999 , nucl .
a663 & 664 * ( 2000 ) ; proc . of 9th int .
conf . on hadron spectroscopy ( hadron 2001 ) , protvino , russia , 2001 , aip conf . proc .
* 619 * ( 2002 ) , also in
protvino 2001 , hadron spectroscopy " .
j. gasser , h. leutwyler , phys .
125b * ( 1983 ) 325 , _ low energy theorems as precision tests of qcd " _ ; g. colangelo , j. gasser , h. leutwyler , nucl . phys .
* b603 * ( 2001 ) 125 , _ @xmath0 scattering " _ ; b. ananthanarayan , p. bttiker , phys . rev .
* d54 * ( 1996 ) 5501 , _ scattering lengths and medium and high energy @xmath0 scattering"_. proc . of workshop on possible existence of the @xmath3-meson and its implications to hadron physics , kyoto , japan , 2000 , eds .
s. ishida _
et al . _ , kek - proceedings 2000 - 4 , dec .
2000 , soryushiron kenkyu ( kyoto ) * 102 * ( 2001 ) .
s. m. roy , phys .
lett . * 36b * ( 1971 ) 353 , _ exact integral equation for pion - pion scattering involving only physical region partial waves " _ ; s. m. roy , helv . phys .
acta * 63 * ( 1990 ) 627 , _ pion - pion scattering"_. m.r .
pennington , s.d .
protopopescu , phys . rev .
* d7 * ( 1973 ) 1429 , _
@xmath0 scattering amplitude in the low - energy region " _ ; m.r .
pennington , s.d .
protopopescu , phys . rev .
* d7 * ( 1973 ) 2591 , _ how roy s equations resolve the up - down ambiguity and reproduce the @xmath130 resonance"_. j .- l .
basdevant , j .- c .
le guillou , h. navelet , nuovo cim . * 7a * ( 1972 ) 363 , _ crossing and physical partial - wave amplitudes " _ ; j .-
basdevant , c.d .
froggatt and j. l. petersen , nucl . phys .
* b72 * ( 1974 ) 413 , _ construction of phenomenological @xmath0 amplitudes"_. | roy s equations , which incorporate crossing symmetry of the @xmath0 scattering amplitudes , are used to resolve the present ambiguity between two solutions for the scalar - isoscalar phase shifts below 1 gev .
it is shown that the down - flat " solution satisfies well roy s equations and consequently crossing symmetry while the other solution called up - flat " does not and thus should be eliminated .
= 15.5 cm = -0.4 cm = -1.2 cm |
the concept of quantum coherence is one of the most fundamental in quantum mechanics . according to the general principles quantum coherence of the wave function
can not be destroyed due to _
elastic _ interaction with a _
static _ external potential . on the other hand ,
_ inelastic _
scattering processes may ( and in general do ) destroy the phase coherence of the wave function . for electrons in a metal
such processes ( e.g. inelastic electron - electron and electron - phonon scattering ) are important at sufficiently high temperatures .
but as the temperature @xmath0 is lowered inelastic processes become less intensive and the corresponding inelastic scattering time @xmath2 for an electron in equilibrium tends to infinity at @xmath3 .
hence , one could consider natural that interaction may cause dephasing in a metal at nonzero @xmath0 , but _ not at _ @xmath1 .
surprizingly , it was found in many experiments with disordered metals ( see @xcite and references therein ) that the effective dephasing time @xmath4 for the electron wave function saturates at the level @xmath5 ns and does not depend on temperature below @xmath6k .
these findings are in a _ clear contradiction _ with the above point of view according to which the time @xmath4 should increase to infinity as @xmath0 approaches zero .
scattering on magnetic impurities , heating and the external noise were suggested as possible reasons for saturation of @xmath4 observed in earlier experiments .
all these reasons were convincingly ruled out by the authors @xcite , and it was argued that it is the _ zero point motion of electrons _ in a disordered conductor that causes dephasing at @xmath1 .
a theory of the effect of quantum decoherence at low temperatures was suggested by the present authors @xcite .
it was demonstrated that at sufficiently low @xmath0 _ the high frequency quantum noise _ of the effective electronic environment is responsible for the effect of quantum decoherence .
it was predicted that at low @xmath0 the decoherence rate @xmath7 for 1d metallic systems saturates at the level @xmath8 where @xmath9 is the wire conductance per unit length , @xmath10 is the fermi velocity . the result ( [ 1 ] )
is in a very good agreement with the experimental findings @xcite as well as with previous experimental results .
the predictions @xcite inevitably lead to another _ fundamental conclusion _ : low dimensional metals _ do not become insulators even at _ @xmath1 .
indeed , e.g. in the case of 1d systems the localization length is @xmath11 , while the result ( [ 1 ] ) yields the effective decoherence length @xmath12 , where @xmath13 is the elastic mean free path and @xmath14 is the effective number of conducting channels .
it is obvious that for @xmath15 ( which is always the case in metallic wires ) @xmath16 is _ parametrically _ smaller than @xmath17 .
this implies that due to interaction with other electrons the phase coherence of the electron wave function is destroyed _ faster _ than the electron can get localized , i.e. localization _ never _ takes place under the above conditions . since the results @xcite
imply a necessity to strongly reconsider the commonly adopted point of view on the role of interaction in disordered metals at low @xmath0 , it is not surprizing that the subject still remains controversial .
moreover , it is sometimes argued that quantum decoherence at @xmath1 would contradict to general principles of quantum mechanics .
what are the arguments against the quantum noise in a disordered metal as a reason for quantum decoherence of electrons ?
usually it is argued on a general level , adopting a much simpler quantum mechanical model in order to illustrate the main statement .
on one hand this makes sense indeed : the nature of a fundamental quantum mechanical effect ( or the absence of it ) can and should be understood without unnecessary complications . on the other hand
, one has to make sure that all significant features which yield the effect in a more complicated problem are still present in a simplified one .
there is an obvious danger in making too many simplifications : one can easily throw out the baby with the bathwater .
consider a quantum particle colliding with a harmonic oscillator with the frequency @xmath18 .
assume that the interaction potential is short range and before the collision the particle was in the state with the energy _
smaller _ than @xmath19 , while the oscillator was in the ground state .
the general features of this _ scattering problem _ are well known : after interaction the oscillator remains in the ground state ( the particle does not have enough energy to excite it ) , the particle energy also remains the same as before interaction , the particle wave function although in general changes as a result of interaction stays fully coherent .
thus no energy exchange between the particle and the oscillator has occured , and no phase coherence has been lost due to interaction .
what remains is to understand if the above consideration is relevant to the problem in question . at the first sight it appears to be relevant indeed .
although in a metal the electron interacts with other electrons instead of oscillators ( let us ignore the electron - phonon interaction for simplicity ) , at @xmath1 all states below / above the fermi energy should be occupied / empty , and no energy can be transferred : the electron can not be scattered into any lower energy state due to the pauli principle as well as to any higher energy state because no energy can be extracted from other electrons due to the same reason .
it follows immediately , that at @xmath3 the phase coherence of the electron wave function in a metal can not be lost due to the very same reason as in the above example with the particle and the oscillator : there exists no inelastic scattering .
this consideration turns out to be too naive , however
. it may be sufficient if the _
infinite _ time limit can be taken already at the very first step of the calculation as e.g. within the standard formulation of the scattering problem . in this case
the result will be proportional to the delta - function of the energy difference between the initial and the final states ensuring the energy conservation .
however , at any finite time the energy - time uncertainty principle should be taken into account .
the energy uncertainty can be not necessarily small , especially if the particle interacts with an _ infinite _ number of oscillators .
this is crucially important for the effect in question . on top of
that our problem is more complicated because we are dealing with an _ interacting _ quantum many body system rather than the scattering problem for two particles with well defined in- and out - states .
the electron energy fluctuates as a result of interaction with other electrons , only the total energy of all electrons plus the interaction energy is conserved .
one may introduce the effective oscillators also in this case , they are just the modes of the electromagnetic field .
the interaction between the electron and one such mode can be naively described as @xmath20 , where @xmath21 is the amplitude of the electric field , @xmath22 is the wave vector and @xmath23 is the electron coordinate .
the interaction potential remains finite for any @xmath23 , i.e. the electron always interacts with the oscillator and the scattering problem can not be formulated .
actually the problem is even more complicated because the electron interacts with an infinite number of such oscillators at the same time .
it is also important that _ in the presence of interaction _ the energy exchange between the low energy particle and the high energy oscillator is possible _ without excitation _ of the latter simply because its lowest energy level acquires a finite width .
this effect depends on the strength of interaction , but it always exists because the interaction is never `` turned off '' .
one might argue that this simply means that electrons are `` bad '' particles in the presence of interaction and one should rather define `` better behaving '' quasiparticles and calculate all measurable quantities in their terms .
this does not always help because of at least two reasons .
one is that quasiparticles ( if they exist ) are usually defined within an approximate procedure . although the approximation may work well for calculation of certain physical quantities
, it may fail for other quantities .
another reason is that the transformation of the basis ( even if it can be done exactly ) may not be convenient if the measurements are done only with a `` bad '' particle . in this case calculation with `` good '' quasiparticles can be by far more complicated , and it is much simpler to `` get rid '' of all but one `` interesting '' degree of freedom at an early stage of calculation by tracing them out in the full density matrix .
this is the key idea of the feynman - vernon theory of the influence functionals @xcite developed further by caldeira and leggett @xcite and schmid @xcite in application to an infinite bath of harmonic oscillators with the ohmic spectrum .
superconducting josephson junctions may serve as an example of a fermionic system where the same ideas have been worked out @xcite .
in this case it is possible to exactly integrate out all electron degrees of freedom and describe the system dynamics in terms of only one collective variable the josephson phase @xmath24 .
of course , @xmath24 is known to behave `` badly '' ( its quantum dynamics is incoherent in almost all cases , see e.g. @xcite ) but namely this variable is of interest in josephson junctions and squids just because the junction current and voltage operators as well as the flux operator in squids are defined in terms of @xmath24 . in this case `` better behaving '' quasiparticles are simply irrelevant .
therefore it appears to be unreasonable to argue in favour of `` coherent '' quasiparticles if all measurements are done only with `` incoherent '' variables .
we believe that a similar situation is encountered if one calculates the conductance of a disordered metal . actually in this case it is even more complicated because it is not clear if `` good '' quasiparticles can be introduced at all .
the paper is organized as follows . in section 2
we will demonstrate how the fundamental effect of quantum decoherence can be derived within a simple model of a quantum mechanical particle interacting with an environment consisting of a collection of harmonic oscillators .
it is remarkable that already this simple model captures all significant features of the effect and allows to understand why the arguments against its existence fail .
we also establish the relation between our results and the standard perturbative treatment of a scattering problem .
section 3 is devoted to a rigorous microscopic analysis of quantum decoherence in a disordered metal .
discussion of recent experiments and comparison with our theory are presented in section 4 .
the main conclusions are outlined in section 5 .
consider a quantum - mechanical particle ( or a more general system ) characterized by a coodinate @xmath25 and the action @xmath26 $ ] .
assume that this particle interacts with another quantum system described by a coordinate @xmath27 and the action @xmath28 .
we will call the latter quantum system `` environment '' .
the total action for the system `` particle+environment '' has the form @xmath29=s_0[q]+s_{\rm
env}[q(t)]+s_{\rm int}[q(t),q(t ) ] , \label{sqq}\ ] ] where @xmath30 describes interaction between @xmath31 and @xmath32 .
suppose we are interested in the probability @xmath33 for the particle @xmath31 to have the coordinate @xmath34 at a time @xmath35 provided at @xmath36 it had the coordinate @xmath37 . a general and elegant way to solve this problem can be formulated within the feynman - vernon theory of the so - called influence functionals .
an extensive discussion of this theory can be found in refs .
@xcite . here
we only repeat the key steps . by definition the probability @xmath33 is given by the square of a transition amplitude @xmath38 . in the absence of interaction ( @xmath39 )
the oscillator coordinate @xmath32 does not enter into the expression for @xmath40 .
however for @xmath41 there appears an additional force acting on a particle @xmath31 .
this force depends on @xmath32 which is itself a quantum variable .
thus we can not anymore restrict ourselves to the dynamics of @xmath31 , but rather should deal with the total hamiltonian for a system `` @xmath42 '' or , equivalently , with a complete action ( [ sqq ] ) .
of course , the probability @xmath33 is again equal to the square of the corresponding transition amplitude @xmath40 which now depends on both @xmath31 and @xmath32 .
it is important , however , that we are not interested in the final state of the oscillator and do not make any measurements of @xmath32 .
therefore we should sum over _ all _ possible final states of @xmath32 and the formula for the probability @xmath33 takes the form @xmath43 this is an important point .
each of the terms in the sum ( [ w1 ] ) represents a probability for a system to come into the state @xmath44 . since the subsystem @xmath32 can be in _ any _ final state @xmath45 , in order to find the total probability @xmath33 we have to add all these probabilities together
one can slightly generalize the problem and describe the evolution of the density matrix of the system @xmath46 from some initial to some final state .
this evolution is described by the equation @xmath47 where @xmath48 is the initial density matrix of the particle @xmath31 . for the sake of simplicity in what follows
we will assume that there is no interaction between @xmath31 and @xmath32 before @xmath36 .
then the total initial density matrix can be factorized as @xmath49 . here
the kernel @xmath50 is again given by the product of two amplitudes .
it is convenient to represent this product in terms of a double path integral @xmath51-{\rm i}s_0[q_2(t)])f[q_1(t),q_2(t ) ] , \label{j}\ ] ] where @xmath52 is the _ influence functional _ which describes the total effect of the subsystem @xmath32 on the particle @xmath31 .
the functional @xmath52 in turn can be represented in terms of a double path integral @xmath53&=&\sum_f\int { \rm d}q_i\int { \rm d}q'_i\int_{q_i}^{q_f } { \rm d}q_1 \int_{q'_i}^{q'_f}{\rm d}q_2\tilde\rho ( q_i , q'_i)\times \nonumber \\ & & \exp ( { \rm i}s_{\rm env}[q_1]-{\rm i}s_{\rm env}[q_2]+ { \rm i}s_{\rm int}[q_1,q_1]-{\rm i}s_{\rm int}[q_2,q_2 ] ) .
\label{f}\end{aligned}\ ] ] here again ( we cite from @xcite ) `` @xmath54 just means that at some final time @xmath55 after we are no longer interested in the interaction we must take @xmath56 and integrate over all @xmath45 '' , i.e. @xmath57 this completes the general analysis of feynman and vernon .
its important advantage is that no approximations have been done so far , i.e. the above formulas are _
exact_. one more citation from @xcite is in order : `` @xmath52 contains the entire effect of the environment including the change in behavior of the environment resulting from reaction with @xmath31 . in the classical analogue
, @xmath52 would correspond to knowing not only what the force is as a function of time , but also what it would be for every possible motion @xmath25 of the object .
the force for a given environmental system depends in general on the motion of @xmath31 , of course , since the environmental system is affected by interaction with the system of interest @xmath31 '' .
thus all changes of the state of the environment resulting from the interaction with a dynamical variable @xmath31 are _ automatically _ taken into account within the above formalism . in order to proceed
we first assume that the environment consists of a single harmonic oscillator @xmath32 which has a unit mass and a frequency @xmath18 . for simplicity , the interaction is chosen bilinear with respect to both the particle and the oscillator coordinates @xmath31 and @xmath32 , so that the total action for the system `` particle+oscillator '' has the form @xmath29=s_0[q]+\int { \rm d}t\left ( \frac12\dot q^2 -\frac{\omega^2}{2}q^2 + cqq\right ) , \label{sqq1}\ ] ] where @xmath58 is a constant which governs the strength of interaction .
we will also assume that the oscillator is initially kept at a temperature @xmath0 , i.e. the probability to occupy the state @xmath59 is @xmath60 .
in the limit @xmath61 the initial state of the oscillator is its ground state @xmath62 . for this model
it is a matter of a simple integration over the @xmath32-variables to obtain the exact expression for the influence functional @xmath52 ( [ f ] ) .
one finds @xcite : @xmath63=\exp ( -{\rm i}s_r[q_1,q_2]-s_i[q_1,q_2 ] ) .
\label{f2}\ ] ] defining @xmath64 and @xmath65 we have @xmath66 @xmath67 let us point out that @xmath68 for _ all _ trajectories @xmath25 .
( [ f]-[si0 ] ) summarize the complete effect of interaction with the oscillator @xmath32 on quantum dynamics of a particle @xmath31 .
let us come back to the probability @xmath33 ( [ w1 ] ) or , more generally , to the kernel @xmath50 ( [ j ] ) .
we will consider two simple examples .
the first example is a freely propagating quantum particle with a mass @xmath69 . in the absence of interaction ( @xmath39 ) we have @xmath70 , and the double path integral ( [ j ] ) decouples into the product of two single integrals . for @xmath71
each of them is dominated by _ the same _ classical path @xmath72 and ( cf .
e.g. @xcite ) @xmath73 ) , \;\;\ ;
s_0[\tilde q]=m(q_f - q_i)^2)/t . \label{amp}\ ] ] the probability @xmath74 does not depend on the phase @xmath75 of each of the amplitudes , these phases enter with the opposite signs and cancel .
now let us turn on the interaction ( @xmath76 ) .
if just for the sake of simplicity
one treats this interaction perturbatively , one immediately observes that the double path integral ( [ j ] ) is again dominated by the same classical path @xmath77 . for _
any _ @xmath78 we have @xmath79 and therefore @xmath80\equiv 1 $ ] like in the absence of interaction .
one can also establish the general form of the kernel @xmath50 ( [ j ] ) .
it is relatively complex and is not presented here .
more interesting situation emerges if we modify our model and consider our particle @xmath31 interacting with @xmath81 oscillators with frequencies @xmath82 .
the corresponding generalization is trivial : one should just substitute @xmath83 in ( [ sri1 ] ) and carry out the summation over all @xmath84 .
if one sends @xmath81 to infinity and assumes a continuous distribution of the oscillator frequencies @xcite : @xmath85= \eta \omega , \;\;\;\
; |\omega | <
\omega_c , \label{ohm}\ ] ] ( @xmath86 defines the high frequency cutoff ) one arrives at the influence functional of the form ( [ f2]-[si0 ] ) where one should substitute @xmath87 the problem defined by eqs . ( [ f2]-[si0],[subs ] ) is gaussian and can be solved exactly , see e.g. @xcite .
performing a straightforward gaussian integration over @xmath31 one arrives at the the exact expression for the kernel @xmath50 ( [ j ] ) : @xmath88 . \label{jint}\end{aligned}\ ] ] here @xmath89 , @xmath90 are initial / final values of @xmath91 , @xmath92 and the functions @xmath93 and @xmath94 tend to the following values in the interesting limit of long times : @xmath95 and @xmath96 .
the obtained exact solution allows to make several important observations .
one of them is completely obvious : the particle @xmath31 looses its coherence due to interaction with the caldeira - leggett bath of oscillators . indeed in the long time
limit we have @xmath97 and the kernel ( [ jint ] ) effectively reduces to @xmath98 where @xmath99 is the system size . in other words , for any initial conditions the density matrix tends to the same equilibrium form @xmath100 which is not sensitive to the initial phase . according to eqs .
( [ jint],[f1 ] ) the decay of off - diagonal elements of the initial density matrix is exponential at any nonzero @xmath0 with the characteristic time @xmath101 . at @xmath1 the off - diagonal elements decay as a power law @xmath102 but also in this case the information about the initial phase is practically lost in the long time limit .
another observation is that at sufficiently long times the average value of the kinetic energy of the particle @xmath103 ( [ energy ] ) is not zero even at @xmath1 _ irrespectively _ to its initial energy . at high temperatures @xmath104
the energy is given by its classical value @xmath105 , but at lower @xmath0 its value is determined by the interaction parameter @xmath106 and the high frequency cutoff parameter @xmath86 .
it is sometimes believed that if initially all the bath oscillators are in the ground states and the particle energy is zero , no energy exchange between the particle and the oscillators will be possible because the particle has no energy to excite the oscillators and the latter in turn can not transfer their zero - point energy to the particle .
this statement is obviously incorrect in the presence of interaction : the low energy particle will increase its average energy while the interaction energy will be lowered to preserve the energy conservation for the whole system .
the presence of @xmath86 in eq .
( [ energy ] ) implies that there exists the energy exchange between the particle and _ all _ oscillators including the high frequency ones with @xmath107 .
one should not think , however , that such oscillators need to be excited in order to make this exchange possible .
the energy transfer mechanism is different : in the presence of interaction the oscillator energy levels ( including the ground state one ) acquire a finite width and the oscillator can exchange energy in arbitrarily small portions .
it is important to emphasize that some of the above effects can not be correctly described within a naive perturbation theory in the interaction based e.g. on the fermi golden rule . just for an illustration let us choose the plane wave @xmath108 as the initial state of the particle @xmath31 and evaluate the transition probability @xmath109 to the state with the momentum @xmath110
. without interaction one has @xmath111 . for small @xmath112 at @xmath1 in the long time
limit one gets from eqs .
( [ rho1 ] ) , ( [ jint ] ) @xmath113 where @xmath114 and @xmath115 .
it is obvious that the result ( [ wpp ] ) can not be recovered in any finite order of the perturbation theory in @xmath116 .
the above model can also serve as an illustration of the role of `` good '' quasiparticles in the effect of quantum decoherence .
it is clear that in this model one can carry out exact diagonalization of the hamiltonian and introduce a new set of independent ( and therefore coherent ) particles / oscillators .
it is also clear that this transformation will by no means influence our result for the density matrix @xmath46 : this result is exact .
thus also the calculation with `` good '' quasiparticles will yield the incoherent dynamics of @xmath31 , however with much more efforts and with loss of physical transparency .
the basic reason for dephasing of @xmath31 is , however , transparent also in this case : @xmath31 will be expressed as a sum of _ infinite _ number of independent particle / oscillator cordinates and therefore will be able to return to its initial state only after infinite time .
the second example is a quantum particle on a ring .
again we would like to calculate the probability @xmath33 , but now we have to choose both @xmath117 and @xmath118 on a ring .
we choose @xmath119 ( see fig.1 ) , i.e. @xmath33 is the probability for a particle to return to the same point @xmath120 . again without interaction
the two amplitudes @xmath40 and @xmath121 decouple and can be evaluated separately .
we have @xmath122 , where @xmath123 is the contribution from a path which traverses @xmath69 times along the ring in a clockwise ( @xmath124 ) or a counterclockwise ( @xmath125 ) way and returns to the starting point @xmath120 .
it is obvious that @xmath126 , where @xmath127 is the ring radius .
if we neglect terms @xmath128 with @xmath129 ( one can argue that e.g. for @xmath130 those are fast oscillating with @xmath131 terms which effectively cancel out in the course of summation over @xmath69 and @xmath132 ) , then the probability @xmath33 can be written as a sum of two terms @xmath133 , where @xmath134 the term @xmath135 is determined by a pair of equivalent paths @xmath78 ( fig .
this contribution does not vanish in the classical limit .
the term @xmath136 comes from a pair of _ time reversed _
paths @xmath137 ( fig .
this term describes the effect of quantum interference and therefore is very sensitive to the presence of the phase coherence in our system . obviously , @xmath136 vanishes in the classical limit . as before , in both cases the phase factors @xmath138 enter with opposite signs and cancel in each of the terms in ( [ w12 ] ) . without interaction
we have @xmath139 .
now let us analyze the effect of interaction .
the expression for the return probability @xmath135 turns out to be insensitive to interaction due to exactly the same reason as in our first example : this probability is determined by the pairs of _ equivalent _ paths @xmath140 with @xmath141 .
in contrast to @xmath135 , the return probability @xmath136 determined by the pairs of time reversed paths _ is _ affected by interaction .
we will restrict ourselves to the most interesting physical situation when the energy of a particle @xmath31 remains conserved during its motion .
then the simplest pair of the time reversed paths is : @xmath142 and @xmath143 is an even function of @xmath144 . substituting these paths into ( [ sr0],[si0 ] ) after a simple integration we find @xmath145 the result
( [ sri1 ] ) demonstrates that the return probability @xmath136 for a particle @xmath31 after a time @xmath35 for the time reversed paths acquires the factor @xmath146 due to interaction with a harmonic oscillator with a frequency @xmath18 .
the probability @xmath135 remains unaffected .
the whole consideration can be trivially generalized to the case of more complicated time reversed paths with an arbitrary winding numbers @xmath69 . in this case
we again find an additional factor ( [ fsi ] ) in the interference term @xmath136 and @xmath147 is determined by an expression similar to ( [ sri1 ] ) which also depends on @xmath69 .
the term @xmath148 is again zero for such paths for all @xmath69 .
since no new effects emerge at @xmath149 , in what follows we will analyze only the simplest case @xmath150 ( [ sri1 ] ) .
the first conclusion one can draw from the result ( [ sri1 ] ) is that no qualitative changes in the system behavior emerges if one varies the temperature @xmath0 .
the value @xmath151 is smaller at low @xmath61 as compared to the high temperature limit @xmath152 but the effect persists even at @xmath1 .
thus the relation between @xmath0 and @xmath18 turns out to be important only in a quantitative sense , no qualitative dependence on this relation should be expected .
a much more important parameter is @xmath153 .
we see that the value @xmath151 and hence @xmath136 oscillate in time with a period @xmath154 . after each such period
the interference term @xmath136 restores its `` nonineracting '' value while at all intermediate times the value @xmath136 is smaller than in the noninteracting case .
in this situation we still can not speak about decoherence : the system keeps information about its initial phase and periodically returns to its initial state . on top of that
if the radius of the ring @xmath127 is constant in time ( see below ) we have @xmath155 , i.e. the oscillations of @xmath136 practically disappear in the long time limit .
all these results are not surprizing : one should not expect to find the decoherence effect in the system of two quantum mechanical particles .
quantum decoherence appears after the next step : we again modify our model coming from the interaction with one oscillator to the infinite set of oscillators with the ohmic spectrum ( [ ohm ] ) . in this case after the integration of ( [ sri1 ] ) over @xmath18 one gets @xmath156 this equation yields @xmath157 at @xmath1 and @xmath158 at @xmath159 .
these results are in a nice qualitative agreement with those obtained in the exactly solvable model studied above . there the exponential decay of the off - diagonal elements of the initial density matrix @xmath160 with @xmath161 was found at any @xmath162 while at @xmath1 a power law decay ( [ power ] ) was observed . in both cases
the similarity is obvious if one interchanges @xmath127 and @xmath163 .
we observe from ( [ fsi ] ) that if @xmath147 is small the influence functional @xmath164 and no decoherence occurs .
however for @xmath165 the interference of the time - reversed paths is completely suppressed , and quantum dephasing takes place . by setting @xmath166 we can , therefore , define the typical size of the ring @xmath167 beyond which the effect of quantum decoherence becomes important . in our particular example
the characteristic dephasing length @xmath16 decreases as @xmath168 at @xmath159 and it is constant @xmath169 at @xmath1 .
this is the effect of quantum decoherence .
note that by fixing @xmath127 and increasing @xmath35 in our problem we effectively decrease the particle velocity @xmath170 which is eventually sent to zero as @xmath35 approaches infinity .
therefore it is quite natural that at @xmath1 the interference term @xmath136 , although suppressed by a factor @xmath171 , does not decay in time . qualitatively the same property is observed in our result for the decoherence time in a disordered metal ( [ 1 ] ) : if we treat @xmath10 as a formal parameter which may be put equal to zero we will immediately arrive at a zero decoherence rate @xmath7 in this limit . the same is true for the high temperature result @xcite .
the above situation is , however , not very relevant for a metal where conducting electrons move with the velocity @xmath172 which absolute value does not change in time .
thus to account for that we should rather keep the velocity of a particle @xmath173 fixed .
this implies that as we increase @xmath35 the radius @xmath127 for a classical return path increases linearly with time . in other words
, we can slightly modify our model allowing our particle to choose the ring with a proper @xmath127 for each time @xmath35 ( see fig .
then we immediately observe that @xmath147 ( [ si5 ] ) grows in time as @xmath174 at @xmath1 and @xmath175 at @xmath159 , i.e. in this case the decay of the quantum interference term @xmath136 caused by interaction with the caldeira - leggett bath of oscillators is _
faster than exponential_. finally , let us note that in a disordered metal , although the length of the electron trajectory increases linearly with time , the dynamics is diffusive and the effective loop size grows as @xmath176 . substituting this expression into ( [ si5 ] ) one obtains @xmath177 where @xmath178 and @xmath179 are unimportant numerical coefficients of order one . although the above simple model can not be directly applied to disordered metals it demonstrates several important properties which will be also observed in a rigorous calculation .
one such property is that the suppression of quantum interference between time reversed paths increases if the size of the loop grows in time .
another property is that oscillators with @xmath180 may give the dominating contribution to dephasing . in the above model the frequencies @xmath181
give the maximum contribution , but if the spectrum of the problem is different ( as e.g. in a disordered metal , see below ) high frequency modes @xmath182 may also become important .
the physical reasons for this conclusion were already clarified above : in order to have energy exchange in an interacting system it is not necessary to excite the high frequency oscillators , broadening of their ground state levels is sufficient .
it is well known that the high frequency cutoff @xmath86 enters the expression for the interaction induced decoherence rate of a quantum particle in the periodic @xcite and the double well potentials @xcite . the same is observed here for an exactly solvable model of a free damped particle in the limit @xmath1 .
our model also demonstrates at which step of our calculation the effect of dephasing appears .
interaction of the particle @xmath31 with one harmonic oscillator leads to the oscillations of the interference term @xmath136 with the oscillator frequency @xmath18 .
these oscillations are natural since the initial state is not the eigenstate of the system .
yet no dephasing appears .
if coupling to _ many _ oscillators with different frequencies is introduced the probability @xmath136 will be _ always _ suppressed and the system will _ never _ return to its initial state .
obviously this mechanism of dephasing has nothing to do with the temperature of the environment and it persists even at @xmath1 . in this limit
the exponential decay of @xmath136 in time is due to increase of the loop size with @xmath35 .
the phase breaking length @xmath16 ( [ l ] ) depends only on the interaction strength and appears to be ( roughly ) insensitive to the particular ( e.g. ballistic or diffusive ) type of a particle motion .
one might wonder what is the relation between the above analysis and the standard perturbative treatment of a scattering problem for a quantum particle interacting with a harmonic oscillator . beside its general importance
the scattering approach is also of a practical relevance because it is frequently applied to conductance calculations in mesoscopic systems .
the usual definition of a scattering problem operates with in- and out - scattering states measured after exactly _ infinite _ time . in other words ,
the limit @xmath183 is taken already at the very first step , and the whole calculation is carried out only in this limit .
this is sufficient in many physical situations , but not for our problem due to the reasons to be clarified below . here
we will keep @xmath35 finite ( although possibly large ) throughout the calculation and let it go to infinity in the end . with this in mind a direct connection to the scattering problem can be easily established . consider a quantum particle @xmath31 scattered on a harmonic oscillator with a frequency @xmath18 . before scattering the oscillator
is assumed to be in its ground state , and the particle is in the state with wave function @xmath184 and the energy @xmath185 . assuming the interaction to be of the same form as in ( [ sqq1 ] ) and proceeding perturbatively in the interaction strength @xmath58
one can easily derive the probability @xmath186 for a particle @xmath31 to leave the state @xmath184 after the time @xmath35 .
it is just the sum of the transition probabilities @xmath187 into all possible final states @xmath188 which are orthogonal to @xmath184 in the absence of interaction .
one easily finds ( see e.g. @xcite ) @xmath189 where @xmath190 .
this result implies @xmath191 , i.e. at large @xmath35 the transition rate experiences fast oscillations and approaches @xmath192 at @xmath183 .
if @xmath193 for all @xmath194 ( i.e. @xmath184 describes the ground state of the noninteracting problem ) the transitions are highly improbable at sufficiently long times and one may conclude that with the dominating probability the particle remains in its initial state and no quantum dephasing takes place .
we would like to emphasize , however , that the time average of the escape probability @xmath33 ( [ gr ] ) is not equal to zero already in this case .
this is a direct consequence of the energy - time uncertainty principle .
let us now consider scattering on _ many _ oscillators . as before we will assume that the frequency spectrum for these oscillators is ohmic ( [ ohm ] ) .
in order to define the scattering problem we also assume that interaction exists only in a certain space region @xmath195 , i.e. we put @xmath196 . without interaction the eigenstates of a problem
are the plane waves @xmath197 where @xmath198 .
the transition matrix elements can be easily evaluated .
they give an important contribution for @xmath199 in which case we get @xmath200 .
subsituting this expression into ( [ gr ] ) and carrying out the summation over the final states @xmath201 and over the oscillator frequencies ( making use of ( [ subs ] ) with @xmath202 ) , we find @xmath203 further assuming that the initial particle energy is small @xmath204 after simple integrations we obtain @xmath205 we observe that the escape probability @xmath186 is not zero and , moreover , it is not necessarily small .
the reason for that is transparent . although the contribution of each oscillator to @xmath186 is small , it is not zero at any finite @xmath35 due to the energy - time uncertainty principle .
the sum of these small contributions from _ many _ oscillators is finite and yields the result ( [ gr3 ] ) .
we would like to emphasize that by no means this result is in contradiction with the energy conservation law , rather it demonstrates that one should be careful applying the energy arguments to describe the time evolution of an interacting quantum system , especially if it consists of an infinite number of degrees of freedom . even at times
much larger than the characteristic time scale ( in our problem the relevant time scale is set by the dephasing time @xmath4 ) the energy uncertainty may be sufficient for @xmath186 to significantly differ from zero . in order to observe the relation of a perturbative expression ( [ gr3 ] ) to our previous results we set @xmath206 to be of order of the system size @xmath207 .
after that the similarity between ( [ gr3 ] ) and e.g. eqs . ( [ power],[si5 ] ) becomes completely obvious . requiring that @xmath208 ( the escape is complete ) and using ( [ gr3 ] ) with @xmath207 one immediately arrives at the estimate ( [ l ] ) for the decoherence length @xmath209 derived previously with the quasiclassical analysis of the exact influence functional
this result leaves no room for doubts concerning the validity of the quasiclassical description of quantum dephasing at @xmath1 .
in fact , the comparison of the results ( [ power],[gr3 ] ) with ( [ si5 ] ) demonstrates that the quasiclassical approximation rather _ underestimates _ the dephasing effect of interaction : @xmath16 may only become shorter if fluctuations around the classical trajectory are taken into account .
another obvious conclusion is that at any @xmath0 ( including @xmath1 ) the effect of quantum decoherence in not only due to low frequency oscillators which , moreover , can be even completely unimportant .
the relative contribution of oscillators with small @xmath18 depends on the particular form of the spectrum ( @xmath210 ) , being more important in the subohmic case @xmath211 and practically irrelevant in the superohmic case @xmath212 when the high frequency oscillators yield the main effect .
the latter situation will be encountered in the next section where it will be shown that for a @xmath213-dimensional disordered metal one has @xmath214 .
again we emphasize that the above results _ do not _ imply that the processes with high energy transfers are important for dephasing at _ any _ @xmath0 .
it is _ erroneous _ to interpret the parameter @xmath18 as describing the energy transfer : in our calculation the integral over @xmath18 always represents the summation over the bath oscillators ( cf .
( [ ohm]-[subs ] ) ) .
no high frequency oscillators need to be excited , ( small ) energy uncertainty for _ infinitely _ many oscillators is sufficient to provide ( large ) dephasing for a particle @xmath31 .
in order to provide a quantitative description of the effect of quantum decoherence in disordered metals it is necessary to go beyond the simple model considered in the previous section and account for @xmath215 fermi statistics and the pauli principle , @xmath216 the specifics of coulomb interaction in a @xmath213-dimensional system and @xmath217 the effect of disorder .
the corresponding analysis is presented below .
our starting point is the standard hamiltonian for electrons in a disordered metal @xmath218 , where @xmath219\psi_\sigma ( { \bf r } ) , \label{h0}\ ] ] @xmath220 here @xmath221 is the chemical potential , @xmath222 accounts for a random potential due to nonmagnetic impurities , and @xmath223 represents the coulomb interaction between electrons .
let us define the electron green - keldysh function @xcite @xmath224 } } { \int { \rm d}v_1{\rm d}v_2\ ; e^{{\rm i}s[v_1,v_2 ] } } , \label{g1}\ ] ] where @xmath225=2{\rm tr}\ln\hat g_v^{-1}+ { \rm i } \int\limits_0^t { \rm d}t'\int { \rm d}{\bf r } \frac{(\nabla v_1)^2-(\nabla v_2)^2}{8\pi}. \label{lndet}\ ] ] here we performed a standard hubbard - stratonovich transformation introducing the path integral over a scalar potential field @xmath226 in order to decouple the @xmath227-interaction in ( [ hint ] ) . in ( [ g1],[lndet ] )
we explicitely defined the fields @xmath228 and @xmath229 equal to @xmath230 respectively on the forward and backward parts on the keldysh contour @xcite .
the matrix function @xmath231 obeys the equation @xmath232 where @xmath233 the solution of ( [ schrodinger ] ) is fixed by the dyson equation @xmath234 the matrix @xmath235 is the electron green - keldysh function without the field @xmath226 .
it is well known that the 1,2-component of the green - keldysh matrix @xmath236 is directly related to the exact electron density matrix @xmath237 where we also defined the `` density matrix '' @xmath238 related to the 1,2-component of the matrix @xmath231 and performed the average over the fields @xmath239 and @xmath240 as defined in ( [ g1 ] ) .
the density matrix @xmath241 contains all necessary information about the system dynamics in the presence of interaction . making use of eqs .
( [ schrodinger]-[dyson ] ) after some formal manipulations ( see @xcite for details ) one arrives at the equation describing the time evolution of the density matrix : @xmath242 - ( 1-\rho_v)\frac{ev^-}{2}\rho_v - \rho_v\frac{ev^-}{2}(1-\rho_v ) , \label{rho5}\ ] ] where we defined @xmath243 and @xmath244 .
it is important to emphasize that the derivation of this equation was performed _ without any approximation _ , i.e. the result ( [ rho5 ] ) _ is exact_. the equation ( [ rho5 ] ) fully accounts for the pauli principle which is important for the fluctuations of the field @xmath245 .
this field is irrelevant for dephasing .
it is quite obvious from ( [ rho5 ] ) that the field @xmath246 plays the same role as an external field .
all electrons move collectively in this field , its presence is equivalent to local fluctuations of the fermi energy @xmath247 .
the pauli principle does not play any role here .
there is no way how the density matrix @xmath241 can `` distinguish '' the _ intrinsic _ fluctuating field @xmath248 from the stochastic _ external _ field , be it classical or quantum .
since the external field is known to lead to dephasing of the wave function , the field @xmath248 should produce the same effect .
as the equation ( [ rho5 ] ) is exact this conclusion is general and _ does not depend on approximations_. in order to proceed further let us assume coulomb interaction to be sufficiently weak and expand the action ( [ lndet ] ) in powers of @xmath226 up to terms proportional to @xmath249 .
after a straightforward calculation ( see e.g. @xcite ) one finds @xmath250= { \rm i}\int\frac{{\rm d}\omega { \rm d}^3k}{(2\pi)^4 } v^-(-\omega ,- k)\frac{k^2\epsilon(\omega , k)}{4\pi}v^+(\omega , k)- \nonumber \\ -\frac{1}{2 } \int\frac{{\rm d}\omega { \rm d}^3k}{(2\pi)^4 } v^-(-\omega ,-
k)\frac{k^2{\rm im}\epsilon(\omega , k)}{4\pi } \coth\left(\frac{\omega}{2t}\right ) v^-(\omega , k ) ; \label{actionf}\end{aligned}\ ] ] where @xmath251 is the dielectric susceptibility of a disordered metal @xmath252 here @xmath253 is the classical drude conductivity , @xmath254 is the metallic density of states and @xmath255 is the diffusion coefficient . for the sake of simplicity in eq .
( [ eps ] ) we disregarded the phonon contribution which will not be important for us here .
the expression ( [ eps ] ) is valid for small wave vectors @xmath256 and small frequencies @xmath257 .
note , that if one considers only nearly uniform in space ( @xmath258 ) fluctuations of the field @xmath226 one immediately observes that eqs .
( [ actionf],[eps ] ) _ exactly _ coincide with the real time version of the caldeira - leggett action @xcite in this limit ( cf .
( [ f2]-[si0],[subs ] ) ) .
taking into account only uniform fluctuations of the electric field one can also derive the caldeira - leggett action expressed in terms of the _ electron coordinate only_. in this case the effective viscosity @xmath116 in the caldeira - leggett influence functional is @xmath259 , where @xmath260 is the sample resistance ( in contrast to the effective viscosity for the field @xmath226 which is proportional to @xmath261 ) .
for our present purposes it is not sufficient to restrict ourselves to uniform fluctuations of the collective coordinate @xmath226 of the electron environment .
the task at hand is to evaluate the kernel of the operator @xmath262 where the sum runs over the states of the electromagnetic environment with all possible @xmath59 and @xmath18 . averaging over @xmath263 amounts to calculating gaussian path integrals with the action ( [ actionf ] ) and
can be easily performed .
we obtain @xmath264 @xmath265- { \rm i}s_0[{\bf r}_2,{\bf
p}_2]- { \rm i}s_r[{\bf r}_1,{\bf p}_1,{\bf r}_2,{\bf p}_2]- s_i[{\bf r}_1,{\bf r}_2]\big\ } ; \label{js}\ ] ] where @xmath266=\int\limits_{t'}^t { \rm d}t '' \bigg({\bf p\dot r } - \frac{{\bf p}^2}{2 m } - u({\bf r})\bigg ) \label{s_0}\ ] ] is the electron action , @xmath267&= & \frac{e^2}{2}\int\limits_{t'}^t { \rm d}t_1 \int\limits_{t'}^t { \rm d}t_2 \big\{r(t_1-t_2,{\bf r}_1(t_1)-{\bf r}_1(t_2 ) ) \big[1 - 2n\big({\bf p}_1(t_2),{\bf r}_1(t_2)\big)\big]- \nonumber \\ & & -r(t_1-t_2,{\bf r}_2(t_1)-{\bf r}_2(t_2 ) ) \big[1 - 2n\big({\bf p}_2(t_2),{\bf r}_2(t_2)\big)\big ] \nonumber \\ & & + r(t_1-t_2,{\bf r}_1(t_1)-{\bf r}_2(t_2 ) ) \big[1 - 2n\big({\bf p}_2(t_2),{\bf r}_2(t_2)\big)\big]- \nonumber \\ & & -r(t_1-t_2,{\bf r}_2(t_1)-{\bf r}_1(t_2 ) ) \big[1 - 2n\big({\bf p}_1(t_2),{\bf r}_1(t_2)\big)\big ] \big\ } , \label{sr}\end{aligned}\ ] ] where @xmath268 is the occupation number and @xmath269&= & \frac{e^2}{2}\int\limits_{t'}^t { \rm d}t_1 \int\limits_{t'}^t { \rm d}t_2 \bigg\{i(t_1-t_2,{\bf r}_1(t_1)-{\bf r}_1(t_2))+ i(t_1-t_2,{\bf r}_2(t_1)-{\bf r}_2(t_2))- \nonumber \\ & & -i(t_1-t_2,{\bf r}_1(t_1)-{\bf r}_2(t_2))- i(t_1-t_2,{\bf r}_2(t_1)-{\bf r}_1(t_2))\bigg\}. \label{si}\end{aligned}\ ] ] in equilibrim @xmath270 is just the fermi function . in this case at the scales @xmath271 the functions @xmath127 and @xmath272 are defined by the equations @xmath273 the expression in the exponent of eq .
( [ js ] ) defines the real time effective action of the electron propagating in a disordered metal and interacting with other electrons .
the first two terms represent the electron action @xmath274 ( [ s_0 ] ) on two branches of the keldysh contour while the last two terms @xmath148 and @xmath147 determine the influence functional ( cf .
( [ j],[f2]-[si0],[subs ] ) ) of the effective electron environment . as can be seen from eqs .
( [ sr]-[ri ] ) this influence functional is not identical to one derived in the caldeira - leggett model .
however on a qualitative level the similarity is obvious : in both models the influence functionals describe the effect of a certain effective _ dissipative _ environment .
let us first neglect the terms @xmath148 and @xmath147 describing the effect of coulomb interaction .
then in the quasiclassical limit @xmath275 the path integral ( [ js ] ) is dominated by the saddle point trajectories for the action @xmath274 : @xmath276 with obvious boundary conditions @xmath277 , @xmath278 for the action @xmath279 $ ] and @xmath280 , @xmath281 for the action @xmath282 $ ] . since in a random potential @xmath222
there is in general no correlation between different classical paths @xmath283 and @xmath284 these paths give no contribution to the integral ( [ js ] ) : the difference of two actions @xmath274 in the exponent may have an arbitrary value and the result averages out after summation .
thus only the paths with @xmath279\simeq s_0[{\bf r}_2,{\bf p}_2]$ ] provide a nonzero contribution to the path integral ( [ js ] ) .
two different classes of paths can be distinguished ( see e.g. @xcite ) : \i ) the two classical paths are almost the same : @xmath285 , @xmath286 ( cf . fig .
for such pairs we obviously have @xmath287 and @xmath288 .
physically this corresponds to the picture of electrons propagating as nearly classical particles . in the diffusive limit
these paths give rize to diffusons ( see e.g. @xcite ) and yield the standard drude conductance .
\ii ) the pairs of time reversed paths . in this case
@xmath289 , @xmath290 ( cf .
1b , c ) . in the path integral ( [ j ] ) the trajectories
@xmath291 and @xmath292 are related as @xmath293 and @xmath294 . in a disordered metal
these paths correspond to cooperons and give rize to the weak localization correction to conductivity @xmath295 .
this correction is expressed in terms of the time integrated probability @xmath186 for all diffusive paths to return to the same point after the time @xmath35 ( see e.g. @xcite ) . in the absence of any kind of interaction which breaks the time reversal symmetry
this value coincides with the classical return probability and is given by the formula @xmath296 , where @xmath213 is the system dimension and @xmath178 is the transversal sample size .
the weak localization correction @xmath297 diverges for @xmath298 .
this divergence can be cured by introducing the upper limit cutoff at a certain time @xmath299 .
this time is usually reffered to as decoherence time .
one finds @xcite : @xmath300 the physical reason for the existence of a finite @xmath4 is the electron - electron interaction which breaks the time reversal symmetry in our problem . to evaluate @xmath299
we first note that the functions @xmath127 and @xmath272 ( [ ri ] ) change slowly at distances of the order of the fermi wavelength @xmath301 .
therefore we may put @xmath302 , @xmath303 . here
@xmath304 is a classical trajectory with the initial point @xmath305 and the final point @xmath306 , i.e. we consider trajectories which return to the vicinity of the initial point .
the contribution from the time reversed paths @xmath136 to the return probability has the form @xmath307 -s_i[t , t';{\bf r},{\bf r}_t ] } \right\rangle_{r } , \label{w}\ ] ] where @xmath308 is the probability without interaction and the average is taken over all diffusive paths returning to the initial point .
the value @xmath52 in ( [ w ] ) decays exponentially in time , therefore we may put the average inside the exponent .
it is easy to observe that the term @xmath148 gives no contribution to this average . working out the average of @xmath147 we obtain @xmath309 \right\rangle_{r } } = \exp \left(-t e^2\int\limits_{-\infty}^{+\infty } { \rm d}t \langle i(t,{\bf r}(t)-{\bf r}(0))\rangle_{r}\right ) \label{tau}\ ] ] to find the average over the diffusive paths , we introduce the fourier transform of the function @xmath310 and replace @xmath311 by @xmath312 .
then we get @xmath313 let us first consider a quasi - one - dimensional system with @xmath314 . making use of eqs .
( [ 3d],[eps ] ) and integrating over @xmath59 we find @xmath315 the upper cutoff in ( [ tau1 ] ) is chosen at the scale @xmath316 because at higher @xmath18 the diffusion approximation becomes incorrect . from ( [ tau1 ] )
we obtain @xmath317 in the long time limit ( which can not be reached at @xmath3 ) the decay of @xmath52 is faster than exponential ( cf .
@xcite ) . at sufficiently low @xmath318
this difference becomes important only at @xmath319 where @xmath52 is already exponentially suppressed .
for smaller @xmath35 the decay is exponential and the decoherence rate increases linearly with @xmath0 . defining the effective decoherence length @xmath320 , in the temperature interval @xmath321 one finds @xmath322 where @xmath323 is the dephasing length at @xmath1 . at temperatures lower than @xmath324 the decoherence time and length are almost temperature independent . with the aid of eqs .
( [ dsigma1],[f1d ] ) it is also easy to find the weak localization correction @xmath325 to the drude conductance . in the limit @xmath1
we obtain @xmath326 i.e. @xmath327 , where @xmath328 is the effective number of conducting channels in a 1d mesoscopic system and @xmath329 is the drude conductance of a @xmath213-dimensional sample .
thus for @xmath15 the weak localization correction is _ parametrically _ smaller than the drude conductance even at @xmath1 . for 2d and 3d systems the same analysis yields @xmath330 ,
\ ; \ ; \ ; \ ; \ ; \ ; \ ; \quad{\rm 2d } , \nonumber \\ \frac{1}{\tau_\varphi } & = & \frac{e^2 } { 3\pi^2\sigma\sqrt{2d}\tau_e^{3/2}}[1 + 6(t\tau_e)^{3/2}],\ ; \quad{\rm 3d } , \label{11}\end{aligned}\ ] ] the influence functional @xmath52 decays exponentially @xmath331 except for the case of 2d systems at high @xmath0 where one has @xmath332 .
we observe that also in this case the difference from a purely exponential decay is not significant and can be ignored .
the above formalism provides a rigorous description of the effect of quantum decoherence in disordered metals and allows to derive the decoherence rate @xmath7 at low @xmath0 for various dimensions .
the physical reasons for dephasing of electron wave functions remain the same as in the case of a particle interacting with the caldeira - leggett bath of oscillators .
the corresponding discussion is presented in the previous section .
here we only add several comments .
it might be interesting to investigate the average kinetic energy of an electron in a disordered metal in the presence of interaction . at low @xmath0
it turns out to be temperature independent due to the same reason as for a quantum particle in the bath of oscillators , i.e. due to interaction .
a naive calculation along the same lines as in section 2 yields @xmath333 at @xmath1 .
a more accurate analysis is beyond the frames of the present paper .
this result is just a manifestation of the well known fact : the distribution function @xmath334 of interacting electrons ( not quasiparticles ) is nonzero for any @xmath335 even at @xmath1 @xcite .
an obvious consequence of our results is that the decoherence rate @xmath7 exceeds temperature at sufficiently low @xmath0 .
does this fact imply the breakdown of the fermi liquid theory ( flt ) hypothesis at such @xmath0 ? from a formal point of view the violation of the inequality @xmath336 is yet insufficient for such a conclusion ( see also @xcite ) , although it obviously does not support the flt hypothesis either .
it was demonstrated in section 2 that a finite decoherence rate for a certain variable of interest has no direct relation to the lifetime of quasiparticles ( the latter is infinite in the case of a free particle in the caldeira - leggett bath ) . in our calculation
@xmath4 determines the dephasing time for _ real electrons _ and not for the landau quasiparticles .
our result implies that interacting electrons are `` bad '' particles since their wave functions dephase even at @xmath1 .
the possibility to construct `` better behaving '' quasiparticles remains questionable for disordered metals .
furthermore , even if such quasiparticles exist their properties are completely unknown
. therefore at this stage they can hardly be used for calculation of any physical quantity .
an important advantage of our method is that it allows for a direct calculation of measurable quantities in terms of interacting electrons without appealing to the fermi liquid hypothesis .
our results for the decoherence time at low @xmath0 turn out to be in a remarkably good agreement with available experimental data obtained in various physical systems . in ref .
@xcite we have already carried out a detailed comparison between our theory and the experimental results @xcite obtained for 1d au wires . the agreement within a numerical factor of order one
was observed for all samples studied in @xcite .
since both @xmath4 and @xmath16 are _ defined _ with such an accuracy no better agreement can be expected in principle . here we will present a comparison of our theory with two other experiments carried out with 2d electron gas in semiconductor structures .
these systems are somewhat different from the metallic ones mainly because of much higher effective resistance .
the parameters of the systems were chosen in a way to realize a quasi 1d conducting system with disorder .
one such experiment was carried out by pooke _
@xcite in narrow si pinched accumulation layer mosfets .
these authors studied samples with resistances ranging from 120 to 360 k@xmath337 and observed a finite decoherence rate @xmath7 at all temperatures . on a log - plot ( not shown ) clear signs of saturation at low @xmath0
the corresponding data for @xmath16 @xcite obtained for 3 samples are presented in fig . 2 and in the table 1 together with our theoretical predictions .
note , that the experimental value of @xmath16 at @xmath1 was obtained by a linear extrapolation @xmath338 of the experimental data @xcite .
table 1 [ cols="^,^,^,^,^",options="header " , ] the temperature dependence of the data @xmath16 @xcite is also in excellent agreement with our predictions at all @xmath0 , see fig .
3 . in ref .
@xcite the measured maximum values for @xmath4 were compared with the theoretical formula suggested in refs . @xcite and
a discrepancy by a factor of 50 was reported .
this fact allowed the authors to conclude @xcite that their experimental results for @xmath4 argue against the idea of decoherence by zero - point fluctuations of the electrons .
the comparison of the data @xcite with our theoretical results clearly demonstrates that this conclusion is simply an artefact of the inadequate choice of a theoretical formula adopted in @xcite .
the measurements of the decoherence rate @xcite _ strongly support _ the idea of decoherence due to intrinsic quantum noise rather than argue against it . in ref .
@xcite also another interesting experimental observation was made : a crossover to a highly resistive state was found around @xmath6 k. this observation was interpreted in @xcite as a thouless crossover to the regime of strong localization and was also qualified as contradicting the very idea of quantum decoherence at @xmath1 .
several comments are in order .
\(i ) for the systems studied in @xcite one has @xmath339 .
hence , the weak localization correction at low @xmath0 ( [ delsig ] ) should be of the same order as the drude conductance ( in fact , this is exactly what was observed in @xcite ) and the thouless crossover can not be ruled out theoretically for such small @xmath14 . also the result for @xmath4 may become more complicated for small @xmath14 @xcite because of a somewhat more important role of capacitive effects .
this may in principle lead to a relative increase of @xmath16 ( it appears , however , that this effect is not very important in @xcite ) .
thus even the presence of a thouless crossover in the samples @xcite would by no means contradict our theoretical results and hence the idea of quantum decoherence at @xmath1 .
\(ii ) the interpretation of the observed effect as a thouless crossover is not quite convincing .
this interpretation is based on two reasons @xcite : ( i ) at the crossover @xmath16 was found to be `` only '' @xmath340 times smaller than @xmath17 and ( ii ) the resistance of a wire segment of the length @xmath341 was found to be @xmath342 k@xmath337 .
the first observation does not contradict to our theory which predicts @xmath343 at low @xmath0 .
as to the resistance of a segment @xmath341 it is of the order of the quantum resistance unit also at temperatures @xmath344k , i.e. well above the crossover .
hence , no definite conclusion can be drawn .
\(iii ) the expression for @xmath345 used in @xcite makes sense only provided the condition @xmath346 is satisfied ( @xmath99 is the wire length ) .
for @xmath347 the wire can be viewed as @xmath348 _ independent _ samples connected in series ( in @xcite @xmath81 is typically of order @xmath349 ) . since the the length of each of such samples @xmath16 is @xmath340 times _ smaller _ than the localization length @xmath345 it is somewhat naive to seriously believe that strong localization can be observed in such samples relying only on the `` order - of - magnitude '' character of the relation between @xmath16 and @xmath345 at the crossover . moreover , assuming the dependence @xmath350 one immediately observes that the `` standard '' crossover condition @xmath351 would hold at temperatures @xmath352 times smaller than the actual crossover temperature , i.e. at @xmath353 mk .
we see no way how the thouless crossover can be expected above this temperature range for the systems studied in @xcite .
\(iv ) the inequality @xmath354 is violated in @xcite at all relevant @xmath0 : at the crossover one has @xmath355 depending on the sample , and @xmath356 remains of order one even at @xmath357k , i.e. well in the weak localization regime .
if one believes that the violation of the above condition signals the breakdown of flt due to interaction the whole discussion of the thouless crossover becomes pointless .
an alternative would be to acknowledge that the value @xmath356 may not be the relevant parameter as far as flt is concerned .
but in any case _ real electrons _ dephase and therefore can hardly be described within the thouless scenario of strong localization .
\(v ) since @xmath358 and @xmath17 is several times larger than @xmath16 at the crossover interaction definitely plays a very important role `` helping '' to localize electrons instead of destroying localization .
if so , why not to assume that the whole effect is solely due to interaction and not due to spacial disorder ?
for instance , it is well known that the mobility of a quantum particle in a _ periodic _ potential can decrease dramatically with @xmath0 if this particle is coupled to a dissipative environment ( see e.g. @xcite ) . at @xmath1
this particle can even get localized due to the effect of quantum noise of the environment @xcite which completely destroys the phase coherence of the wave function . within this scenario
the crossover to a highly resistive state @xcite can be considered as _ supporting _ the idea of quantum decoherence due to intrinsic quantum noise .
thus , although a detailed interpretation of the crossover @xcite is still an open problem , presently we see no way to use it as an argument against quantum decoherence at @xmath1 . it was argued in @xcite that the effect of saturation of @xmath4 observed in many experiments at low @xmath0 can be caused by the external microwave noise .
although filtering of external noise is indeed a serious experimental problem it is quite obvious that the above explanation faces several severe problems . without going into details let us just indicate some of them .
firstly , according to the arguments @xcite the dephasing effect of the external noise may not be accompanied by heating for low resistive samples with @xmath359 k@xmath337 , whereas in the opposite case of highly resistive samples heating is unavoidable .
it is not clear how to match this conclusion with the experimental results @xcite where clear signs of saturation of @xmath16 at @xmath360 1k were seen for samples with resistances up to @xmath361 k@xmath362 k@xmath337 .
secondly , the formula ( [ 1 ] ) _ quantitatively _ ( within a factor of order one ) describes the low temperature value of the decoherence time @xmath4 measured in different experiments in at least 10 1d semiconductor and metallic samples . in order to interpret the results of all these measurements in terms of external noise
one should assume that external noise always adjusts itself to a particular value of the sample conductance ( in various experiments these values differ by many orders of magnitude ) and the fermi velocity .
more than that , the corresponding electric field produced by the external noise inside the sample should be always of the same order as one due to the intrinsic quantum noise
. it would be interesting to estimate the probability for such a coincidence in ( at least ) 10 different samples .
thirdly , the presence of the external noise can only be proven experimentally by making experiments with and without necessary filtering and observing different results in these two cases . in ref .
@xcite the external noise power was estimated , however no evidence for its existence in experiments was presented .
in contrast , quantum noise with @xmath363 is well observable reality , see e.g. @xcite .
thus the explanation of the existing experimental results in terms of external noise turns out to be problematic .
perhaps new experiments are needed to unambiguously rule this issue out .
in the present paper we have discussed the fundamental effect of interaction induced quantum decoherence in disordered metals .
we have considered a simple model of a quantum particle interacting with the caldeira - leggett bath of oscillators .
an exact solution in the case of a free damped particle demonstrates that the off - diagonal elements of the particle density matrix decay in the long time limit at all temperatures including @xmath1 . for a particle on a ring
similar results are found .
a very transparent physical picture of the effect of quantum decoherence due to interaction with a quantum bath of oscillators emerges from our analysis .
the interference contribution to the return probability @xmath136 for a particle interacting with _
one _ oscillator with a frequency @xmath18 oscillates in time and is smaller than one for all time moments except @xmath364 when the system returns to its initial state .
if interaction occurs with _ infinitely many _
oscillators with a continuous distribution of frequencies the particle will _ never _ return exactly to its initial state . at @xmath1
the return probability will be suppressed by a factor @xmath365 where @xmath116 is the viscosity of the environment and @xmath127 is the size of the return path .
this defines the typical dephasing length @xmath366 . for a particle in a diffusive environment
the typical size of such a path grows with time as @xmath367 , and the interference probability will decay as @xmath368 .
this is the effect of quantum dephasing .
it has an essentially quantum mechanical nature , therefore its existence at @xmath1 is by no means surprizing . at nonzero @xmath0 the effect increases due to increasing fluctuations in the environment .
the main features of the effect derived by means of our simple model are reproduced within the rigorous analysis developed for a disordered metal , in this case one should account for the pauli principle and the sample dimension .
the electron interacts with the fluctuating electromagnetic field produced by other electrons .
this field can be again represented as a collection of oscillators with a somewhat more complicated spectrum than in the caldeira - leggett model .
quantitatively the results will depend on that , but the physical nature of the effect remains the same .
the effect of quantum dephasing studied here has no direct relation to the question about the existence of `` coherent '' quasiparticles in the problem .
it is obvious that only the behavior of measurable quantities is of physical importance . if these quantities ( e.g. the current or conductance in the case of disordered metals ) are expressed in terms of `` incoherent '' variables , quantum dephasing yields directly measurable consequencies . a convincing illustration for that is provided by the existing experimental data .
our results indicate a necessity to reconsider the commonly adopted point of view on the role of interactions in disordered metallic systems . in particular
, one arrives at the conclusion that no electron localization takes place and low dimensional disordered metals with generic parameters do not become insulators even at @xmath1 .
we would like to dedicate this paper to the memory of the late albert schmid whose works constitute one of the major contributions to the field .
it is a pleasure to acknowledge numerous stimulating discussions with c. bruder , y. imry , yu .
makhlin , a. mirlin , g. schn and p. wlfle in the course of this work .
we also profitted from discussions with ya .
blanter , g. blatter , t. costi , yu .
gefen , v. geshkenbein , v. kravtsov , r. laughlin , a. van otterlo , m. paalanen , d. polyakov and a. rosch .
this work was supported by the deutsche forschungsgemeinschaft within sfb 195 and by the intas - rfbr grant no .
95 - 1305 .
p. mohanty , e.m.q .
jariwala , and r.a .
webb , phys .
lett . * 78 * , 3366 ( 1997 ) .
golubev and a.d .
zaikin , preprints ( cond - mat/9710079 and cond - mat/9712203 ) .
r.p . feynman and f.l .
vernon jr . , ann .
( ny ) * 24 * , 118 ( 1963 ) .
r.p . feynman and a.r .
hibbs , _ quantum mechanics and path integrals _ ( mcgraw hill , ny , 1965 ) .
a.o . caldeira and a.j .
leggett , phys .
* 46 * , 211 ( 1981 ) ; ann .
( ny),*149 * , 374 ( 1983 ) .
a. schmid , j. low temp .
phys . * 49 * , 609 ( 1982 ) .
v. ambegaokar , u. eckern , and g. schn , phys .
. lett . * 48 * , 1745 ( 1982 ) ; u. eckern , g. schn , and v. ambegaokar , phys .
b * 30 * , 6419 ( 1984 ) .
g. schn and a.d .
zaikin , phys .
rep . * 198 * , 237 ( 1990 ) and references therein .
a. schmid , phys .
* 51 * , 1506 ( 1983 ) ; s.a .
bulgadaev , pisma zh .
39 * , 264 ( 1984 ) [ jetp lett . * 39 * , 315 ( 1984 ) ] ; f. guinea , v. hakim , and a. muramatsu , phys .
* 54 * , 263 ( 1985 ) ; m.p.a .
fisher and w. zwerger , phys .
b * 32 * , 6190 ( 1985 ) .
s. chakravarty and a.j .
leggett , phys .
. lett . * 53 * , 5 ( 1984 ) ; a.j .
et al . _ , rev .
phys . * 59 * , 1 ( 1986 ) ; for an exact solution in the long time limit see f. lesage and h. saleur , preprint ( cond - mat/9712019 ) .
a.o . caldeira and a.j .
leggett , physica a * 121 * , 587 ( 1983 ) ; * 130 * , 374 ( 1985 ) .
altshuler , a.g .
aronov , and d.e .
khmelnitskii , j. phys .
c * 15 * , 7367 ( 1982 ) .
s. chakravarty and a. schmid , phys . rep . *
140 * , 193 ( 1986 ) .
a. stern , y. aharonov , and y. imry , phys .
a * 41 * , 3436 ( 1990 ) .
y. imry , _ introduction to mesoscopic physics _
( oxford university press , oxford , 1997 ) .
keldysh , zh .
. fiz . * 47 * , 1515 ( 1964 ) [ sov .
jetp * 20 * , 1018 ( 1965 ) ] .
e.m . lifshitz and l.p .
pitaevskii , _ statistical physics _ , vol . 2 ( pergamon , new york , 1980 ) .
a.g . aronov and p. wlfle , phys .
b * 50 * , 16574 ( 1994 ) .
pooke , n. paquin , m. pepper , and a. gundlach , j. phys .
cond . mat . * 1 * , 3289 ( 1989 ) .
khavin , m.e .
gershenzon , and a.l .
bogdanov , preprint ( cond - mat/9803067 ) .
p. mohanty and r.a .
webb , phys .
b * 55 * , r13452 ( 1997 ) . r.h .
koch , d.j .
van haarlingen , and j. clarke , phys . rev .
lett . * 47 * , 1216 ( 1981 ) ; r.j .
lett . * 78 * , 3370 ( 1997 ) . | we discuss a fundamental effect of the interaction - induced decoherence of the electron wave function in disordered metals . in the first part of the paper we consider a simple model of a quantum particle interacting with a bath of harmonic oscillators and analyze the physical origin of the effect .
this exactly solvable model also allows to understand why the arguments against the existence of the effect at low temperatures fail .
the second part of the paper is devoted to a rigorous analysis of quantum decoherence in disordered metals .
we also discuss the relation of our results to the recent experiments on gaas structures .
the existence of a finite quantum decoherence rate at low @xmath0 implies that low dimensional disordered metals with generic parameters do not become insulators even at @xmath1 . |
within the past decades great progress has been achieved in hadron spectroscopy @xcite . especially , inspired by the observation of exotic states @xcite ( such as the candidate for the tetraquark @xcite or pentaquark @xcite state etc . ) , the underlying structure of these states attracts much attention both in theory and experiment .
observation of the exclusive photoproduction of exotic hadronic states off baryons , proposed in the refs .
@xcite , is the most direct way to get information about their nature . at higher energies
such processes can be described in terms of regge trajectory exchanges pd77,vg09 . in ref .
@xcite , the @xmath2 and @xmath3 photoproduction were studied with the reggeized model .
it is found that the reggeized model gives a good description for these reactions in and beyond the resonance region .
thus the photoproduction reaction at high energy may be appropriate to study the role of reggeized treatment . in the 1950s , regge proved the importance of extending the angular momentum @xmath4 to the complex field @xcite . for more general reviews about the regge theory ,
later , the exchange of dominant meson regge trajectories was used to successfully describe the hadron photoproduction @xcite .
however , there is one question : has the regge trajectory approach been well tested by experiment at higher energies ? in the past , the @xmath0 ( @xmath5 ) photoproduction
was extensively studied .
the estimation of the exclusive photoproduction cross section for @xmath6 has been performed according to the one - pion exchange ( ope ) mechanism with absorption in @xcite .
the experimental results for the values and energy dependence of this cross section at relatively low ( @xmath720 gev ) energies are quite consistent with the prediction of this model @xcite .
nevertheless the energy range covered by the existing experimental data is not enough to distinguish between the ope prediction and the regge trajectory approach @xcite .
the compass experiment at cern , uses the muon beam , can significantly enlarge the available energy range of virtual photons up to about 150 gev .
compass has a good opportunity to contribute to the study of exotic charmonia via their photoproduction .
however the uncertainties of the theoretical description of photon - nucleon interaction at high energies complicate this task . the process of @xmath8 , where @xmath9 is a virtual photon , has quite good experimental signature and can be used as a benchmark .
possibility to use @xmath6 photoproduction as a benchmark for study of exotic hadrons is discussed also in @xcite .
it is also significant to carry out more theoretical studies on the @xmath1 process in order to clarify the role of the reggeized treatment .
moreover , due to the vector meson dominance ( vmd ) assumption , a photon can interact with a vector meson , which means that the @xmath10 reaction can also proceed through the vector meson dominance ( vmd ) mechanism @xcite .
thus the @xmath11 photoproduction mechanism is also an interesting issue . in this work
, the @xmath1 reaction is investigated using an effective lagrangian approach and the regge model .
in addition to the @xmath12 exchange , the contributions from the vmd mechanism is also considered .
the differential cross section of the @xmath10 reaction is also calculated , which could be tested by further compass experiment .
this paper is organized as follows .
after the introduction , we present the formalism and the main ingredients which are used in our calculation .
the numerical results and discussions are given in sec .
iii . in sec .
iv , we give a detailed illustration of the possibility of the experimental test at compass . finally , the paper ends with a brief summary .
in the present work , an effective lagrangian approach in terms of hadrons is adopted , which is an important theoretical method in investigating various scattering processes @xcite .
figure 1 describes the basic tree - level feynman diagrams for the @xmath0 photoproduction process through general @xmath12 exchange [ fig .
1(a ) ] and vector meson dominance ( vmd ) mechanism [ fig .
1(b ) ] tb65,tb69,th78 .
reaction via @xmath13 exchange .
( b ) same as in ( a ) , but for the case in the frame of the vmd model.,title="fig : " ] reaction via @xmath13 exchange .
( b ) same as in ( a ) , but for the case in the frame of the vmd model.,title="fig : " ] to gauge the contribution of this diagram , we need to know the relevant effective lagrangian densities . for the @xmath14 interaction vertex
we take the effective pseudoscalar coupling @xcite,@xmath15where @xmath16 is the pauli matrix , while @xmath17 and @xmath12 stand for the fields of the nucleon and the pion , respectively .
the coupling constant of the @xmath14 interaction was given in many theoretical works , and we take @xmath18 @xcite .
the commonly employed lagrangian densities for @xmath19 and @xmath20 couplings are @xcite : @xmath21where @xmath22 , @xmath23 , @xmath24 and @xmath25 are the photon , @xmath11 meson , @xmath26 and @xmath12 fields .
@xmath27 is the mass of the @xmath12 meson .
the coupling constant @xmath28 and @xmath29 can be determined by the partial decay widths @xmath30 and @xmath31 , respectively . with the above lagrangian densities , we obtain @xmath32with@xmath33where
@xmath34 is the k@xmath35llen function with @xmath36 . using the partial decay widths of @xmath0 as listed in the pdg book @xcite , we get @xmath37 gev@xmath38 and @xmath39 gev@xmath38 . for the @xmath40 interaction vertex we can also derive it using the vector meson dominance ( vmd ) mechanism @xcite on the assumption that the coupling is due to a sum of intermediate vector mesons . in the vmd mechanism for photoproduction , a real photon can fluctuate into a virtual vector meson , which subsequently scatters from the target proton . under the vmd mechanism ,
the lagrangian of depicting the coupling of the intermediate vector meson @xmath26 with a photon is written as @xmath41where @xmath42 and @xmath43 are the mass and the decay constant of the @xmath26 meson , respectively . with the above equation
, we get the expression for the @xmath44 decay,@xmath45where @xmath46 indicates the three - momentum of an electron in the rest frame of the @xmath26 meson , while @xmath47 is the electromagnetic fine structure constant .
thus , with the partial decay width of @xmath44 @xcite@xmath48we get the constant @xmath49 . to account for the internal structure of hadrons , we introduce phenomenological form factors . for the vertex of @xmath20 ,
the following form factor is adopted @xcite,@xmath50 for the vertices of @xmath14 and @xmath19 , three types of the form factors are considered @xcite : ( i ) the monopole form factor@xmath51(ii ) the dipole form factor@xmath52(iii ) the exponential form factor@xmath53,\ ] ] where @xmath54 and @xmath55 are the free parameters , which can be determined from the data in this work .
@xmath56 is momentum fraction of the proton carried by the neutron . with the effective lagrangian densities as listed above , the invariant scattering amplitudes for the @xmath57 process
can be written as @xmath58 for fig .
1(a ) , and@xmath59for fig .
here @xmath60 and @xmath61 are the photon polarization vector and the polarization vector of the @xmath11 , respectively , @xmath62 and @xmath63 are the dirac spinors for the initial proton and final the neutron , respectively . to describe the behavior at high photon energy , we introduce the regge trajectories @xcite@xmath64\sin [ \pi \alpha _ { k}(t)]},\]]where @xmath65 is the slope of the trajectory and the scale factor @xmath66 is fixed at 1 gev@xmath67 , while @xmath68 and @xmath69 are the mandelstam variables . in addition , the kaonic regge trajectory @xmath70 is @xcite@xmath71 with @xmath72 , the unpolarized differential cross section for the @xmath73 process at the center of mass ( c.m . )
frame is given by @xmath74where @xmath75 denotes the angle of the outgoing @xmath76 meson relative to the beam direction in the c.m .
frame , while @xmath77 and @xmath78 are the three - momenta of the initial photon beam and the final @xmath76 , respectively .
energy dependence of the cross section calculated above for each of the models for the fixed parameter @xmath80 gev of the monopole form factor is shown in fig fig : total_fix_l .
one can see that the difference between the models is about an order of magnitude .
fine tuning of the parameter @xmath81 can be performed on the basis of the experimental results .
the existing experimental data @xcite for @xmath11 photoproduction at low energies are summarized in table i and fig .
the original data from @xcite have been reanalysed in @xcite to take into account the actual branching ratio for @xmath82 decay channel , since it was taken to be 100% .
to fix the same problem we scale the result for the cross section and the error presented in @xcite by the factor of 1.5 .
we also skip the result for @xmath83 gev , which seems are in tension with the shape of the expected theoretical curves .
the minuit code of the cernlib library was used to perform one - parameter @xmath84 fits of the theoretical curves to the @xmath85 data .
the free parameters involved and their fitted values are listed in table ii .
.the experimental data for @xmath11 photoproduction cross section . here the beam energy @xmath86 is in the units of gev , while the cross section @xmath87 is the units of @xmath88b . [ cols="^,^,^",options="header " , ] the fitted parameter @xmath81 satisfys the expectation with a reasonable @xmath89 . and
yet , it is found that the cases with reggeized treatment need a larger @xmath81 . for comparison ,
we also calculate the result with dipole or exponential form factor .
the fitted parameter @xmath90 and @xmath55 are listed in table iii and table iv , respectively .
3 show that the fitted results with the three types of form factor .
it is found that the difference is small the result of improvements in three types of form factor .
therefore , in the following calculation , we only consider the case of monopole form factor . from fig .
3(a ) one can see that the experimental data ( except the point at @xmath91 gev)@xcite for the total cross section of the @xmath1 reaction are well reproduced with a small value of @xmath92 .
the shape of the total cross section via @xmath12 exchange is different from that of the vmd mechanism . with the above equations and the fitted parameters as listed in table ii , the relevant physical results are calculated , as shown in fig .
4 - 6 . in fig .
4 we also present the variation of the total cross section of the @xmath1 reaction within the typical uncertainties of the @xmath81 values . from fig .
4 ( a ) it is seen that the total cross section via @xmath12 exchange is more sensitive than that of the vmd mechanism to the values of @xmath81 . moreover , a comparison of the results from fig .
4 ( a ) and fig .
4 ( b ) reveals that the total cross section becomes less sensitive to the @xmath81 values when the reggeized treatment is added to the process of @xmath10 . in fig .
5 , we show the differential cross section of @xmath10 as a function of @xmath93 .
it is obvious that there is a significant peak structure in the region of low @xmath93 , which increases rapidly near the threshold and then decreases slowly with increasing @xmath93 .
however , it is seen , that the shapes of the differential cross section @xmath94 with the reggeized treatment are much different from that without the reggeized treatment at higher @xmath93 .
the reggeized treatment can lead to that the differential cross section @xmath94 decreases rapidly with increasing @xmath93 , especially at higher energies .
figure 6 presents the differential cross section for the @xmath95 reaction with or without the reggeized treatment at different energies .
it is seen that the differential cross section with the reggeized treatment is very sensitive to the angle @xmath75 and makes a considerable contribution at forward angles .
reaction for the fixed parameter @xmath80 gev . ]
the datas are from ws75,ye72,mn09,gt93,ei69 .
here , ( a ) is the result with monopole form factor , while ( b ) and ( c ) are the results related to the dipole and exponential form factors correspondently.,title="fig : " ] reaction .
the datas are from ws75,ye72,mn09,gt93,ei69 . here , ( a ) is the result with monopole form factor , while ( b ) and ( c ) are the results related to the dipole and exponential form factors correspondently.,title="fig : " ] reaction .
the datas are from ws75,ye72,mn09,gt93,ei69 . here , ( a ) is the result with monopole form factor , while ( b ) and ( c ) are the results related to the dipole and exponential form factors correspondently.,title="fig : " ] reaction for the different values of cutoff parameter @xmath81 .
( b ) same as in ( a ) , but for the case of the regge trajectory exchange.,title="fig : " ] reaction for the different values of cutoff parameter @xmath81 .
( b ) same as in ( a ) , but for the case of the regge trajectory exchange.,title="fig : " ] considering the @xmath11 are usually detected in experiment via the @xmath97 invariant mass , it would be useful to give the theoretical predictions of the differential cross section @xmath98 as a function of the beam energy @xmath99 , which could be tested by further experiment .
since the full decay width of the @xmath0 is small enough in comparison to its mass , the invariant mass distribution for the dalitz process @xmath100 can be defined with the two - body process @xcite@xmath101where the full width @xmath102 mev and the partial width @xmath103 mev are taken @xcite . with the above equations and the fitted parameters as listed in table ii , the invariant - mass distribution @xmath98 for @xmath104 gev is calculated , as shown in fig .
it is seen that there exists an obvious peak at @xmath105 gev .
the compass experiment @xcite is situated at the m2 beam line of the cern super proton synchrotron . since 2002
it has obtained experimental data for positive muons scattering of 160 @xmath106 ( 2002 - 2010 ) or 200 @xmath106 momentum ( 2011 ) off solid @xmath107lid ( 2002 - 2004 ) or nh@xmath108 polarized targets ( 2006 - 2011 ) .
particle tracking and identification is performed in a two - stage spectrometer covering a wide kinematical range .
the trigger system comprises hodoscope counters and hadron calorimeters . according to the presented calculations of the @xmath11 production cross section and previously published compass results for exclusive photoproduction of @xmath109 @xcite and @xmath110 @xcite
we can conclude that thousands of @xmath111 mesons could be produced per year of data taking via the exclusive charge - exchange reactions @xmath112 and @xmath113 ( but the recoil nucleon can not be detected ) .
the energy of a virtual photon covers the range from about 20 gev and up to 180 gev .
the obtained data can be used to clarify the mechanism of the @xmath11 production and the role of the reggeized treatment at high energies . nevertheless , most of the @xmath111 mesons at such energies are produced non exclusively via the pomeron exchange mechanism .
such events could produce strong background under poor exclusivity control .
additional systematics could come from the process @xmath114 ( the cross section of this reaction is of the same order of magnitude @xcite ) because the compass setup is not able to reconstruct a decay of low - energy @xmath115 in a regular way .
nuclear effects in @xmath11 photoproduction off the lithium-6 , deuterium and nitrogen nuclei should also be taken into account in an appropriate way .
the forthcoming upgrade of the compass setup related to the planned data taking within the framework of the gpd program @xcite could provide better conditions for experimental study of the reaction @xmath8 and partially eliminate problems mentioned above .
the new 2.5 m long liquid hydrogen target surrounded by a 4 m long recoil proton detector will be used .
absence of the neutrons in the target will remove one exclusive production channel for @xmath116 .
the recoil proton detector serves the double purpose : to reconstruct and identify recoil protons via time - of - flight and energy loss measurements .
since the reaction @xmath8 does not have a recoil proton in the final state , any activity in the recoil proton detector could be used as the veto in the offline analysis .
in addition , the recoil proton detector will be able to detect the @xmath117 decay .
significant impact on the exclusivity control efficiency will be made by the planned upgrade of the electromagnetic calorimetry system .
within the framework of the effective lagrangian approach and regge model , the @xmath0 photoproduction from the proton is investigated .
the obtained numerical results indicate the following : * the total cross section @xmath1 related to the experimental data @xcite is well reproduced with reasonable value of @xmath92 . although the monopole form factor was used in the calculations , the dipole and exponential form factors were also tested .
it is found that the total cross section becomes less sensitive to the @xmath81 values when the reggeized treatment is added . *
the shapes of the differential cross section @xmath94 with the reggeized treatment are much different from that without the reggeized treatment at higher @xmath93 .
the reggeized treatment can lead to that the differential cross section @xmath94 decreases rapidly with the increasing @xmath93 , especially at higher energies . *
the differential cross section @xmath118 with the reggeized treatment is very sensitive to the angle @xmath75 and makes a considerable contribution at forward angles . *
the invariant mass distribution for the dalitz process @xmath119 shows an obvious peak at @xmath120 gev , which can be checked by further experiment . * for comparison
, it is found that the cross section of the @xmath121 process is almost the same as that of the @xmath1 reaction .
thus , the above theoretical results are valid to the @xmath121 channel . to sum up , we suggest testing our prediction for the cross section of the @xmath1 process at the compass facility at cern .
such a test could provide important information for clarifying the production mechanism of the @xmath0 and the role of the reggeized treatment at high energies .
nevertheless , the precise measurements near the threshold , where the difference between the predictions of the production models is maximal , are also important .
the author x. y. wang is grateful helmut haberzettl and jun he for useful discussions about regge theory . x. liu _
et al_. phys . rev .
d 77 094005 ( 2008 ) .
j. he , x. liu , phys . rev .
d 80 114007 ( 2009 ) .
et al_. phys .
d 89 034016 ( 2014 ) .
et al_. phys . rev .
d 92 034022 ( 2015 ) .
m. karliner , j. rosner , phys .
b 752 329 ( 2016 ) .
d. morrison , phys .
22 , 528 ( 1966 ) . | in this work , the photoproduction of @xmath0 off a proton target is investigated within an effective lagrangian approach and the regge model .
the theoretical result indicates that the shapes of the total and differential cross sections of the @xmath1 reaction within the feynman ( isobar ) model are much different from that of the reggeized treatment .
the obtained cross section is compared with the existing experimental results at low energies .
the @xmath0 production cross section at high energies can be tested by the compass experiment , which can provide important information for clarifying the role of the reggeized treatment at that energy range . |
given a graph @xmath0 , a * perfect dominating set * is a subset of vertices @xmath1 such that each vertex @xmath2 is dominated by exactly one vertex @xmath3 . an * efficient dominating set * is a perfect vertex dominating set @xmath4 where @xmath4 is also an independent set .
every graph @xmath5 contains a perfect dominating set , for instance , take @xmath6 .
but not every graph contains an efficient vertex dominating set .
these problems consists in searching the sets with minimum number of vertices .
all of them are np - hard , even when restricted to certain graph families .
the weighted version of these problems , where each vertex @xmath7 has a weight assigned @xmath8 , consists on finding a perfect vertex dominating set where the sum of the weights is minimum .
we denote these problems as minimum weighted perfect vertex domination ( mwpvd ) , minimum weighted efficient vertex domination ( mwevd ) .
we denote the edge - versions of these problems as minimum weight perfect edge domination ( mwped ) and minimum weight efficient edge domination ( mweed ) .
efficient edge dominating sets are also known as dominating induced matchings , and denoted as dim s .
note that for these edge - versions the dominating set consists of edges instead of vertices , hence the weights are on the edges , and the adjacency of two edges is defined as two edges that shares a vertex .
we say a _ pendant _ vertex ( also known as _ leaf _ ) is one whose degree is exactly one . in this paper
we show results for the weighted perfect domination problem , and for the efficient domination problem , restricted to circular - arc graphs .
the following definitions and results come from @xcite given a circular - arc model @xmath9 where @xmath10 , two points @xmath11 are equivalent if @xmath12 . the @xmath13 extreme points from the @xmath14 arcs of @xmath15
divide the circle @xmath16 in @xmath13 segments of the following types : ( i ) @xmath17 ( ii ) @xmath18 ( iii ) @xmath19 $ ] ( iv ) @xmath20 $ ] .
we say the segments of type ( i ) are _ intersection segments_. it is easy to see that all points inside one of the @xmath13 segments are equivalent .
@xcitethere are at most @xmath13 distinct @xmath21 .
@xcite[lemmalms ] given a ca model @xmath22 , if there are no two or three arcs of @xmath15 that covers the entire circle @xmath16 then @xmath23 is an hca model .
given a circular - arc model @xmath22 , the following algorithms can be achieved in @xmath24 time : * search a universal arc .
a universal arc has common intersection with every arc from @xmath15 .
* search two arcs @xmath25 such that @xmath26
. * search three arcs @xmath27 such that @xmath28 .
* search a point @xmath29 such that @xmath30 is maximum or minimum . *
if @xmath31 and @xmath32 and do not exists two arcs from @xmath15 that covers the entire circle @xmath16 then exists an induced cycle @xmath33 with @xmath34 such that arcs corresponding to the vertices cover the circle @xmath16 in @xmath23 and the rest of the arcs from @xmath15 are pairwise disjoint and are contained in exactly one of the arcs from @xmath33 .
thus each arc is either part of the @xmath33 or a leaf with a parent arc from @xmath33 .
it is possible to identify each arc from the @xmath33 .
note that @xmath35 if @xmath36 .
for instance , let @xmath37 , @xmath38 and @xmath39 .
hence @xmath40 .
we consider the four variants of the mentioned problems for circular - arc graphs , we propose efficient algorithms to solve them , except for the mwevd for which already exists a linear time algorithm .
we assume that our input is a circular - arc graph @xmath5 and a weight function @xmath41 over the vertices or edges depending on the problem we were solving . for simplicity
, we may use the circular - arc model from @xmath5 . for these cases
, there is an implicit previous step which is applying a linear time algorithm @xcite in order to obtain a circular - arc model from @xmath5 . for mwevd and mweed , we consider the function @xmath41 nonnegative .
the minimum weighted efficient vertex domination problem ( mwevd ) on a graph @xmath5 can be expressed as an instance of the minimum weighted dominating set problem ( mwds ) on the same graph @xmath5 making some minor adjustments on the way described in @xcite for unweighted version of mwevd .
there is an @xmath42 time algorithm to solve mwds for circular - arc graphs @xcite .
hence , mwevd can be solved for circular - arc graphs in linear time .
the minimum weighted efficient edge domination problem ( mweed ) for a graph @xmath5 is equivalent to either : 1 .
mwevd for the line graph @xmath43 2 .
mwis for the square of the line graph @xmath44 if @xmath5 is a circular - arc graph , then the graph @xmath45 is also a circular - arc graph .
the graph @xmath46 has exactly @xmath47 vertices and up to @xmath48 edges .
the algorithm from @xcite solves mwis problem in linear time for circular - arc graphs .
it is known that a graph that admits a dim is @xmath49-free .
this property imposes a bound on the number of edges in circular - arc graphs , thus the algorithm complexity bound can be improved from @xmath48 to @xmath50 .
[ lem : bound1 ] @xcite every @xmath49-free graph @xmath5 where @xmath51 such that @xmath5 is either chordal or interval graph has at most @xmath52 edges .
[ lem : bound2 ] every @xmath49-free graph @xmath5 where @xmath53 such that @xmath5 is circular - arc graph has at most @xmath13 edges . if @xmath54 is trivially true , and if @xmath5 is an interval graph the property is satisfied by lemma [ lem : bound1 ] . thus ,
suppose @xmath55 and @xmath5 is not an interval graph .
if exists @xmath56 such that @xmath57 then @xmath5 is an interval graph and if @xmath58 then @xmath5 contains a @xmath49 . in consequence ,
@xmath59 for any point @xmath56 .
let @xmath56 be the point with maximum value of @xmath60 .
clearly , @xmath56 belongs to an intersection segment @xmath17 where @xmath61 .
hence , @xmath62 and we can restrict the study to @xmath63 .
it is easy to see that at least one of these exists .
we analyze each possible case : * @xmath64 : let @xmath17 the intersection segment that contains @xmath56 : 1 .
@xmath65 : in this case @xmath66 where @xmath67 contains arc @xmath68 .
if we cut arc @xmath67 at point @xmath69 converting it in two different arcs @xmath70 and @xmath71 we get an interval graph @xmath46 with @xmath72 intervals . by lemma [ lem : bound1 ]
the number of edges of @xmath46 is at most 2(n+1)-3 = 2n-1 edges .
since @xmath46 has at least the same amount of edges that @xmath5 we conclude that @xmath73 .
@xmath74 : in this case @xmath75 .
if we replace @xmath68 by @xmath76 and @xmath77 by @xmath78 , we get an interval graph @xmath46 with @xmath14 vertices . however the edge that connects @xmath79 have been lost unless @xmath80 cover the entire circle
. therefore @xmath81 and @xmath82 .
* @xmath83 : + 1 .
@xmath84 in this case @xmath85 where @xmath86 contain arc @xmath68 .
if we cut arcs @xmath86 at point @xmath69 , converting each one into two different arcs , we get an interval graph @xmath46 with @xmath87 intervals and at least one more edge since @xmath86 have been converted to two different edges . by lemma [ lem : bound1 ]
the number of edges of @xmath46 is at most @xmath88 , hence @xmath89 2 .
@xmath74 : in this case @xmath90 where @xmath67 contains the segment @xmath17 .
we can replace @xmath68 by ( @xmath91 and @xmath77 by @xmath92 ( and may lose an edge unless @xmath80 covers the entire circle ) .
we cut the arc @xmath67 in either @xmath69 or @xmath93 , any one works .
we get an interval graph @xmath46 with @xmath72 vertices and at least @xmath94 edges . by lemma [ lem : bound1 ]
the number of edges of @xmath46 is at most @xmath95 .
hence @xmath89 .
the graph @xmath96 has 6 vertices and 12 edges , hence it shows tightness of the bound .
therefore any algorithm to solve dim problem can first check the amount of edges to ensure the existence of a dim . since @xmath97 then any @xmath42 time algorithm for circular - arc graphs can be easily converted to an @xmath24 time algorithm . a linear - time algorithm to solve mweed for general graphs given a fixed dominating set was presented in @xcite . if there exists a set of at most three arcs that covers the entire circle
, then there is a dominating set of size at most @xmath98 , thus the problem can be solved using the mentioned algorithm in linear time .
we assume a model @xmath23 without a set @xmath99 that covers the entire circle such that @xmath100 .
therefore @xmath23 is a helly circular - arc ( hca ) model and the original graph @xmath5 is an @xmath101 .
we analyze the model @xmath23 ( each case can be implemented in linear time ) : * ( a ) @xmath102 : * then @xmath5 is not @xmath49-free , hence it does not admit a dim * ( b ) @xmath103 : * then @xmath23 is an interval model .
algorithm from @xcite can be applied * ( c ) @xmath31 : * since @xmath104 , then @xmath105 . we can locate an induced cycle @xmath106 on @xmath5 such that arcs from those vertices covers the entire circle @xmath16 in the model @xmath23 , and the rest of arcs from @xmath107 are pairwise - disjoint and are contained in exactly one arc from @xmath33 .
if a vertex @xmath108 is connected with a set of pendant vertices @xmath109 , then @xmath7 should be colored with black in order to get a coloring of the graph that represents a valid dim .
note that the only edge that belongs to a minimum weighted dim will be an edge of minimum weight between @xmath7 and @xmath109 .
therefore , all edges between @xmath7 and @xmath110 such that its weight is not minimum can be removed from the graph @xmath5 in order to obtain @xmath46 .
every minimum weighted dim from @xmath5 is a valid minimum weighted dim in @xmath46 .
note that if exists @xmath111 edges with minimum weight , then any @xmath112 of those edges can be erased . in the new graph @xmath46
every vertex @xmath108 contains at most one pendant vertex @xmath113 in his neighborhood . since hca is an hereditary property
, @xmath46 is still an hca graph , and every clique from @xmath46 is a @xmath114 , therefore @xmath115 .
the graph @xmath116 is hca if @xmath46 is hca and it contains @xmath24 vertices and @xmath24 edges since maximum @xmath117 .
the @xmath116 model can be obtained within @xmath24 time @xcite .
therefore mweed can be solved for @xmath46 , by solving mwevd for @xmath118 , which can be done in linear time . *
( d ) @xmath119 : * denote @xmath120 . then @xmath121 form a triangle in @xmath5 , which means that any dim of @xmath5 must contain exactly one of the three edges of this triangle : @xmath122 . by removing point @xmath56 from @xmath16 ( every arc @xmath123 containing @xmath56
is now two disjoint arcs @xmath124 and @xmath125 ) , we obtain an interval graph @xmath126 , in which the vertices @xmath121 of @xmath5 correspond to two triangles formed by @xmath127 and @xmath128 , respectively .
so , any dim of @xmath126 must contain one edge of each of these triangles .
furthermore , our task would become simpler if @xmath127 and @xmath128 would correspond to triangles having neither common nor adjacent vertices .
the latter condition is fulfilled because there are no two arcs that covers the entire circle .
[ lt ] the triangles @xmath129 and @xmath130 have neither common nor adjacent vertices .
clearly , @xmath131 are six different vertices . hence there are no common vertex in both triangles .
suppose that @xmath132 is adjacent to @xmath133 , @xmath134 .
consider the following cases .
( a ) @xmath135 , : : which means @xmath68 covers the circle @xmath16 , that is absurd .
( b ) @xmath136 , : : in this case , @xmath68 and @xmath77 cover the circle which is a contradiction .
so , we now consider that the triangles @xmath127 and @xmath137 contain neither common nor adjacent vertices
. we will apply algorithm from @xcite to @xmath126 , which solves mweed for chordal graphs ( hence interval graphs ) .
the algorithm will be executed three times . in each of the applications ,
the graph @xmath126 remains the same , but the weighting changes , to @xmath138 , @xmath139 and @xmath140 , respectively . in order to avoid edges of triangles @xmath129 and @xmath141 incident to @xmath132 or @xmath142 to be included in the dim , in the weighting @xmath143 , @xmath144
, we assign to each of these edges a high weight , for instance more than twice the sum of all weights of the edges of @xmath5 .
all the other edges of @xmath126 remain the same as in the corresponding edges of @xmath5 .
if the value of the minimum weighted dim of @xmath126 is above that high weight assigned to the edges we want to avoid , then we know that @xmath5 contains no dim .
otherwise , we have solved our problem , and we only need to subtract the duplicated weight among the edges of the triangles @xmath127 and @xmath145 which is part of the solution for @xmath126 , but not for @xmath5 .
the following are the formal definitions of the weights .
let @xmath146 be the arcs containing @xmath56 .
assign to @xmath126 the weighting @xmath143 , @xmath144 , which defines a weight @xmath147 for each edge @xmath148 , as follows .
+ for @xmath149 and @xmath148 , + @xmath150 + + for @xmath149 and @xmath151 , + @xmath152 + + for @xmath153 , @xmath136 , @xmath154 and @xmath155 , + @xmath156 + + for @xmath157 and @xmath136 , + @xmath158 + + let @xmath159 , @xmath160 . then @xmath161 @xmath162 _ where _
@xmath163 let @xmath160 .
cut the circle at point @xmath56 and consider the model @xmath164 of the interval graph @xmath126 . for lemma
[ lt ] , @xmath129 and @xmath141 are triangles with neither common nor adjacent vertices .
suppose @xmath5 has a dim @xmath165 , weighted by @xmath166 , having total weight @xmath167 .
we know that exactly one edge of the triangle @xmath168 , say @xmath169 , belong to @xmath165
. then choose the weighting @xmath140 for @xmath126 and consider the following subset of edges @xmath170 .
@xmath171 + we claim the @xmath172 is a dim for @xmath126 .
since @xmath165 is a matching , @xmath172 is clearly so .
assume by contrary that it is not induced , and let the violating edge be @xmath173 , where @xmath174 are vertices incident to distinct edges of @xmath172 .
since @xmath165 is an induced matching of @xmath5 , it follows that @xmath175 and @xmath176 .
the latter means that @xmath127 and @xmath128 have a common or an adjacent vertex , a contradiction .
consequently , @xmath172 is indeed an induced matching .
it remains to show that it is dominating .
let @xmath99 be the subset of vertices of @xmath126 , not incident to the edges of @xmath172 .
let @xmath177 , @xmath178 . because @xmath165 is a dominating matching , the only possibility for @xmath174 to be adjacent is @xmath179 and @xmath180 , which contradicts @xmath181 not covering the circle .
then @xmath172 is a dim having weight equal to @xmath182 .
conversely , assume that @xmath126 has a dim weighted by , say @xmath140 , having weight @xmath183 . we know
that exactly one edge of each of the triangles @xmath129 and @xmath130 belong to @xmath172 . furthermore , the edges of these triangles which are incident either to @xmath184 or @xmath185 all have weight @xmath186 . therefore any dominating set of edges with weight @xmath187 contains the edges @xmath188 and @xmath189 .
let @xmath190 + we claim that @xmath165 is a dim of @xmath5 . clearly , for any @xmath191 , @xmath144 and @xmath192 , @xmath193 if and only if @xmath194 or @xmath195 .
consequently , @xmath172 being a dim of @xmath126 implies that @xmath165 is a dim of @xmath5 .
furthermore , the weight of @xmath165 is precisely the weight of @xmath172 less @xmath196 , since it was counted twice in @xmath172 .
the lemma follows .
an @xmath42 time algorithm to solve mwpvd for interval graphs was presented in @xcite .
the same paper shows the only known algorithm to solve mwpvd for circular - arc graphs in @xmath197 time .
note that any efficient vertex dominating set is also a perfect vertex dominating set . given a circular - arc graph @xmath5 , a circular - arc model @xmath23 from @xmath5
can be obtained in @xmath42 time and universal arcs can be identified in @xmath24 time . if a universal arc exists , we can solve the problem using the following procedure : [ approachmwpvd2 ] [ universalvertex ] given a graph @xmath198 where @xmath199 is a universal vertex : 1 .
if @xmath5 contains another universal vertex @xmath200 , then the unique perfect vertex dominating sets of @xmath5 are : @xmath201 , @xmath202 where @xmath132 may be any universal vertex .
if @xmath203 is the unique universal vertex from @xmath5 , then every perfect vertex dominating set from @xmath5 contains @xmath203 .
moreover , each solution can be computed on the following way : let @xmath204 the connected components from @xmath205 .
then every perfect vertex dominating set @xmath206 from @xmath5 should verify : 1 .
a vertex @xmath207 belongs to @xmath206 if and only if @xmath208 2 .
a subset @xmath209 that verifies ( a ) and contains @xmath203 is a perfect vertex dominating set from @xmath5 the proof is separated in two cases : 1 . if @xmath5 contains universal vertex @xmath200 : + it is clear that @xmath201 and each universal vertex @xmath202 are perfect vertex dominating sets from @xmath5 .
assume there exists another perfect vertex dominating set @xmath206 .
if @xmath210 then it should be a universal vertex and was one of the above mentioned sets .
hence @xmath206 contains at least two vertices and there is at one least vertex @xmath211 . since @xmath212
then every universal vertex @xmath132 should belong to @xmath206 , otherwise the vertex @xmath132 will be dominated by more than one vertex inside @xmath206 , which contradicts the definition of perfect vertex domination . moreover
, every vertex @xmath211 will be dominated by at least two universal vertices from @xmath206 , then again , this contradicts the definition of the set @xmath206 .
therefore , there is not any other perfect vertex dominating set .
2 . if @xmath203 is the unique universal vertex from @xmath5 : + assume exists a perfect vertex dominating set @xmath206 that does not contains @xmath203 .
it is clear that @xmath206 should have at least two vertices since there is not another universal vertex , but then the vertex @xmath203 is dominated by @xmath212 vertices , which is absurd .
thus every perfect vertex dominating set contains @xmath203 .
suppose @xmath206 is a perfect vertex dominating set such that @xmath213 , @xmath211 and @xmath214 .
since @xmath215 belongs to the same connected component @xmath216 , there exists a pair of adjacent vertices @xmath217 such that @xmath218 and @xmath219 . thus @xmath220 is dominated by @xmath203 and @xmath221 , which is a contradiction .
thus every perfect vertex dominating set satisfies ( a ) .
let @xmath206 be a subset of vertices that satisfies ( a ) and contains @xmath203 and @xmath211 .
it is easy to see that @xmath203 dominates @xmath113 .
the set of vertices @xmath222 belongs to the same connected component . by ( a )
, none of them are in @xmath206 .
therefore , @xmath203 is the unique vertex from @xmath206 that dominates @xmath113 if @xmath5 contains at least two universal vertices , then the amount of perfect vertex dominating sets is at most @xmath24 , and the one with minimum weight can be obtained in linear time . in case
@xmath5 contains exactly one universal vertex the minimum weighted perfect vertex dominating set can be obtained with the following procedure : @xmath223 .
we define the weight of a set of vertices as the sum of the weights of its vertices .
if the weight of a connected component @xmath216 of @xmath224 is negative , then @xmath225 .
thus we assume @xmath5 does not contain a universal vertex .
it is possible to solve mwevd in linear time for circular - arc graphs .
hence we can save the best efficient vertex dominating set as a candidate solution , and search for the perfect vertex dominating sets that are not efficient vertex dominating sets .
we determine in @xmath24 time the point @xmath56 such that @xmath30 is minimum , and according to this value a different approach can be used : * ( i ) @xmath226 , * in this case @xmath23 is an interval model and @xmath5 an interval graph .
a linear - time algorithm @xcite can be applied to solve mwpvd . *
( ii ) @xmath227 , * thus for every point @xmath228 , @xmath229 . let @xmath206 be a perfect vertex dominating set which is not an efficient vertex dominating set , and with minimum weight .
there should be two adjacent vertices @xmath230 , otherwise @xmath206 is an efficient vertex dominating set .
let @xmath231 the corresponding arcs .
note that @xmath232 , and let @xmath233 . for
any arc @xmath234 the vertex @xmath235 ( corresponding to @xmath236 ) should belong to @xmath206 , otherwise ( otherwise it will be dominated by two vertices ) .
let @xmath237 the first point from @xmath238 in clock - wise order such that @xmath239 ( hence @xmath238 and @xmath240 belong to different segments ) .
we will show that for any arc @xmath241 , the vertex @xmath235 corresponding to @xmath236 should be in @xmath206 we consider two cases according to segment finalization @xmath238 . * if it has open - end at @xmath242 , then @xmath243 .
thus , every arc @xmath244 , the vertex @xmath235 corresponding to @xmath245 . * if it has close - end at @xmath246 , then @xmath247 .
thus @xmath248 contains arcs from @xmath249 and an additional arc @xmath250 .
since @xmath250 is adjacent to arcs from @xmath249 and @xmath251 then @xmath250 should be in @xmath206 , since @xmath249 belongs to @xmath206 and @xmath250 could not have more than one adjacent from @xmath206 , unless is part of @xmath206 .
let @xmath252 and apply iteratively the same procedure @xmath13 times , this is the amount of different segments .
this shows that all vertices are in @xmath206 .
thus @xmath253 , and the solution will be the best among @xmath254 and the best solution previously found .
* ( iii ) @xmath255 , * in this case @xmath256 where @xmath257 corresponds to a vertex @xmath258 .
recall that @xmath15 has no universal arc .
we generate three interval models ( @xmath259 y @xmath260 ) from @xmath23 and replace @xmath261 with @xmath262 and @xmath263 , adding arcs for each one of them and modifying the weight function as we describe next : * ( @xmath264 ) * we replace @xmath261 with @xmath265 and @xmath266 .
we also add @xmath267 and @xmath268 , assigning weights to the vertices that corresponds to new arcs : @xmath269 , @xmath270 . * ( @xmath271 ) * we replace @xmath261 with @xmath265 y @xmath266 .
we also add @xmath267 , assigning weights to the vertices that corresponds to the new arcs : @xmath272 , @xmath273 . * ( @xmath260 ) * we replace @xmath261 with @xmath265 y @xmath266 .
we also add @xmath268 , assigning weights to the vertices that corresponds to the new arcs : @xmath272 , @xmath274 .
it is easy to see that described models are interval models .
hence mwpvd can be solved for those three interval models using a linear time algorithm from @xcite .
we show how to map the perfect vertex dominating set @xmath275 from @xmath276 , @xmath144 , to a perfect vertex dominating set @xmath206 from @xmath5 maintaining the weight .
* the weight of @xmath277 is bounded , thus @xmath278 .
since both are leaves , they should be dominated by their parents @xmath279 y @xmath280 , respectively . then @xmath281 . in this case , we take @xmath282 . note that the weight of @xmath206 and @xmath277 is the same since @xmath283 .
+ each vertex from @xmath284 such that @xmath285 is dominated by exactly one vertex from @xmath277 .
each dominating vertex remains except for @xmath279 and @xmath280 .
but vertices dominated by @xmath279 and @xmath280 are now dominated by @xmath7 .
hence @xmath206 is a perfect vertex dominating set from @xmath5 .
* the weight of @xmath286 is bounded , hence @xmath287 . since @xmath288 is a leave @xmath289 , then @xmath290 .
the vertex @xmath279 is dominated by @xmath288 , thus the rest of the neighbors of @xmath279 are not in @xmath286 .
+ in this case , let @xmath291 . note that @xmath206 and @xmath286 has the same weight since @xmath292 .
+ each vertex @xmath293 is dominated by exactly one vertex from @xmath206 .
all these vertices , except for @xmath7 were dominated by vertices from @xmath286 which remains at @xmath206 , except for @xmath288 , but this vertex dominates only @xmath279 which is not a vertex from @xmath5 .
@xmath280 is dominated by a vertex from @xmath291 .
clearly , this dominating vertex dominates @xmath7 and is the unique vertex that dominates @xmath7 . as a consequence
, @xmath206 is a perfect vertex dominating set from @xmath5 . *
the weight of @xmath294 is bounded , this case is symmetric to the previous case .
let @xmath295 be a perfect vertex dominating set from @xmath5 with the same weight of @xmath294 .
now the prefect vertex dominating set @xmath206 from @xmath5 should be maped to a perfect vertex dominating set with the same weight of @xmath206 in some model @xmath276 , @xmath296 .
the following describes the mapping : * if @xmath213 , @xmath297 in the model @xmath264
. then @xmath277 and @xmath206 has the same weight .
@xmath277 is a perfect vertex dominating set in @xmath264 , unless there exists a vertex @xmath298 such that is dominated by @xmath279 and @xmath280 . in this case , the arcs @xmath261 and @xmath236 ( arc corresponding to @xmath235 ) covers the entire circle @xmath16 . clearly , @xmath299 and is dominated by @xmath300 .
but @xmath206 is a perfect vertex dominating set from @xmath5 , thus @xmath7 is a universal vertex .
if this is not the case , then exists a vertex @xmath113 which is not adjacent to @xmath7 and the corresponding arc @xmath301 should be contained in @xmath302 , but since @xmath235 is dominated by @xmath7 , all the intersecting arcs with @xmath235 corresponds to vertices outside @xmath206
. then @xmath113 can not be dominated by any of the vertices from @xmath206 because every intersecting arc of @xmath301 intersects @xmath236 .
therefore @xmath7 is a universal vertex , but it contradicts the hipothesis that no universal vertices exists at @xmath5 .
hence , it does not exists the vertex @xmath235 and @xmath277 is a perfect vertex dominating set from @xmath303 . *
if @xmath285 , @xmath7 is dominated by @xmath304 and @xmath305 , where @xmath236 corresponds to vertex @xmath235 .
it is clear that @xmath306 since @xmath307 . in this case ,
again , @xmath206 and @xmath286 has the same weight .
if @xmath286 is not a perfect vertex dominating set from @xmath271 , then @xmath279 is dominated by @xmath309 and another vertex @xmath310 .
moreover , @xmath311 and dominates @xmath7 in @xmath5 but @xmath7 was dominated by @xmath304 .
hence @xmath312 and @xmath236 and @xmath261 covers the entire circle @xmath16 . applying the same reasoning of the previous case
we can conclude that @xmath235 is a universal vertex and contradicts the hypothesis that @xmath5 do not contain universal vertices .
therefore , @xmath286 is a perfect vertex dominating set in @xmath271 . *
if @xmath313 , @xmath7 is dominated by @xmath314 and @xmath315 .
it s symmetric to the previous case , we obtain that @xmath316 is a perfect vertex dominating set in @xmath260 with the same weight that @xmath206 .
we give an @xmath42 time algorithm to solve mwped for circular - arc graphs . to the best of our knowledge
there is no known polynomial time algorithm to solve this problem on circular - arc graphs .
it is proved in @xcite the np - completeness of unweighted version of this problem for bipartite graphs . given a graph @xmath198 and a perfect edge dominating set @xmath317 from @xmath5 ,
we denote @xmath318 the vertices incident to an edge from @xmath319 .
we can define the following 3-coloring for the vertices of @xmath5 : the black vertices @xmath320 \subseteq d\}$ ] , the gray vertices @xmath321 and the white vertices @xmath322 .
it is easy to see that for any 3-coloring of vertices that satisfies properties ( p1 ) and ( p2 ) @xmath325 is a perfect edge dominating set and for any 3-coloring of @xmath5 that satisfies ( p1 ) and ( p2 ) , if it contains @xmath326 , with @xmath327 , then vertices from @xmath326 should be black . given a circular - arc graph @xmath198 , we show how to solve mwped in linear time .
first , solve mweed in @xmath24 time . if there is a solution ( dim ) , we save the one with minimum weight as a candidate solution .
therefore the candidates that should be explored are the perfect edge dominating sets that are not dim .
note that the set @xmath328 is also a candidate solution .
it is clear that @xmath319 determines a 3-coloring of the vertices from @xmath201 that verifies ( p1 ) and ( p2 ) .
let @xmath7 and @xmath113 be the corresponding vertices to @xmath261 and @xmath301 .
the following possibilities form a valid color combination of @xmath7 and @xmath113 in order to satisfy ( p1 ) and ( p2 ) : ( black , black ) .
: : it is clear that any other arc @xmath236 from the model that corresponds to the vertex @xmath332 has common intersection with @xmath261 or @xmath301 . + using ( p2 )
, @xmath235 is not white .
hence , there is no white vertex in @xmath201 .
therefore , @xmath333 and @xmath334 , absurd .
this combination is not valid .
( black , gray ) and ( gray , black ) .
: : since both are symmetric , we consider one of them : ( black , gray ) . the vertex @xmath113 is gray and must verify ( p1 ) .
the only non - white neighbor is @xmath7 , hence the rest of his neighborhood should be white . by ( p2 )
the rest of neighbors of @xmath7 should not be white . then there are no common vertices of @xmath7 and @xmath113 . then for each vertex we know the color it must have , since every arc @xmath236 intersects with either @xmath261 or @xmath301 .
if there is no conflict with the colors each vertex have and verifies ( p1 ) and ( p2 ) , we can save this solution as one more candidate solution .
( gray , gray ) .
: : every vertex @xmath235 ( different from @xmath7 and @xmath113 ) should be adjacent to @xmath7 and/or @xmath113 . since @xmath7 and @xmath113 are gray adjacent vertices and verify ( p1 ) , @xmath235 is a white vertex .
thus @xmath319 contains the edge @xmath335 only , thus @xmath319 is a dim , which contradicts the election of @xmath319 .
therefore this color combination is not valid .
( gray , white ) and ( white , gray ) .
: : the vertex @xmath7 is gray and should verify ( p1 ) , hence it must have exactly one non - white neighbor @xmath336 . on the other hand , vertices from @xmath337
are gray because ( p2 ) .
we analyze the following two cases : + 1 .
@xmath338 : every neighbor from @xmath336 is also neighbor from @xmath7 , hence are all white .
vertices that are not neighbors from @xmath7 are neighbors from @xmath113 since their corresponding arcs are contained at @xmath301 and should induce a matching since they satisfy ( p2 ) .
it is easy to see that @xmath319 of this combination is exactly the edges from the induced matching plus the edge @xmath339 .
thus @xmath319 is a dim , which contradicts the election of @xmath319 . as a consequence ,
this combination is invalid for this case . 2 .
@xmath340 , then @xmath336 is a gray vertex ( p2 ) . by ( p1 ) vertices from @xmath341
are white .
vertices from @xmath337 should be gray ( p2 ) .
thus there are no black vertices in @xmath5 .
therefore @xmath319 is a dim .
the combination is invalid for this case .
+ both combinations are invalid according to the previous cases there are at most two extra solutions that should be compared to other candidate solutions .
one of this solutions is given by taking @xmath342 , and the other is a dim of minimum weight ) .
it is clear that @xmath319 determines a 3-coloring of @xmath201 that verifies ( p1 ) and ( p2 ) .
let @xmath345 the corresponding vertices to @xmath346 .
the possible combinations of colors that satisfies ( p1 ) and ( p2 ) are : ( black , black , black ) . : : in this combination , every arc @xmath250 should have non empty intersection with one of @xmath346 , then @xmath336 is not white . thus @xmath333 and @xmath342 .
this combination is not valid .
( white , gray , gray),(gray , white , gray),(gray , gray , white ) .
: : these three combinations are symmetric so we analyze one of them .
assume @xmath336 is a black vertex
. then @xmath347 by ( p2 ) . also @xmath348 and @xmath349 since @xmath113 and @xmath235 are gray adjacent vertices that should satisfy ( p1 ) .
thus @xmath250 can not be anywhere and @xmath336 is not a black vertex .
hence @xmath350 .
it is clear that @xmath319 is a dim and therefore this combination is invalid . *
( ii ) @xmath353 : * thus there is a @xmath354 formed by vertices that corresponds to arcs that belongs to @xmath21 , where @xmath355 .
let @xmath356 the vertices and @xmath357 the corresponding arcs .
the vertices @xmath356 must be black vertices for any 3-coloring that corresponds to a perfect edge dominating set . we can replace each arc @xmath358 with two arcs @xmath359 and @xmath360 . clearly , the resulting model @xmath361 is an interval model because the point @xmath56 do not belongs to any arc .
the intersection graph @xmath362 from @xmath361 is an interval graph where each vertex @xmath363 is replaced by two non - adjacent vertices @xmath364 y @xmath365 and the neighbors of @xmath366 in @xmath5 ends being either neighbors of @xmath364 or neighbors of @xmath365 in @xmath367 since @xmath23 do not contains two arcs that covers the entire circle @xmath16 .
in addition , the vertices @xmath368 induces a @xmath354 in @xmath367 and the vertices @xmath369 induces another @xmath354 in @xmath367 . we can change the weight function @xmath41 to @xmath370 for @xmath367 for each @xmath371 , in the following way : we will show that each perfect edge dominating set @xmath319 from @xmath5 corresponds to one perfect edge dominating set @xmath373 from @xmath367 with the same weight .
thus , the problem is reduced to solve mwped for interval graph @xmath367 with the linear time algorithm @xcite .
given a perfect edge dominating set @xmath319 from @xmath5 , we show how to generate @xmath373 .
let the 3-coloring of @xmath5 corresponding to @xmath319 , where @xmath374 are black vertices . in @xmath367
those vertices are replaced by @xmath375 and @xmath376 .
the new 3-coloring for @xmath367 consists on assign as the black vertices to @xmath375 and @xmath376 , while the original vertices maintain the same color .
this coloring must satisfy ( p1 ) and ( p2 ) in @xmath367 .
otherwise , assume ( p1 ) is not satisfied by a vertex @xmath113 .
clearly , @xmath377 and verifies ( p1 ) .
since white and gray vertices as their adjacencies remains in both graphs , then @xmath378 contains exactly one black vertex @xmath366 , @xmath379 in @xmath5 , and in @xmath367 it has two black neighbors @xmath364 and @xmath365 .
this is a contradiction since in this case @xmath380 covers the entire circle @xmath16 from @xmath23 . thus ( p1 ) is satisfied . on the other hand , every white vertex @xmath377 verifies ( p2 ) .
the vertices from @xmath337 are gray and @xmath337 remains the same at @xmath367 , hence it satisfies ( p2 ) .
thus the 3-coloring of @xmath367 corresponds to a perfect edge dominating set @xmath373 from @xmath367 .
the sets @xmath319 and @xmath373 have the same weight .
clearly , in the 3-coloring corresponding to @xmath373 , every vertex @xmath368 and @xmath369 are black vertices because are part of @xmath354 with @xmath381 . in order to generate the 3-coloring for @xmath5 ,
assign @xmath356 as black vertices and let the remaining vertices with the same color .
it can be verified in a similar way that this 3-coloring verifies ( p1 ) and ( p2 ) from @xmath5 , hence it corresponds to a perfect edge dominating set @xmath319 from @xmath5 .
moreover the map of @xmath319 and @xmath373 is the same as the previous one .
let @xmath121 and their corresponding arcs @xmath384 .
for every 3-coloring of @xmath5 that satisfies ( p1 ) and ( p2 ) , those three vertices must be : ( i ) three black vertices or ( ii ) exactly one of them white and the other two gray .
we add an arc @xmath385 to the model , and call it @xmath386 and @xmath387 as his corresponding graph .
it is easy to check that the 3-colorings of @xmath5 that satisfies ( p1 ) , ( p2 ) and * ( i ) * have a one - to - one correspondence with 3-colorings of @xmath387 that satisfies ( p1 ) and ( p2 ) since @xmath121 and the new vertex @xmath388 corresponding to the additional arc are black vertices since those four vertices form a @xmath49 and @xmath388 is not adjacent to any other vertex from @xmath387 . in order to preserve the weight between the perfect edge dominating sets among @xmath5 and @xmath387 , the new edges @xmath389 and @xmath390 must have weight 0 .
thus , mwped can be solved for @xmath387 where the model @xmath391 do not contain a set of two or three arcs that covers the entire circle @xmath16 , is not an interval model and @xmath392 , which can be solved using the previous case in linear time . an additional consideration for 3-colorings of @xmath5 that satisfies ( p1 ) , ( p2 ) and not ( i ) must be added . for the 3-colorings that satisfies ( p1 ) , ( p2 ) but not ( i ) , each arc @xmath393 can be replaced by two arcs @xmath359 and @xmath360 just as in ( ii ) , but instead of @xmath381 arcs , now is just for 3 arcs .
again , the model @xmath361 is an interval model and the intersection graph @xmath362 is an interval graph where each vertex @xmath394 is replaced by two non - adjacent vertices @xmath364 and @xmath365 .
a 3-coloring of @xmath367 can be obtained from the previous one , by keeping the same colors for common vertices among both graphs , and @xmath398 get colors from @xmath366 , @xmath399 .
the 3-coloring obtained corresponds to a perfect edge dominating set @xmath373 from @xmath367 .
if the 3-coloring of @xmath5 contains the vertex @xmath400 , @xmath401 with color white then the weight of @xmath319 using @xmath41 is the same weight that @xmath373 using @xmath402 .
thus , the mwped problem can be solved three times for the interval graph @xmath367 , each time with a different function @xmath402 , @xmath403 , applying linear algorithm from @xcite .
if the best of these solutions has bounded weight then the solution forms a perfect edge dominating set of minimum weight in @xmath5 that satisfies ( ii ) .
otherwise , there is no perfect edge dominating set of @xmath5 whose three - coloring satisfies ( ii ) .
the minimum perfect edge dominating set from @xmath5 is the minimum returned from the solutions given by @xcite for @xmath367 , the solution from @xmath387 , @xmath328 and the mweed for @xmath5 .
all of them can be obtained in linear time . *
( iv ) @xmath31 y @xmath404 : * since @xmath35 , then @xmath32 . in this case
, we can find the induced cycle @xmath33 with @xmath405 of @xmath5 such that the arcs corresponding to the vertices covers the circle @xmath16 in @xmath23 and the rest of arcs from @xmath15 are pairwise disjoint and are contained in exactly one of the arcs from @xmath33 .
the vertices from @xmath33 are @xmath406 , @xmath407 , @xmath408 , y @xmath409 , and the corresponding arcs of @xmath406 are @xmath410 and
the circular order according to the extreme is : @xmath411 . 1 .
@xmath5 is @xmath412 , thus @xmath413 . in this case
@xmath43 is exactly@xmath5 . instead of a weighted function over the arcs ,
the weighted function is over the vertices .
( note that @xmath414 is a circular - arc graph ) 2 .
there exists @xmath366 with leaves . without lost of generality , we assume @xmath415 and @xmath416 are their leaves , @xmath417 is the degree of @xmath418 . in this case , the vertex @xmath418 is father of @xmath419 , in any other 3-coloring corresponding to a perfect edge dominating set of @xmath5 , @xmath418 is not white since otherwise @xmath419 would be gray but could not have a non white neighbor to satisfy ( p1 )
. then @xmath418 must be black or gray
@xmath418 is black : then @xmath420 can not be all white , so @xmath421 can not be white .
but at the same time this set can not be gray because are adjacent only to @xmath418 , and has no white vertices to satisfy ( p1 )
. then @xmath416 are black vertices . to solve this subcase , we can alterate the model in the following way : add two identical arcs @xmath422 . in this new model ,
any leave @xmath423 , @xmath418 and the two new corresponding vertices to the additional arcs form a @xmath49 , thus in any 3-coloring , these vertices must be black .
but these two new vertices are adjacent only to @xmath418 and his leaves .
hence the remaining vertices can have any coloring independent of these new vertices .
therefore the perfect edge dominating sets from @xmath5 in this subcase corresponds one - to - one to the modified model .
the modified model do not contain a set of 2 or 3 arcs that cover the circle @xmath16 and there is a point @xmath56 such that is contained in 4 different arcs . algorithm ( ii )
can be used to solve it . by assigning weight 0 to the added arcs
then the solutions can be mapped and the weight will be the same .
@xmath418 is gray : then it contains exactly one non - white vertex which must be one of @xmath424 , and the rest are white vertices .
assume @xmath425 is the non - white vertex , then it is black because there is no other neighbor , thus this 3-coloring satisfies ( p1 ) and ( p2 ) .
a new 3-coloring can be obtained , swapping color of @xmath425 with the color of another leave @xmath423 .
the new coloring will hold properties ( p1 ) and ( p2 ) , hence it is convenient to choose the leave @xmath423 such that @xmath426 is minimum .
therefore there are 3 candidates to be considered as the non - white neighbor of @xmath418 .
these candidates are @xmath427 and the best @xmath423 . for each case we may alter the model , adding an arc that intersects only to @xmath428 and the arc corresponding to the non - white vertex chosen .
the new added edges have weight @xmath429 in order to prevent that in the new model the added vertex is chosen as the non - white vertex of @xmath418 . this forces to choose the same non - white vertex from the original model .
the new vertex has no other neighbors , hence it does not affect the coloring of the rest of the vertices .
the resulting model has a point @xmath56 which is contained by 3 different arcs , then the linear time algorithm ( iii ) can be used to solve mwped on this new model .
the algorithm must be run 3 times , once for each election of the non - white vertex for @xmath418 .
+ it applies at most once the algorithm from ( ii ) and three time the algorithm from ( iii ) , hence the algorithm time complexity is @xmath42 .
a. brandstdt , a. leitert , and d. rautenbach .
efficient dominating and edge dominating sets for graphs and hypergraphs . in _
isaac12 international symposium on algorithms and computation_. lncs 7676 , pp . 267277 , 2012 . | given a graph @xmath0 , a _ perfect dominating set _ is a subset of vertices @xmath1 such that each vertex @xmath2 is dominated by exactly one vertex @xmath3 .
an _ efficient dominating set _ is a perfect dominating set @xmath4 where @xmath4 is also an independent set .
these problems are usually posed in terms of edges instead of vertices .
both problems , either for the vertex or edge variant , remains np - hard , even when restricted to certain graphs families .
we study both variants of the problems for the circular - arc graphs , and show efficient algorithms for all of them . ....
.... efficient domination , perfect domination , circular - arc graphs |
considerable attention has been paid to the ability of intense laser fields to align or orient polar molecules ; in such fields , the molecules experience torque arising from the dipolar interaction with electric fields @xcite .
one approach to achieve adiabatic molecular alignment is to use a nanosecond - pulse laser @xcite . while adiabatic alignment disappears after the pulse is turned off , ultrashort pulses ( 100 fs or less ) can excite rotational wave packets of quantum rotors , thus yielding a noticeably aligned shape after the pulse is off @xcite .
furthermore , a specially designed sequence of pulses is known to achieve an enhanced angular focusing in quantum rotors @xcite ; this has been realized experimentally in optical lattices @xcite .
these types of alignment , which differ from the adiabatic , are important for manifold applications requiring transient molecular alignment under field - free conditions , such as the generation of laser pulses @xcite and the control of high harmonic generation as a source of coherent radiation @xcite . in most studies of the molecular alignment , dipolar interaction between polar molecules has been neglected . in this case ,
the rotational dynamics of polar molecules can be simply analyzed by using an isolated kicked - rotor model @xcite .
the quantum kicked rotor and its classical analog have long served as a paradigm for quantum and classical chaos . in contrast , the quantum dynamics of _ interacting _ kicked rotors has been less well studied so far , though intriguing phenomena of interacting quantum rotors are expected to emerge .
very recently , a study of the center - of - mass motion in two coupled kicked rotors has revealed that the decoherence effect induced by the internal degree of freedom enhances the quantum - classical correspondence in the dynamics of rotors @xcite .
in addition , anomalous dielectric responses have been pointed out in two coupled dipolar rotors @xcite .
these results imply that coupled kicked rotors , even in _ only two _ rotors , exhibit peculiar behaviors different from those of isolated kicked rotors . in the present work , we theoretically investigate the quantum dynamics of coupled ( paired ) kicked rotors subject to an ultrashort laser pulse ( @xmath0-function kick ) .
we find that the dipolar interaction between rotors remarkably enhances the transient orientation of paired rotors produced by a @xmath0-kick .
this enhancement of the orientation of _ quantum _ rotors contrasts with the results with the _ classical _ treatment for paired rotors ; in the latter case , the dipolar interaction inevitably hinders the orientation of coupled rotors .
furthermore , we demonstrate that the orientation of paired rotors can be further enhanced by applying the accumulative squeezing scheme proposed in ref . @xcite .
our findings enlighten the study of transient orientation of interacting polar molecules .
this paper is organized as follows .
section ii describes the hamiltonian of the paired - rotor system together with its analytical solutions for eigenenergies and their eigenfunctions .
the time development of paired rotors for _ post_-kicked times is given in this section .
section iii analyzes the time dependence of the orientation factor in both quantum and classical paired rotors .
the physical origin of the enhanced orientation in quantum rotors is discussed on the basis of the spatial profile of the probability density of paired - rotor wavefunctions .
section iv gives the conclusion .
the paper contains two appendices with the details of the calculations .
suppose that two rotors carrying dipole moments @xmath1 are arranged as shown in fig .
1 . for arrangement ( a ) , both rotors rotate in a plane , while in the case of ( b ) , both rotors belong to an identical rotation axis @xcite .
the hamiltonian for the system is given by @xmath2 , where @xmath3 is the hamiltonian for the @xmath4-th kicked rotor and @xmath5 represents dipolar interaction between rotors .
the term @xmath3 can be written as @xmath6 where @xmath7 is the angular momentum operator and @xmath8 is the moment of inertia of the rotor .
when rotors are driven by a linearly polarized field , we can set @xmath9 and interact via the dipolar interaction @xmath5 .
the direction of the electric field @xmath10 is defined as shown.,width=377 ] where @xmath11 is the field amplitude of short laser pulses .
the direction of the field is fixed as shown in fig . 1 , causing the angular focusing of paired rotors at @xmath12 . assuming that the rotational radius of rotors is sufficiently smaller than the separation @xmath13 between rotors , the dipolar interaction @xmath5 is expressed by @xmath14 \label{eq03}\ ] ] with the dielectric constant @xmath15 .
the vector @xmath16 connects two rotational centers as denoted in fig . 1 .
for simplicity , we rewrite the term ( [ eq03 ] ) as @xmath17 with the definition , @xmath18 the quantity @xmath19 determines the magnitude of the dipolar interaction between rotors , and plays a key role in the dynamics of coupled rotors , as we discuss later .
the explicit form of @xmath20 is obtained straightforwardly from fig . 1 .
the arrangement ( a ) gives the form , @xmath21 while , for arrangement ( b ) , we have @xmath22 in the absence of the field @xmath11 , eigenstates of paired rotors are analytically obtained by transforming variables into @xmath23 and @xmath24 . substituting them into eqs .
( [ eq01 ] ) and ( [ eq04 ] ) , the hamiltonian @xmath25 is separated as @xcite , @xmath26 @xmath27 here the quantity @xmath28 represents the kinetic energy .
the parameters @xmath29 equal @xmath30 for arrangement ( a ) and @xmath31 for ( b ) .
the separability of the hamiltonian @xmath25 allows us to write the paired - rotor wavefunction in the form @xmath32 .
consequently , the schrdinger equation of paired rotors @xmath33 can be decomposed into two independent eigenvalue equations expressed by @xmath34\ph_{\alpha } = 0 ; \;\ ; \alpha=\xi,\eta , \label{eq09}\ ] ] where @xmath35 and @xmath36 .
the solution of eq.([eq09 ] ) is given by the mathieu function @xcite , whose explicit forms are given in appendix a. the eigenenergies @xmath37 of paired rotors are thus expressed by @xmath38 .
let us consider the time development of the wavefunction in paired - rotor systems after a @xmath0-function kick at @xmath39 .
the wavefunction @xmath40 immediately after the kick is related to that just before the kick , @xmath41 , with a phase determined by @xmath42 \ps(\th_1,\th_2,\tau^{- } ) .
\label{eq10}\end{aligned}\ ] ] substituting the definition ( [ eq02 ] ) into eq .
( [ eq10 ] ) and transforming variables @xmath43 , we obtain @xmath44 where @xmath45 is the bessel function of @xmath46-th order . the quantity , @xmath47 represents the strength of the pulse . in actual calculations for eq .
( [ eq11 ] ) , the summation of @xmath46 can be truncated at the finite value @xmath48 , since the magnitude of the bessel function @xmath49 rapidly decays with increasing @xmath50 .
the time development of paired rotors for _ post_-kicked times is described by @xmath51 \ps(\xi,\eta,\tau^{+ } ) .
\label{eq13}\ ] ] the right - hand side of eq .
( [ eq13 ] ) can be analytically calculated by expanding the function @xmath52 by the mathieu function .
the details of the calculation are presented in appendix b. in the following , we set @xmath53 , and take @xmath54 and @xmath55 as units of energy and time , respectively .
the initial state just before the kick is fixed in the ground state .
in quantum paired rotors . the figures ( a ) and ( b ) correspond to arrangements ( a ) and ( b ) in fig . 1 .
the quantity @xmath55 is taken as the unit of time . left : the time dependence for @xmath56 within a long time scale .
the parameter @xmath57 is varied as @xmath58 ( dotted line ) and @xmath59 ( solid line ) fixing the kick strength @xmath60 .
right : the behavior of @xmath56 within a short time scale .
@xmath57 is varied as @xmath58 ( dotted line ) , @xmath61 ( dashed - dotted ) , @xmath62 ( dashed - dotted - dotted ) , and @xmath59 ( solid).,title="fig:",width=332 ] in quantum paired rotors . the figures ( a ) and ( b ) correspond to arrangements ( a ) and ( b ) in fig . 1 .
the quantity @xmath55 is taken as the unit of time .
left : the time dependence for @xmath56 within a long time scale .
the parameter @xmath57 is varied as @xmath58 ( dotted line ) and @xmath59 ( solid line ) fixing the kick strength @xmath60 .
right : the behavior of @xmath56 within a short time scale .
@xmath57 is varied as @xmath58 ( dotted line ) , @xmath61 ( dashed - dotted ) , @xmath62 ( dashed - dotted - dotted ) , and @xmath59 ( solid).,title="fig:",width=332 ] the degree of orientation for paired rotors is characterized by the orientation factor @xmath63 , where the angular bracket indicates to take the expectation value .
the factor @xmath56 tends to be zero when the orientation of the rotors in the field direction of @xmath64 is perfect . on the other hand ,
the factor equals @xmath65 when the amplitude of wavefunction @xmath66 is uniformly distributed in the @xmath67-@xmath68 space .
the strength of the dipolar interaction is characterized by the parameter @xmath69 , _
i.e. _ , the ratio of the interaction energy @xmath19 to the kinetic energy @xmath54 .
figure 2 ( a ) and ( b ) show the time dependence for the orientation factor @xmath56 for paired rotors fixing the kick strength @xmath60 .
indices ( a ) and ( b ) in fig .
2 correspond to arrangement of rotors ( a ) and ( b ) in fig . 1 .
in both ( a ) and ( b ) , the left figure plots the orientation factor for a long time scale , @xmath70 , while the right one does so for a short time scale , @xmath71 .
we first discuss the two figures on the left , where two values of @xmath57 are taken ; the dotted lines display the orientation factor for @xmath58 , and the solid lines display that for @xmath59 . for @xmath58 ,
the time dependence for @xmath56 for arrangement ( a ) is identical to that for ( b ) , because the two rotors are no longer correlated via dipolar interaction . in this case
, the orientation factor yields a simple form @xcite of @xmath72 with the lowest value of @xmath73 at the focal time @xmath74 . for finite @xmath57 s ,
the orientation factor exhibits somewhat complicated behavior different from that for isolated rotors .
the left two figures in fig .
2 exhibit the difference of the time dependence for @xmath56 between the case of @xmath59 and that of @xmath58 . for arrangement ( a )
, the magnitude of @xmath56 for @xmath59 exceeds that for @xmath58 at a time @xmath75 and @xmath76 .
we must notice that the lowest value of @xmath56 for @xmath59 located at @xmath77 is remarkably smaller than that for @xmath58 .
this indicates that , in arrangement ( a ) , the orientation of paired rotors is enhanced by introducing strong dipolar interaction . for arrangement ( b ) , in contrast , the magnitude of @xmath56 for @xmath59 does not exceed that for @xmath58 at any @xmath78 .
in addition , the lowest value of @xmath56 at @xmath79 seems to be invariant to the change of @xmath57 , implying that the orientation of paired rotors in arrangement ( b ) is not much affected by dipolar interaction . in order to examine the effect of the interaction for the lowest value of @xmath56
, we investigate in detail the behavior of @xmath56 around the focal time @xmath80 with varying @xmath57 .
the calculated results are shown in the right two images in fig . 2 , where the value of @xmath57 is increased from @xmath58 ( dotted line ) up to @xmath59 ( solid line ) .
in case ( a ) , the increase in @xmath57 monotonously reduces the lowest value of @xmath56 . for @xmath59 ,
the factor eventually takes the lowest value @xmath81 at the focal time @xmath82 , which is much smaller than the lowest value of @xmath56 for @xmath58 . in case
( b ) , on the other hand , the time dependence for @xmath56 hardly changes with varying @xmath57 . we thus conclude that , as far as arrangement ( a ) is concerned , the orientation of paired rotors can be efficiently enhanced by taking into account the dipolar interaction between rotors .
this is one of main findings of the present study .
as we see in the next subsection , the enhanced orientation in _ quantum _ rotors can not be interpreted from the _ classical _ dynamics for paired rotors , indicating that the enhanced orientation stems from a purely quantum effect . before proceeding to a further investigation of _ quantum _ paired rotors
, we consider the effect of dipolar interaction on the orientation for _ classical _ paired rotors . for classical kicked rotors ,
the term @xmath3 defined in ( [ eq02 ] ) is rewritten as @xmath83 while the interaction term @xmath84 is the same as that defined in ( [ eq04 ] ) . under field - free conditions ,
the equations of motion for arrangement ( a ) are given by @xmath85 \label{eq15a } , \\
i\ddot{\th}_2 & = & \frac{e_d}{2}\left [ 3 \sin(\th_1+\th_2 ) + \sin(\th_1-\th_2 ) \right ] \label{eq15b}.\end{aligned}\ ] ] by transforming the variables into @xmath86 and @xmath87 , we obtain the following equations : @xmath88 here we define the parameter @xmath89 , showing the strength of the dipolar interaction between classical rotors . for arrangement ( b ) , the same procedure yields @xmath90 solutions of eqs .
( [ eq16])-([eq19 ] ) are expressed by jacobi s elliptic functions . the orientation factor for classical paired rotors
is calculated by @xmath91 , where the bracket @xmath92 means averaging over initial angles @xmath93 @xcite . in classical paired rotors .
figures ( a ) and ( b ) correspond to arrangements ( a ) and ( b ) in fig . 1 . the quantity @xmath94 is taken as the unit of time .
the parameter @xmath89 , determining the strength of dipolar interaction , is varied as @xmath95 from the bottom ( solid line ) to the top ( dashed line).,width=313 ] figure 3 shows the time dependence of the orientation factor for the classical paired rotors .
quantities @xmath94 and @xmath96 are taken as the unit of time and the parameter @xmath97 , respectively .
the value of @xmath97 is incrementally increased from @xmath98 ( dotted line ) up to @xmath99 ( solid line ) as denoted in the figure caption .
we see that the increase in @xmath97 inevitably raises the minimal value of @xmath56 in both arrangements ( a ) and ( b ) .
this leads to the conclusion that , in _ classical _ systems , the strong dipolar interaction interferes with the orientation of paired rotors for both arrangements ( a ) and ( b ) .
the physical interpretation is given as follows .
when two dipolar rotors are assigned in arrangement ( a ) , strong dipolar interaction forces them to be parallel in the direction normal to the field direction ( see fig . 1 ) .
hence , the interaction hinders the orientation of rotors in the field direction . for arrangement ( b ) , on the other hand , two dipolar rotors tend to be anti - parallel to each other .
this increases the minimal orientation factor at the focal time . as a consequence ,
the dipolar interaction in _ classical _ systems certainly prevents the rotors from becoming oriented in the field direction defined in fig . 1 .
these facts naturally lead us to the following question : why is it that a strong dipolar interaction can enhance the orientation of rotors in _ quantum _ systems ( see fig . 2 ( a ) ) ? comparing the behavior of @xmath56 shown in fig . 3 ( a ) with that shown in the right of fig . 2 ( a ) , we clearly see that the effect of dipolar interaction on the orientation of paired rotors differs completely between classical and quantum systems . to settle this point
, we consider the time development of wavefunctions in the paired - rotor system as follows .
to understand the mechanism underlying the enhanced orientation in quantum paired rotors , we examine the time development of the probability density @xmath100 in the @xmath67-@xmath68 space .
figure 4 ( i ) and ( ii ) give contour plots of the probability density for arrangement ( a ) with @xmath59 . at @xmath101 ( fig . 4 ( i ) )
, the amplitude of @xmath100 is spatially confined around the two symmetric positions @xmath102 and @xmath103 .
when the kick is applied , these two wave packets move toward the origin @xmath12 while retaining their shapes , and finally collide head - on with each other at the origin at a focal time @xmath104 .
the resultant angular focusing is demonstrated in fig .
4 ( ii ) , where the probability density of @xmath105 is well localized at around the origin .
this transient angular focusing at the origin enhances the orientation of quantum paired rotors . for arrangement ( a ) in fig . 1
top : the parameter @xmath59 with ( i ) @xmath101 , and ( ii ) @xmath104 .
bottom : @xmath61 with ( iii ) @xmath101 , and ( iv ) @xmath104 .
white regions indicate large amplitudes of the probability density.,title="fig:",width=321 ] for arrangement ( a ) in fig . 1 .
top : the parameter @xmath59 with ( i ) @xmath101 , and ( ii ) @xmath104 .
bottom : @xmath61 with ( iii ) @xmath101 , and ( iv ) @xmath104 .
white regions indicate large amplitudes of the probability density.,title="fig:",width=321 ] we must note that angular focusing is enhanced only when @xmath106 , namely , when the two rotors are strongly correlated via dipolar interaction . if the dipolar interaction is sufficiently weak ( @xmath107 ) , the probability density for the initial state is broadly distributed in the @xmath67-@xmath68 plane ( fig .
4 ( iii ) ) . after a kick is applied , the amplitude of the wavefunction spreads out over the @xmath67-@xmath68 space , followed by the formation of a rainbow structure " @xcite at a focal time @xmath80 ( fig .
4 ( iv ) ) .
the degree of angular focusing for the rainbow structure is obviously inferior to that for the case of strongly interacting rotors ( fig .
4 ( ii ) ) . in summary ,
two factors are essential for enhancing the orientation of paired rotors : i ) the initial state before the kick consists of two wave packets strongly confined at symmetric positions with respect to the origin , ii ) these wave packets move translationally toward the origin after the kick . it should be mentioned that the translational motion of two wave packets after a @xmath0-pulse is not trivial .
we can analytically trace the motion of those wave packets by calculating the expansion coefficients @xmath108 appearing in appendix b. details of the calculations will be published elsewhere @xcite .
the spatial profile of the initial eigenstate is determined by the potential term @xmath5 . for arrangement ( a ) , the potential @xmath5 as a function of @xmath67 and @xmath68 gives two potential minima , at @xmath102 and @xmath103 , and a maximum at @xmath109 @xcite .
the energy difference between the minimum and the maximum is determined by the interaction energy @xmath19 or , equivalently , the parameter @xmath69 .
when @xmath57 is much larger than unity , the energy difference becomes so large that the initial eigenstate is strongly localized at the two potential minima , as shown in fig .
in addition , the amplitude of the wavefunction at the origin becomes almost zero due to the large potential maximum . for post - kicked time
, however , the kick creates a large number of excited states , so that a superposition of them can produce a transient angular focusing at the potential maximum @xmath109 .
this leads to a minimal orientation factor at a focal time . on the other hand , in the _ classical _ limit
, the orientation of paired rotors in the field direction @xmath109 can not occur when the energy difference between the minimum and the maximum is larger than the kinetic energy of the rotors immediately after the kick .
in other words , the strong dipolar interaction prevents the paired - rotor state from being located at the origin @xmath109 .
this follows that the enhanced orientation of paired rotors is a purely quantum phenomenon . for paired rotors kicked with a sequence of seven pulses of the strength @xmath60 ; the values of @xmath57
are varied as @xmath110 from the top ( the solid line ) to the bottom ( the dotted one ) .
the quantity @xmath55 is taken as unit of time . , width=309 ] the orientation of paired rotors can be further enhanced by applying the `` accumulative squeezing '' scheme proposed in ref.@xcite .
this strategy is based on a specially designed series of short laser pulses leading to a dramatic narrowing of the rotor angular distribution .
figure 5 shows the orientation factor @xmath56 of paired rotors kicked by a sequence of seven pulses of the strength @xmath60 .
the values of @xmath57 increase from @xmath111 ( solid ) to @xmath112 ( dotted ) . for both arrangement ,
the strategy works well to achieve the angular squeezing in paired rotors .
moreover , in arrangement ( a ) , a considerable reduction of the factor @xmath56 is seen for strongly interacting paired rotors with @xmath106 .
this result provides a new prospect for the scheme of multiple - pulse angular squeezing in interacting quantum rotors .
in conclusion , we have theoretically investigated the quantum dynamics of paired kicked rotors .
the orientation of paired rotors after the @xmath0-function kick is remarkably enhanced by introducing dipolar interaction between rotors , when the rotors are deposited in an identical plane .
the enhanced orientation is attributable mainly to two factors : i ) the initial state before the kick consists of two wave packets strongly confined at symmetric positions with respect to the origin in the @xmath67-@xmath68 space ; ii ) these wave packets move translationally toward the origin after the kick . we have also demonstrated that the orientation of quantum paired rotors can be further enhanced by applying a specially designed sequence of pulses .
our findings will stimulate experimental works aimed at the orientation of polar molecules correlated via dipolar interaction .
we acknowledge the support for this research by a grant in
aid for scientific research from the ministry of education , science , sports and culture , japan .
numerical calculations were performed on the hitachi sr8000 machine at the supercomputer center , issp , university of tokyo .
four kinds of mathieu functions , @xmath113 , @xmath114 , @xmath115 , and @xmath116 , can be expressed in terms of the fourier expansion as follows @xcite : @xmath117 by substituting eqs .
( [ eqa01]-[eqa04 ] ) into eq .
( [ eq09 ] ) , we obtain successive relations that determine the expansion coefficients . for @xmath118 , as an example , we obtain the following relation : @xmath119 the orthogonality of the mathieu functions is described by @xmath120
the explicit form of eq .
( [ eq13 ] ) can be obtained by expanding the state @xmath52 in terms of the mathieu functions . using these relations ,
the function ( [ eq13 ] ) is expanded as @xmath121 where each type of the mathieu function is abbreviated as @xmath122 the definition of @xmath123 is the same as that of @xmath124 .
the expansion coefficients @xmath125 are calculated straightforwardly as @xmath126 substituting eq .
( [ eqb01 ] ) into eq .
( [ eq13 ] ) , we obtain the explicit form of @xmath127 as @xmath128 \nonumber \\ & \times & d_{ll ' } f_l ( \xi , v_{\xi } ) g_{l'}(\eta , v_{\eta } ) .
\label{eqb05}\end{aligned}\ ] ] in actual calculations for eq .
( [ eqb05 ] ) , the double summation with respect to @xmath129 and @xmath130 can be truncated at a finite value , because the expansion coefficient @xmath108 rapidly decay with increasing @xmath129 and @xmath130 .
surface - mounted polar molecules having an in - plane dipolar moment and dipole rotors attached to a one - dimensional solid , respectively , are thought to be candidates for realizing arrangements ( a ) and ( b ) in fig . 1 . for arrangement ( a ) , the initial state of the rotors just before the @xmath0-kick is set to @xmath131 or @xmath132 with the same probability . for case ( b )
, @xmath133 can take any value with a uniform probability , while @xmath134 is determined by the relation @xmath135 .
the initial angular velocity of @xmath136 immediately after the kick is determined by @xmath137 in both ( a ) and ( b ) . | we point out that the non - adiabatic orientation of quantum rotors , produced by ultrashort laser pulses , is remarkably enhanced by introducing dipolar interaction between the rotors .
this enhanced orientation of quantum rotors is in contrast with the behavior of classical paired rotors , in which dipolar interactions prevent the orientation of the rotors .
we demonstrate also that a specially designed sequence of pulses can most efficiently enhances the orientation of quantum paired rotors . |
multiplex networks have become one of the most active area of recent network research @xcite mainly due to the fact that many real - world systems like public transport or social networks consists of many layers . after six years of studies we know much more about the structure and the function of multiplex networks @xcite .
a lot of attention has been devoted to the analysis of various dynamics on multiplex networks , including diffusion processes @xcite , epidemic spreading @xcite and voter dynamics @xcite .
however , there is still an open question how the number of levels influence the macroscopic properties of the system , such as dynamics of global variables , phase transitions or other emerging patterns . in order to find a general answer to this question systematic studies of various models
are needed .
however , one of the fundamental problems when dealing with models on the multiplex network is how to define the model on such a network , because usually there are several possibilities .
for example , in case of models defined by hamiltonians @xcite inter - layer interactions can be introduced by the relations between the coupling constants . however , the results of the model can strongly depend on this inter - layer interaction .
jang at al thoroughly analyzed the behavior of the ashkin - teller model in various types of inter - layer interaction , and presented a rich phase diagram containing three types of phase transitions
1st order , 2nd order , and mixed order . for the models defined by the transition probabilities , like the voter model
, the threshold model @xcite or the @xmath0-ising model @xcite one way of defining interactions between layers is the adoption of so called ` and ` and ` or ` rules , proposed for the first time by brummitt et al . in @xcite where the generalized threshold cascade model @xcite on multiplex networks has been studied .
authors introduced two kinds on nodes : ( 1 ) an ` or ` node was activated as soon as a sufficiently large fraction of its neighbors were active in at least one level , ( 2 ) an ` and ` node was activated only if in each and every layer a sufficiently large fraction of its neighbors were active .
the concept of ` and ` and ` or ` rules has been recently adopted in the @xmath0-voter model with independence @xcite . for systems consisting only ` and ` nodes the phase transition becomes discontinuous for @xmath10 if the number of layers @xmath11 , @xmath12 for duplex , whereas for a monoplex such behavior is observed for @xmath13 . the ising model has always played a very special role in the statistical physics .
for the original ising model a continuous phase transition is observed for both the regular lattices as well as monoplex complex networks @xcite . on the other hand ,
in the case of network of network topology suchecki and hoyst observed a discontinuous phase transition @xcite .
moreover , in @xcite it has been shown that a seemingly small modification of the kinetic ising model in which a randomly chosen spin interacts only with its @xmath0 neighbors leads to the surprising result on monoplex complete graph , i.e. a switch from a continuous to a discontinuous phase transition at @xmath10 . in this paper
we ask the same question that has been asked in @xcite , but this time within the @xmath0- neighbor ising model , namely `` how the additional level will introduce the type of the phase transition ? '' .
analogously , as in @xcite we focus on trivial topology duplex clique .
the experience gained from the @xmath0-voter model allows to predict that the switch from a continuous to a discontinuous phase transition will appear for a smaller value of @xmath0 in case of a duplex clique than for a monoplex .
such a result would be also expected from the theory of equilibrium phase transitions , if the additional layer could be treated analogously to the additional dimension @xcite .
however , we will show that our simple prediction fails in case of the @xmath0- neighbor ising model and results are exactly opposite . adding a second layer causes that only continuous phase transitions are observed for all values of @xmath0 . because for a monoplex clique a discontinuous phase transition is observed for @xmath2 and for a duplex cliques the phase transition is continuous
, we expect that there is a intermediate topology for which a switch from a discontinuous to a continuous phase transition appears .
therefore we investigate the q - neighbor ising model on a generalized topology of a partially duplex clique , parametrized by @xmath5 $ ] , which allows us to tune the network from monoplex ( @xmath6 ) to duplex ( @xmath7 ) .
such a generalized topology , in which a fraction @xmath8 of all nodes appear on both levels , allows to estimate a critical value of @xmath9 at which a switch from continuous to discontinuous phase transition is observed .
multiplex networks consist of distinct levels ( layers ) and the interconnections between levels are only between a node and its counterpart in the other layer ( i.e. the same node ) . * a duplex clique * is a particular case of a multiplex networks , which consists of two distinct levels , each of which is represented by a complete graph ( i.e. a clique ) of size @xmath14 . the same topology was consider to analyze the @xmath0-voter model with independence @xcite .
levels can represent two different communities ( e.g. facebook and school class ) and are composed of exactly the same people each node possesses a counterpart node in the second layer .
analogously as in @xcite , we assume that each node possesses the same state on each level . in the * partially duplex clique *
only the fraction @xmath8 of @xmath14 nodes have a counterpart in the other layer , remaining nodes belong only to one community ( layer ) .
this means that at each level we have @xmath15 duplex - type nodes and @xmath16 monoplex - type nodes , which means that in total the system consists of @xmath17 distinguishable nodes ; an example of such a topology is shown in fig .
the fraction @xmath8 of individuals who are active in both layers was introduced in @xcite .
it has been suggested that the parameter @xmath8 can be interpreted as _ interlayer connectivity _ or the degree of _ structural multiplexity _ of the system @xcite
. 0.3 cm
0.3 cm in @xcite we have modified the kinetic ising model with metropolis dynamics allowing each spin to interact with @xmath0 spins randomly chosen from the whole system , ( i.e complete graph ) . here
we consider a set of @xmath14 spin described by the binary variables @xmath18 , which are duplicated on the second level .
the algorithm of a single update of the @xmath0-neighbor ising model on a duplex clique consists of eight consecutive steps [ alg ] : 1 .
randomly choose a spin @xmath19 2 . from all neighbors of @xmath19
choose a subset of @xmath0 neighbors on the first level : @xmath20 3 .
calculate the value of the following function : @xmath21 and the value of the same function for the flipped spin , i.e. @xmath22 , finally calculate @xmath23 4 .
select randomly a real number @xmath24 $ ] and if @xmath25 $ ] then set flag @xmath26 else @xmath27 5 . from all neighbors of @xmath19 choose a subset of @xmath0 neighbors on the second level : @xmath28 6 .
calculate the value of the following function : @xmath29 and the value of the same function for the flipped spin , i.e. @xmath30 , finally calculate @xmath31 7 .
select randomly a real number @xmath32 $ ] and if @xmath33 $ ] then set flag @xmath34 else @xmath35 8 .
if @xmath26 and @xmath34 then flip the spin @xmath36 and its counterpart in the second layer as usual , a single time step consists of @xmath14 elementary updates , i.e. @xmath37 , which means that one time unit corresponds to the mean update time of a single individual . as an order parameter we choose magnetization : @xmath38 in a single update the number of spins up @xmath39 can change according to the following process : @xmath40 simultaneously with @xmath39 , magnetization @xmath41 increases / decreases by @xmath42 or remains constant with the above probabilities . to calculate transition probabilities @xmath43 and @xmath44 for @xmath45
it is convenient to use the concentration of up spins , which is related to the magnetization by the simple formula : @xmath46 the transition probabilities as a function of @xmath47 and model s parameters @xmath48 and @xmath0 have the following form : @xmath49 ^ 2\nonumber \\ \gamma^-(c , t , q ) = c \left[\sum_{k=0}^{k = q } { { q \choose k}(1-c)^{q - k}c^{k } e(q , k)}\right ] ^2\nonumber \\ \label{eq : gamma1}\end{aligned}\ ] ] where @xmath50.\ ] ] for the average values of concentration we can also write the rate equation @xcite , which has the following form in the rescaled time @xmath51 : @xmath52 in the stationary state @xmath53 , which is equivalent to the condition that the effective force @xcite : @xmath54 .critical temperature for duplex cliques for first values of @xmath0 . [ cols="<,^,^",options="header " , ] in fig . [ duplex_p ] we compare the results obtained from the monte carlo simulations and numerical solutions of eq .
( [ f0 ] ) [ recovered by plotting the implicit function @xmath55 .
both methods give consistent results and continuous phase transition is visible for all values of @xmath0 . in order to find analytically the value of the critical temperature we can use the method proposed in @xcite ,
namely we calculate @xmath48 for which the following condition is fulfilled : @xmath56 which in our case takes the form of @xmath57 for certain small values of @xmath0 it is possible to show @xmath58 in a compact form ( see table [ tab1 ] ) while for @xmath59 the solution can be well approximated by @xmath60 .
let us recall here that for the @xmath0-neighbor ising model on a monoplex clique the behavior is much richer : for @xmath61 and @xmath4 there is no phase transition , for @xmath1 the phase transition is continuous and for @xmath2 discontinuous @xcite .
moreover , for @xmath62 the hysteresis exhibits oscillatory behavior expanding for even values of @xmath0 and shrinking for odd values of @xmath0 .
it is worth to stress that the rule used here corresponds to the ` and ` dynamics @xcite which means that a change of a spin is possible only if both layers indicate the change . in case of the second rule ,
a so - called ` or ` dynamics , where for the spins s flip indication from only one level is sufficient , there is no phase transition for the @xmath0-neighbor ising model on a duplex clique , regardless of the value of @xmath0 .
as it has been already mentioned in sec . [ sec : net ] in the partially duplex clique only the fraction @xmath8 of @xmath14 nodes has a counterpart in the other layer , remaining nodes belong only to one level ( see fig . [ sch ] ) .
therefore for intermediate topologies , i.e. @xmath63 there are two types of nodes at each level : we have @xmath64 duplex - type nodes ( the state is the same on both levels ) and @xmath16 monoplex - type nodes .
the algorithm of a single update of the @xmath0-neighbor ising model on a partially duplex clique can be described as follows : 1 .
choose randomly a level ; the first or the second with equal probability @xmath65 .
2 . from @xmath14 spins on the selected level
choose randomly a single spin @xmath19 .
if the chosen spin belongs to the subset of duplex nodes , the algorithm looks the same as for the duplex clique .
if the spin belongs to the subset of monoplex nodes then : 1 . from all the neighbors of @xmath19
choose a subset of @xmath0 neighbors , @xmath66 ( monoplex nodes have neighbors only on one level ) .
2 . calculate the value of the following function for the original state of the @xmath67-th spin : @xmath68 and the value of the same function for the flipped @xmath67-th spin , i.e. @xmath69 .
3 . flip the @xmath67-th spin with probability @xmath70 $ ] , where @xmath71 we calculate separately concentration of ` up'-spins on a single layer for the monoplex nodes ( @xmath72 ) and duplex nodes ( @xmath73 ) : @xmath74 since the average magnetization on both layers is the same we can restrict ourselves to analysis of concentrations on one level . the transition probabilities for the monoplex nodes that @xmath75 are given by : @xmath76 and the corresponding rate equation in the rescaled time @xmath51 : @xmath77 the transition probabilities for the duplex nodes that @xmath78 are given by : @xmath79 ^ 2\nonumber \\ \beta_{d } ^-(c , c_d , t , q)= r c_d \left[\sum_{k=0}^{k = q } { { q \choose k}(1-c)^{q - k}c^{k } e(q , k)}\right ] ^2\nonumber \\\end{aligned}\ ] ] and the corresponding rate equation in the rescaled time @xmath51 : @xmath80 finally the total number of ` up ' spins in the single layer is equal to @xmath81 which after dividing by @xmath14 gives @xmath82 for a monoplex network the phase transition is not present for @xmath3 and @xmath4 .
we solve analytically eq .
( [ rr ] ) , for an arbitrary value of @xmath8 in case of @xmath3 and @xmath4 ; for @xmath84 we can find solutions numerically applying runge - kutta iv fourth - order method .
for @xmath3 we obtain the following relation between the critical temperature @xmath58 and @xmath8 ( see appendix [ app : q1 ] for details ) : @xmath85 phase transition appears for @xmath86 and it is continuous ( when @xmath87 we have @xmath88 and no phase transition ) .
for @xmath7 we have @xmath89 confirming the result from table [ tab1 ] .
0.3 cm 0.3 cm in the case of @xmath4 adding a fraction of duplex nodes causes the appearance of a phase transition for @xmath90 . for the range
@xmath91 we observe an interesting setting with seven real solutions ( see fig .
[ stq2_r486 ] ) .
the simplest one @xmath92 is stable ( attracting ) for the whole regime of temperature while next two ( @xmath93 ) are unstable ( repellent ) for @xmath94 $ ] .
we also have two stable ( @xmath95 ) and two unstable ( @xmath96 ) symmetric solutions which create shapes of two tubes . analyzing solutions for @xmath97 we observe an increase of the tube width : in fig .
[ q2_r486_495 ] we show specific solutions for @xmath98 and @xmath99 .
larger tube width corresponds directly to larger area of initial conditions which lead to stable solutions different than @xmath100 .
solutions @xmath100 and @xmath101 overlap for any value of @xmath102 , e.g. , @xmath103 lines cover @xmath99 ones in fig .
[ q2_r486_495 ] . finally for @xmath104 tubes cover the whole space of solutions ( see fig
. [ q2_an]a ) and @xmath105 solution is absorbed by the tube and becomes stable . for @xmath106
we observe a standard schema for a discontinuous phase transition .
the point where two unstable solutions disappear and @xmath100 becomes a stable one is clearly seen for @xmath107 in fig .
[ q2_an]b . with increasing @xmath8
this transition becomes weakly discontinuous stable solutions dominate while the unstable one is visible only for a very small range of @xmath8 ( see consecutive panels in fig .
[ q2_an]c - f ) . for further analysis
we define the width of the hysteresis as @xmath108 where @xmath109 is the temperature for which stable solutions @xmath110 disappear whereas @xmath111 is the temperature for which @xmath100 becomes a stable solution and two unstable solutions disappear .
similarly the jump of the order parameter @xmath41 is defined as the value of magnetization for stable solution @xmath112 .
figure [ dt_dm_q2 ] shows how the value of the jump of magnetization and the width of the hysteresis depend on @xmath8 .
both relations are decreasing and for @xmath7 both @xmath113 as well as @xmath114 are equal to zero .
these plots confirm the above presented results that the phase transition becomes continuous only for @xmath7 .
0.3 cm 0.3 cm 0.3 cm 0.3 cm 0.3 cm 0.3 cm in the case of @xmath1 the phase transition is continuous for all values of @xmath8 .
this observation is confirmed by two different methods : numerical solutions of eq .
( [ rr ] ) using runge - kutta method @xcite and monte carlo simulations of the model ( see left panel in fig .
[ q4s ] ) .
this is an expected result since for @xmath1 continuous phase transitions have already been observed for both monoplex @xcite and duplex cliques ( see fig .
[ duplex_p ] ) .
on the other hand we have different expectations for @xmath10 where for a monoplex clique the phase transition is discontinuous @xcite while for the duplex case its character is continuous ( fig .
[ duplex_p ] ) . in fig .
[ q4s ] ( right panel ) we show magnetization versus temperature : for @xmath116 the order parameter jumps , whereas it behaves continuously for @xmath104 .
although these results are obtained numerically , they are consistent with our expectations based on analytic results for monoplex and duplex cliques : as there is a first order phase transition for @xmath6 and continuous for @xmath7 there needs to be a threshold value of @xmath117 which separates these two regimes .
a natural question that emerges here can be formed in the following way `` for which value of @xmath118 the transition becomes continuous ? ''
the direct answer might be troublesome especially that one can expect a weak first order phase transition on a border between a first order and continuous phase transitions , analogously to the two - dimensional @xmath0-state potts models @xcite .
however we can measure simultaneously two quantities the jump of the order parameter and the width of the hysteresis ( defined as a distance between spinodal line @xmath119 and @xmath120 @xcite ) as a function of @xmath8 and then estimate @xmath118 . taking into account fine agreement between monte carlo results and numerical solutions of eq .
( [ rr ] ) by runge - kutta method we further use the second approach as being faster as well as more precise . to illustrate the phenomenon of the hysteresis width decreasing and moving we plot fig .
[ q6 ] where we present average magnetization @xmath121 for @xmath122 as a function of temperature @xmath48 for several values of inter - layer connectivity @xmath6 , @xmath123 , @xmath116 and @xmath124 .
we also compare a semi - analytic result for @xmath6 with numerical results obtain by runge - kutta method full compliance among the results is visible in fig .
[ q6 ] . to calculate the width of the hysteresis @xmath125 we measure two quantities : @xmath126 the temperature for which magnetization is equal zero when we start from ordered initial conditions @xmath127 and @xmath128 the temperature for which magnetization is equal zero when we start from random initial conditions @xmath129 ( see fig .
[ q4q6_t ] ) .
both quantities increase with @xmath8 , moreover the difference between them , i.e. , the width of the hysteresis @xmath125 , also decays with @xmath8 .
a more detailed analysis of the size of the hysteresis is show in the right panel of fig .
[ dm_q4q6 ] . in both cases
we observe a quadratic decay with @xmath8 until reaching @xmath130 ( for @xmath10 ) and @xmath131 ( for @xmath122 ) . after crossing these values
the width of hysteresis drops down to zero for @xmath132 and @xmath10 while it is equal to @xmath133 ( i.e. , the sampling step of @xmath48 ) for @xmath134 and @xmath122 . in the case of the jump of magnetization
( see left panel in fig .
[ dm_q4q6 ] ) we also observe a decrease in @xmath113 with growing @xmath8 until reaching @xmath132 and @xmath134 , respectively for @xmath10 and @xmath122 . to sum it up : for @xmath10 phase transition is discontinuous for @xmath135 $ ] , and @xmath136 $ ] for @xmath122 . after crossing these values
the transitions become continuous .
0.3 cm in the case of @xmath137 even for @xmath6 the transition is weakly discontinuous the width of the hysteresis equals to @xmath138 , decreasing very fast with @xmath8 and for @xmath139 dropping down to @xmath133 , i.e. , temperature sampling step .
the jump of the order parameter decays with @xmath8 as long as @xmath140 .
thus we conclude that for small values of @xmath8 we observe a weakly discontinuous phase transition which changes its character at @xmath141 becoming continuous .
it is essential to notice here that in this case the range of @xmath8 where the phase transition changes its character is very short .
therefore we need to stress that our result about the specific value of @xmath8 for which the we observe a switch of phase transition character should be treated as approximate .
all the above discussed observations regarding the character of phase transitions for @xmath142 are gathered in a @xmath143-space phase diagram presented in fig .
[ diagram ] where show different behavior of the model : discontinuous phase transition ( marked by ` dc ' ) , continuous phase transition , and a no transition phase .
the phase diagram underlines differences between even and odd values of @xmath0 .
in particular for odd values continuous phase transition dominates : for @xmath3 phase transition is continuous for all @xmath86 , in case of @xmath1 for the whole regime of @xmath8 while for @xmath137 we have discontinuous phase transition only for a small range of @xmath144 $ ] . in case of even values of @xmath0
the situation looks strikingly different .
firstly @xmath4 is a special case with a special phase transition occurring for @xmath145 , classic discontinuous phase for @xmath146 and finally continuous phase transition for @xmath7 . in the case of @xmath10 and @xmath122
we have discontinuous phase transitions for @xmath147 with @xmath148 for @xmath10 and @xmath149 for @xmath122 ; for @xmath150 the transition is continuous
. 0.3 cm
in this study we analyzed modified kinetic ising model @xcite on duplex clique . adding second level radically changes the behavior observed in the model whereas for a monoplex network model @xmath2 exhibits discontinuous phase transition , for duplex network we have continuous phase transitions for all value of @xmath0 .
the mechanism that leads to a discontinuous transition is usually related to fluctuations and discontinuous phase transition is usually observed above a certain critical dimension .
switch from a continuous to a discontinuous phase transition is observed also for larger number of states ( like in the potts model @xcite or in a model of tactical voting @xcite ) or even for a larger number of neighbors ( like in the @xmath0-voter model with independence on a complete graph @xcite ) . in
@xcite authors suggest that an additional level plays a similar role to an extra dimension and therefore an increase in the number of levels supports discontinuity . here for the @xmath0-neighbor ising model we observed opposite situation .
it might suggest that the role of another layer is not trivial and specific for a non - equilibrium model like @xmath0-ising and @xmath0-voter or threshold model and thus it can not be easily predicted on the basis of the observation of monoplex network behavior . analyzing the dynamics of the @xmath0-ising model on a partially duplex clique we observed very smooth change of the character of phase transition from discontinuous to continuous with increasing inter - layer connectivity @xmath8 .
similar like in the @xmath0-neighbor ising model on monoplex topology we observed different behavior for even and odd values of @xmath0 .
we need to underline that as the parameter @xmath8 is not discrete like in potts model @xcite , the continuous character of @xmath8 significantly impedes analysis of model by numerical methods . in this paper we deliberately focused on a basic topology of duplex cliques .
a possible and open future task is finding answers to the question _ how topology of network influence on phase diagram for q - ising model ? if we will also observed a changing character of phase transition from discontinuous to continuous one with adding second layer of network ? _
in the case of @xmath3 combining the set of equations ( [ rr ] ) with @xmath151 and @xmath152 leads to the following three solutions @xmath153 the nominator under the radical allows us to get the formula ( [ tc : q1 ] ) from the main text .
7 s. boccaletti , g. bianconi , r. criado , c.i .
del genio , j. gmez - gardees , m. romance , i. sendia - nadal , z. wang , and m. zanin , _ the structure and dynamics of multilayer networks _ , phys . rep . * 544 * , 1 - 122 ( 2014 ) .
m. kivel , a. arenas , m. barthelemy , j. p. gleeson , y. moreno , and m. a. porter , _ multilayer networks _ ,
journal of complex networks * 2 * , 203 ( 2014 ) .
f. battiston , v. nicosia , v. latora,_structural measures for multiplex networks _ ,
e * 89 * , 032804 ( 2014 ) . v. nicosia , v. latora , _ measuring and modeling correlations in multiplex networks_,phys .
e * 90 * , 032805 ( 2015 ) .
m. de domenico , v. nicosia , a. arenas , v. latora , _ structural reducibility of multilayer networks _ nature communications * 6 * 6864 ( 2015 ) .
s. gmez , a. daz - guilera , j. gmez - gardees , c. j. prez - vicente , y. moreno , and a. arenas , _ diffusion dynamics on multiplex networks _ ,
lett . * 110 * , 028701 ( 2013 ) . c. buono ,
l. g. alvarez - zuzek , p. a. macri , and l. a. braunstein , _ epidemics in partially overlapped multiplex networks _ , plos one * 9 * , e92200 ( 2014 ) . c. granell , s. gmez , and a. arenas , _
dynamical interplay between awareness and epidemic spreading in multiplex networks _
* 111 * , 128701 ( 2013 ) .
j. sanz , ch .- y .
xia , s. meloni , and y. moreno , _ dynamics of interacting diseases _ , phys .
x * 4 * , 041005 ( 2014 ) .
m. diakonova , m. san miguel , and v. eguluz , _ absorbing and shattered fragmentation transitions in multilayer coevolution _
e * 89 * , 06218 ( 2014 ) .
a. chmiel , k. sznajd - weron _ phase transitions in the @xmath0-voter model with independence on a duplex clique _ ,
e * 92 * , 052812 ( 2015 ) .
s. jang , j.s .
hwang , b .
khang _ ashkin - teller model and diverse opinion phase transitions on multiplex network _ ,
e * 92 * , 022110 ( 2015 ) . c. d. brummitt , k - m lee , and k .-
goh , _ multiplexity - facilitated cascades in network _ , phys . rev .
e * 85 * , 045102 ( 2012 ) .
d. j. watts , _ a simple model of global cascades on random networks _ , proc . natl .
usa * 99 * , 5766 ( 2002 ) .
lee , c. d. brummitt , and k .-
goh , _ threshold cascades with response heterogeneity in multiplex networks _ , phys .
e * 90 * , 062816 ( 2014 ) .
a. jdrzejewski , a. chmiel , k. sznajd - weron _ oscillating hysteresis in the @xmath0-neighbor ising model _ ,
e * 92 * , 052105 ( 2015 ) .
p. nyczka , k. sznajd - weron , _ anticonformity or independence ?
-insights from statistical physics _ , j stat phys * 151 * , 174 - 202 ( 2013 ) .
a. pekalski _ ising model on small world networks _ phys .
e * 64*,057104 ( 2001 ) .
a. aleksiejuk , j. a. hoyst , d.stauffer , _ ferromagnetic phase transition in barabasi albert networks _ , physica a * 310 * 260 266 ( 2002 ) s.n dorogovtsev , a.v .
goltsev , j.f.f mendes _ _ ising model on networks with an arbitrary distribution of connections__phys .
e * 66*,016104 ( 2002 ) .
k. suchecki , janusz a. hoyst _ bistable - monostable transition in the ising model on two connected complex networks _ phys .
e * 80*,031110 ( 2009 ) .
m. diakonova , v. nicosia , v. latora , m. sanmiguel,_irreducibility of multilayer network dynamics : the case of the voter model _
new j. phys * 18 * , 023010 ( 2016 ) .
v. spirin , p.l .
krapivsky , s. redner , _ freezing in ising ferromagnets _ , physical review e * 65 * , 016119 ( 2001 ) .
f. y. wu , _ the potts model _ , rev .
* 54 * , 235 ( 1982 ) .
n. a. m. araujo , j. s. andrade jr . , and h. j. herrmann , _ tactical voting in plurality elections _ , plos one * 5 * , e12446 ( 2010 ) .
p. nyczka , k. sznajd - weron , j. cislo , _ phase transitions in the q - voter model with two types of stochastic driving _ ,
e * 86 * , 011105 ( 2012 ) .
fernndez , j.j .
ruiz - lorenzo , m.p .
lombardo , a. tarancn , _ weak first order transitions . the two - dimensional potts model _ ,
b**4 * * 485 - 490 ( 1992 ) l. schlke and b. zheng , _ dynamic approach to weak first - order phase transitions _ ,
e * 62 * 7482 ( 2000 ) p. grassberger , c. christensen , g. bizhani , s. son nd m. paczuski , _ explosive percolation is continuous , but with unusual finite size behavior _
, prl * 106 * , 225701 ( 2011 ) . j. lee and j. m. kosterlitz _ finite - size scaling and monte carlo simulations of first - order phase transitions _ phys .
b 43 , number 4 3265 43 , ( 1991 ) k. binder__theory of first - order phase transitions__rep .
50 783 - 859 ( 1987 ) . | we analyze a modified kinetic ising model , so called @xmath0- neighbor ising model , with metropolis dynamics,[phys .
rev .
e * 92 * 052105 ] on a duplex clique and a partially duplex clique . in the @xmath0-ising model
each spin interacts only with @xmath0 spins randomly chosen from the whole neighborhood . in the case of a duplex clique the change of a spin is allowed only if both levels simultaneously induce this change .
due to the mean - field like nature of the model we are able to derive the analytic form of transition probabilities and solve the corresponding master equation .
the existence of the second level changes dramatically the character of the phase transition . in the case of the monoplex clique ,
the @xmath0-neighbor ising model exhibits continuous phase transition for @xmath1 , discontinuous phase transition for @xmath2 and for @xmath3 and @xmath4 the phase transition is not observed .
on the other hand , in the case of the duplex clique continuous phase transitions are observed for all values of @xmath0 , even for @xmath3 and @xmath4 .
subsequently we introduce a partially duplex clique , parametrized by @xmath5 $ ] , which allows us to tune the network from monoplex ( @xmath6 ) to duplex ( @xmath7 ) .
such a generalized topology , in which a fraction @xmath8 of all nodes appear on both levels , allows to estimate a critical value of @xmath9 at which a switch from continuous to discontinuous phase transition is observed . |
the ability to communicate secretly is considered one of the most important challenges of the information era @xcite . for all practical purposes most modern public key systems
can be considered secure @xcite . however , this security is not based on any mathematical proof but on the belief that there is no classical algorithm to factorize huge prime numbers in a reasonable amount of time .
some classical cyphers , such as the one - time pad @xcite , do not have the aforementioned problem .
they are private key protocols whose security is entirely based on a random string of bits , the key , which only the sender ( alice ) and the receiver ( bob ) should know .
once the key is secretly transmitted the communication is absolutely secure .
the drawback with these cyphers is that any key transmitted through a classical channel can be passively monitored .
although it may be technologically difficult to get the key without being noticed , it is in principle possible .
the solution of the key transmission problem based on the laws of physics was presented for the first time by bennett and brassard in a seminal work @xcite . making use of quantum channels ( polarized photons )
the authors theoretically showed the possibility to share a secret key with absolute security if the laws of quantum mechanics are correct .
they have shown that any interference of an eavesdropper ( eve ) on the quantum channel can be detected by alice and bob at the end of the protocol .
other interesting schemes were later proposed @xcite , in particular the e91 protocol which was the first qkd scheme that employed maximally entangled states @xcite . for an extensive review on other protocols and on experimental feasibilities
we refer the reader to ref .
@xcite . in this contribution
we present a new qkd scheme that uses directly non - maximally entangled states ( no entanglement concentration needed ) and the probabilistic quantum teleportation protocol ( pqt ) @xcite .
nevertheless , the present scheme resembles the bb84 @xcite rather than the e91 @xcite protocol , i.e. , although we make use of entanglement there is no need to check for violation of any bell inequality to assure the security of the shared key .
it is the probabilistic aspect of the pqt that guarantees the security of the teleported key and , as we show later , also allows it to be encoded in a set of orthogonal states .
note that in bb84-like protocols it is mandatory to encode the key in non - orthogonal states to make sure its transmission is secure .
in contrast to other protocols , where departure from maximal entanglement makes them inoperable , our scheme exploits partial entanglement and , using a special generalized bell measurement , ensures flawless key distribution .
furthermore , this new qkd scheme takes advantage of a new free parameter , namely the degree of partial entanglement , that enables more control over the security and transmission rate of the protocol .
indeed , as explained later , this freedom allows us to introduce a minor modification in the qkd scheme that turns it into a controlled qkd protocol , where a third party ( charlie ) has the final word on whether or not alice and bob are allowed to share a secret key , even after all steps of the protocol were implemented .
it is also worth mentioning that charlie decides whether or not alice and bob will share a key without ever knowing it , a feature that has practical applications .
we also show other possible interesting extensions of the basic protocol and how we can improve its security and the reliability of the transmitted key .
one important ingredient in this qkd protocol , and the one that allows it to depart from e91-like protocols , is the use of _ partially _ entangled states to transmit the secret key from one party to the other . indeed , as we will show , by playing with different kinds of partially entangled states and with different joint measurement basis , alice can teleport to bob a secret key .
the other ingredient is , as anticipated in the last sentence , the proper use of a probabilistic teleportation protocol , which allows us to harness the teleporting power of a non - maximally pure entangled state .
let us start by recalling the pqt as developed in ref .
@xcite and extended in ref . @xcite . as usual
, alice wants to teleport the following qubit to bob , @xmath0 where @xmath1 and @xmath2 are arbitrary complex numbers such that @xmath3 + @xmath4 @xmath5 @xmath6 .
contrary to the original proposal @xcite alice and bob now share a non - maximally entangled state , @xmath7 with @xmath8 and @xmath9 , which naturally leads to the following orthonormal basis , & = & ( + m ) / , [ phiplus ] + & = & ( m^ * - ) / , + & = & ( + m ) / , + & = & ( m^ * - ) / , [ psiminus ] with @xmath10 denoting the complex conjugate of @xmath11 and @xmath12 . using the generalized bell states ( gbs ) above
we can rewrite the three qubit state belonging to alice and bob as , @xmath13 @xmath5 @xmath14 @xmath15 @xmath16 @xmath5 @xmath17 @xmath18 @xmath19 @xmath20 @xmath21 @xmath20 @xmath22 @xmath23 @xmath24 @xmath25 @xmath20 @xmath26 @xmath27 @xmath20 @xmath28 @xmath20 @xmath29 @xmath30 @xmath20 @xmath31 , with the first two qubits being with alice and the last one with bob .
alice now proceeds by implementing a generalized bell measurement ( gbm ) , i.e. , she projects her two qubits onto one of the four gbs .
@xcite discusses three possible ways to experimentally implement a gbm . )
the probability to obtain a given gbs is , p_^+_m&=&(||^2+|mn|^2)/[(1+|m|^2)(1+|n|^2 ) ] , + p_^-_m&=&(|m|^2+|n|^2)/[(1+|m|^2)(1+|n|^2 ) ] , + p_^+_m&=&(|n|^2+|m|^2)/[(1+|m|^2)(1+|n|^2 ) ] , + p_^-_m&=&(|mn|^2+||^2)/[(1+|m|^2)(1+|n|^2 ) ] .
alice then sends bob the result of her measurement ( two bits ) via a classical channel who , whereupon , applies a unitary transformation on his qubit according to this information .
these transformations are the same ones given in the original teleportation protocol @xcite : if alice gets @xmath32 then bob does nothing , if she gets @xmath33 he applies a @xmath34 operation , if alice measures @xmath35 then bob applies @xmath36 and finally for @xmath37 he applies @xmath38 .
here , @xmath39 and @xmath40 are the usual pauli matrices ( @xmath41 @xmath5 @xmath42 and @xmath43 @xmath5 @xmath44 ) .
after the correct transformation bob s qubit is given by one of the following possibilities , & & = , [ a ] + & & = , [ b ] + & & = , [ c ] + & & = .
[ d ] from now on , and without loss of generality @xcite , we consider @xmath45 and @xmath11 to be real quantities . therefore , looking at eqs .
( [ b ] ) and ( [ c ] ) we realize that if @xmath46 the state with bob is @xmath47 and the protocol works perfectly .
there exist other possibilities , which come from eqs .
( [ a ] ) and ( [ d ] ) , namely , @xmath48 or @xmath49 .
but this is only possible if we have maximally entangled states since those relations imply @xmath50 . for @xmath46
the probability of success is simply p_suc = p_^-_n+p_^+_n=. [ probsuc ] therefore , if alice knows the entanglement of the channel , which increases monotonically with @xmath45 , she can match her measurement basis ( @xmath51 ) in order to make the protocol work with probability @xmath52 .
it is important to note that if @xmath53 bob obtains a different state and the protocol fails .
it is this property that we explore in order to build our qkd scheme .
let us assume that bob prepares with equal probability two partially entangled states , @xmath54 @xmath5 @xmath55 and @xmath56 @xmath5 @xmath57 , where @xmath58 and @xmath59 , @xmath60 .
for each state he keeps one qubit and send the other one to alice ( see fig .
[ fig1 ] ) .
[ fig1 ] both parties previously agreed on the values of @xmath61 and @xmath62 but at this stage bob does not tell alice the respective value of @xmath45 for each entangled state he prepared .
alice , on the other hand , prepares randomly two types of single qubit states , @xmath63 @xmath5 @xmath64 , which are to be associated with the secret key she wants to share with bob .
for example , the parties use the convention that @xmath65 represents the bit @xmath66 and that @xmath67 the bit @xmath6 . note that we do not need another encoding for the bits @xmath66 and @xmath6 that is non - orthogonal to the previous one , as required by the bb84 protocol . and @xmath6 .
this would increase the security of the present protocol on the expense of the key transmission rate .
however , as it is clear from the security analysis , this is not mandatory . ]
alice then uses each qubit received from bob to implement the pqt for each one of her randomly generated states @xmath63 . in doing so , she also chooses in a random way whether to project each pair of qubits ( hers and bob s ) to the gbs with @xmath68 or @xmath69 .
alice , however , does not inform bob of the value of @xmath11 but only which gbs she gets . at the end of this stage bob
knows what her measurement results were ( but not @xmath11 ) , which allows him to implement the right unitary operation on his qubits .
after that , each one of his qubits are described by one of the four states given by eqs .
( [ a])-([d ] ) , with @xmath70 .
the last six steps of the protocol are as follows .
first , bob projects his qubits onto the @xmath63 basis .
second , bob and alice reveal in a public channel the following information .
bob tells alice which value of @xmath45 ( @xmath61 or @xmath62 ) he has assigned to each partially entangled state whilst alice tells him the value of @xmath11 ( @xmath61 or @xmath62 ) for each gbm she made .
third , they keep all the cases where she has rightly matched the entanglement of the channel with the entanglement of the measuring basis , i.e. , whenever the shared entangled state was @xmath71 and alice chose @xmath72 , @xmath60 . fourth , they discard all the other cases since the pqt fails there ( @xmath73 ) .
fifth , within the cases where the matching condition is satisfied , bob and alice keep only those instances where her measurement results were @xmath74 or @xmath75 , the so called successful runs of the pqt .
for those , and only those runs of the protocol , alice and bob are sure they agree on the teleported state and consequently on the random string of zeros and ones .
finally , the last step consists of using half of the successful cases to test whether or not eve has tried to tamper with the key . in an idealized situation , i.e. , perfect detectors and no noise , they simply broadcast half of their valid results in a public classical channel and check if they always get the same bits .
if they do , the remaining half of bits are their secret key . if they fail to agree on the public data , they discard everything and repeat the whole protocol again .
however , noise and non - ideal detectors will introduce some errors even when all the steps of the protocol are successful .
nevertheless , alice and bob can still achieve any desired level of security by increasing the size of the shared key and employing classical reconciliation protocols and privacy amplification techniques already developed for other qkd schemes @xcite .
assuming an ideal scenario , for instance , excellent detectors and efficient measurement processes , we can calculate the maximum rate of how many teleported qubits constitute the final secret key .
we know that alice implements the pqt half of the times making a gbm with @xmath68 and half with @xmath69 .
therefore , the total probability of success for all pqt is , according to eq .
( [ probsuc ] ) , @xmath76 but half of the successful cases are discarded to check for the presence of eve , and the final rate becomes @xmath77 on the other hand , if we look at the bb84 protocol @xcite , we see that half of the times bob measures the qubits he receives from alice using the same basis she employed to prepare them and , within these successful runs , the other half is used to test for the security of the protocol .
this gives us , assuming no loss during the transmission of the qubit and ideal detectors , a total idealized rate of @xmath78 . returning to the protocol presented here , it is not difficult to see that @xmath79 , no matter what the values of @xmath61 and @xmath62 are .
( if they are equal to one we have @xmath78 but then the protocol is useless . )
however , for modest values of @xmath61 and @xmath62 ( a little greater than @xmath80 ) we get rates above @xmath81 .
if we allow one of them to approach unity we do even better . for example , if we have @xmath82 and @xmath83 we already obtain rates higher than @xmath84
. there exists , nevertheless , an important feature that we can easily achieve employing this protocol that is unattainable using the bb84 protocol
. we can transform it into a sort of controlled qkd scheme introducing another party ( charlie ) who can decide whether or not alice and bob are allowed to share a secret key even after they finished all steps of the protocol . in order to do that
, we let charlie prepare and distribute the entangled states @xmath85 and @xmath86 to alice and bob . hence
, if charlie publicly announces the values of @xmath45 for each entangled state he prepared he can make the protocol work without ever knowing the key .
otherwise , if he does not broadcast this information , the protocol ultimately fails .
note that the probability of success in this scenario , assuming charlie broadcast all the values of @xmath45 for each entangled pair he prepares , is the same we had before , eq .
( [ finalrate ] ) .
his role here is simply to distribute the entangled states between alice and bob , without changing the final success rate for the protocol .
we also remark that a similar third party control can be achieved using a different qkd scheme based on maximally entangled bell states @xcite . at this point
we wish to emphasize the main differences between the present scheme and the bb84 and the e91 protocol .
as described above , here we can achieve a level of third party control that is unattainable using the former two protocols .
this is an important and practical characteristic of this scheme that , as we show below , can also be extended to a fourth , fifth , , @xmath45-th party level of control .
moreover , in the present protocol the key is never transmitted from alice to bob as in the bb84 protocol .
rather , it is teleported from one party to the other , which gives an additional flexibility for this protocol in its third party formulation . indeed , once alice and bob have shared the partially entangled states distributed by charlie they can easily exchange their roles . instead of alice teleporting the key to bob ,
he is the one who teleports the key to her . also , contrary to the e91 protocol where a maximally entangled state is directly responsible to the generation of the secret key , here we use a non - maximally
entangled state as a channel through which the key is teleported .
in other words , the non - maximally entangled states of the present scheme have no direct role on the generation of the secret key .
the security of this protocol is based on the same premises of the bb84 protocol and , therefore , we can understand the security of the former by recalling the security analysis @xcite of the latter . the key ingredient here is the recognition that there are two unknown sets of actions throughout the implementation of the bb84 protocol that are only publicly revealed at the end of it : the basis in which alice prepared her qubits and the basis in which bob measured the qubits received from alice .
a similar thing happens for the present protocol .
we have two unknown sets of actions throughout each run of the protocol that are revealed only at its end : the entanglement of the shared qubits between alice and bob and what basis alice used to implement the gbm .
this lack of information prevents eve from always obtaining the right bit being sent from alice to bob without being noticed .
as we show below , the laws of quantum mechanics forbid eve from acquiring information about the key being transmitted without disturbing the quantum state carrying it if she does not know which entangled state is shared between alice and bob .
let us assume , for definiteness , that in one of the runs of the protocol alice prepared the state @xmath87 and that eve , somehow , replaced the entangled state produced by bob with one produced by her .
eve wants the state prepared by alice to be teleported to her . by doing
so she thinks she can obtain information on the key .
however , she does not know which gbm alice implemented to perform the pqt ( @xmath68 or @xmath69 ) .
this information is only revealed after bob confirms he measured his qubit .
she only knows that the measurement result of alice is , say , @xmath88 .
( this is the best scenario for eve . )
therefore , eve s qubit is described by eq .
( [ b ] ) , @xmath89 which can be written as , @xmath90 .
\label{stateeve}\ ] ] looking at eq .
( [ stateeve ] ) we see that unless eve guessed correctly alice s choice for @xmath11 ( and this only happens half of the times ) , preparing the right entangled state with @xmath46 , we have a superposition of the states @xmath65 and @xmath67 .
this implies that she obtains the wrong bit being transmitted with probability @xmath91 @xmath5 @xmath92 . in other words , since we have a superposition of the right and wrong answers quantum mechanics forbids eve from always getting the right one with a single measurement .
it is clear now that this is similar to the argument used to prove the security of the bb84 protocol .
hence , no matter what eve does , if she prepared the wrong entangled state and bob the right one , she will be caught trying to tamper with the key when alice and bob publicly compare part of it .
this is true since eve can not with certainty send bob another qubit which mimics the right one .
furthermore , eq . ( [ stateeve ] ) tells us that the greater the difference between @xmath61 and @xmath62 the more likely will eve be detected .
we can see this by noting that @xmath93 increases as a function of @xmath94 or , equivalently , as a function of the difference in entanglement between the channels .
lastly , the chances of eve being caught also increases with the size of the string of bits being publicly announced .
we can also estimate the optimal range of parameters ( @xmath61 and @xmath62 ) for this protocol , assuming we want to maximize the transmission rate while at the same time minimizing the chances of eve guessing the correct qubit being teleported .
in other words , we want to maximize a function proportional to @xmath95 , where @xmath93 is , as given above , the probability of eve guessing the wrong qubit and @xmath52 is eq .
( [ finalrate ] ) , the total rate of success in the transmission of the key .
both @xmath93 and @xmath52 are now considered functions of @xmath61 and @xmath62 .
a simple numerical analysis shows that the best strategies occur for @xmath96 and @xmath97 ( and vice - versa ) , while the worst cases for alice and bob occur when @xmath98 .
note that the security check outlined above is an idealization . in real - life situations we always have noise and imperfect devices that give wrong answers even in the absence of eve .
however , this can be controlled using classical reconciliation protocols and privacy amplification @xcite . a more detailed security analysis based on bounds for the mutual information between alice , bob , and eve using , for example , the techniques of refs .
@xcite , is beyond our goals here and is left for future work .
the qkd scheme presented here is very versatile and allows for arbitrary control over the protocol parameters .
this can be achieved by introducing two extensions , where one increases its security and the other increases the distance of reliable transmission of the key .
the security of the protocol is increased by allowing bob to generate more than two partially entangled states .
for example , instead of just creating the states @xmath99 , @xmath60 , he can create three or more states with different @xmath45 . with only two states ,
eve can guess the right gbm in half of the successful runs of the pqt .
however , with more entangled states , her chances are reduced to @xmath100 , where @xmath101 is the number of partially entangled states produced by bob . on the other hand , this increase in security reduces the transmission rate of the key since it becomes less likely that alice and bob achieve the matching condition ( @xmath51 ) .
to extend the distance of reliable key transmissions we can use quantum repeater @xcite stations . in this scenario
it is the first station ( the closest to alice ) that generates the partially entangled states and then publicize the values of @xmath45 , only after bob measures his qubits .
the other stations use maximally entangled states to successively teleport alice s qubit to bob . note that security increases if other repeater stations use partially entangled states too .
the repeater stations can also be used to extend the third party control described before to any number of parties . indeed , if we allow each station to freely choose its own partially entangled states we are increasing the number of parties that can decide whether or not alice and bob will share a secret key .
this is true for the protocol will work if , and only if , all the repeater stations disclose to alice and bob which partially entangled states they generated at each run of the protocol .
while noise and decoherence of entangled qubits usually result in mixed states @xcite , partially - entangled states used in the aforementioned qkd protocol can be considered in the scenario of coupling to a zero - temperature bath @xcite . in this regime dynamical control of decoherence @xcite
allows one to determine the amount of partial entanglement of the channel by properly tuning the relative decoherence between the qubits .
thus , the party sending the partially entangled states ( either bob in the standard qkd or charlie in the controlled version ) can select the degree of partial entanglement and is not restricted by the amount of noise in the system .
we showed that partially entangled states are useful resources for the construction of a direct qkd scheme by the proper use of probabilistic teleportation protocols .
this has an interesting implication on the practical implementation of entangled based qkd schemes , as it is extremely difficult to produce maximally entangled qubits . using the protocol presented here
, one can alleviate the experimental demands on the production of entangled pairs without rendering qkd inoperable .
furthermore , the present partially entangled state based qkd scheme is flexible enough that we were able extend it in at least three directions , each one augmenting its usability .
the first one turned the protocol into a controlled qkd scheme , where a third party decides whether or not alice and bob are able to share a secret key .
then we demonstrated how one can increase its security by letting the parties use more and more different partially entangled states to implement one of the steps of the qkd protocol , namely , the probabilistic teleportation protocol .
and third , we discussed how the use of quantum repeaters extends the distance of reliable key transmission without diminishing the key rate .
finally , the present qkd protocol naturally leads to new interesting questions .
for instance , can we increase the key transmission rate using partially entangled qudits instead of qubits ? is there any possible way of devising a similar approach using mixed entangled states ? or using continuous variable entangled systems ?
can we do better by using different types of entangled qubit - like states , such as the cluster state @xcite or the cluster - type coherent entangled states @xcite ?
it is our hope that the ideas presented here might lead to clues on how to answer these questions .
g. r. thanks the brazilian agency coordenao de aperfeioamento de pessoal de nvel superior ( capes ) for funding this research . c. h. bennett and g. brassard , _
quantum cryptography : public key distribution and coin tossing _ ( proceedings of ieee international conference on computers systems and signal processing : bangalore , india , 1984 ) | we show that non - maximally entangled states can be used to build a quantum key distribution ( qkd ) scheme where the key is probabilistically teleported from alice to bob .
this probabilistic aspect of the protocol ensures the security of the key without the need of non - orthogonal states to encode it , in contrast to other qkd schemes .
also , the security and key transmission rate of the present protocol is nearly equivalent to those of standard qkd schemes and these aspects can be controlled by properly harnessing the new free parameter in the present proposal , namely , the degree of partial entanglement .
furthermore , we discuss how to build a controlled qkd scheme , also based on partially entangled states , where a third party can decide whether or not alice and bob are allowed to share a key . quantum communication , quantum cryptography and communication security , entanglement production and manipulation 03.67.hk , 03.67.dd , 03.67.bg |
in the standard model ( sm ) , the charged weak interaction is described by the exchange of a @xmath6 boson with a pure vector coupling to only left - chirality fermions .
thus , in the low - energy four - fermion framework , the lorentz structure of the matrix element is predicted to be of the `` v - a@xmath7v - a '' type .
deviations from this behavior would indicate new physics and might be caused either by changes in the w - boson couplings or through interactions mediated by new gauge bosons .
leptonic decays such as @xmath8 ( unless specified otherwise , charge - conjugated decays are implied throughout the paper ) are the only ones in which the electroweak couplings can be probed without disturbance from the strong interaction .
this makes them an ideal system to study the lorentz structure of the charged weak current .
the most general , lorentz invariant , derivative - free and lepton - number - conserving four - lepton point interaction matrix element for this decay can be written as @xcite : @xmath9 \biggl[\bar{u}_m(\nu_{\tau})\gamma_n u_j(\tau^-)\biggr],\ ] ] @xmath10 the @xmath11 matrices define the properties of the two currents under a lorentz transformation with @xmath12 for scalar , vector and tensor interactions , respectively .
the indices @xmath13 and @xmath14 label the right- or left - handedness ( r , l ) of the charged leptons . for a given @xmath13 , @xmath14 and @xmath15 ,
the handedness of the neutrinos ( @xmath16 , @xmath17 ) is fixed .
ten non - trivial terms are characterized by ten complex coupling constants @xmath18 ; those with @xmath19 and @xmath20 are identically zero . in the sm , the only non - zero coupling constant is @xmath21 .
as the couplings can be complex , with arbitrary overall phase , there are 19 independent parameters .
the total strength of the weak interaction ( charged weak current sector ) is determined by the fermi constant @xmath22 , hence , the @xmath18 are normalized as : @xmath23 @xmath24 this constrains the coupling constants to be @xmath25 , @xmath26 and @xmath27 . in the case
where neutrinos are not detected and the spin of the outgoing charged lepton is not determined , only four michel parameters ( mp ) @xmath4 , @xmath5 , @xmath28 and @xmath29 are experimentally accessible .
they are bilinear combinations of the @xmath18 coupling constants @xcite : @xmath30 @xmath31 @xmath32 @xmath33 @xmath34 and appear in the predicted energy spectrum of the charged lepton . in the @xmath0 rest frame , neglecting radiative corrections , this spectrum is given by @xcite : @xmath35 @xmath36\biggr),\ ] ] @xmath37 where @xmath38 is @xmath0 polarization , and @xmath39 is the angle between the @xmath0 spin and the lepton momentum . in the sm , the `` v - a '' charged weak current is characterized by @xmath40 , @xmath41 , @xmath42 and @xmath43
measurement of @xmath28 and @xmath29 requires knowledge of the @xmath0 spin direction . in experiments at @xmath44 colliders with unpolarized @xmath45 beams ,
the average polarization of a single @xmath0 is zero .
however , spin - spin correlations between the @xmath46 and @xmath47 produced in the reaction @xmath48 can be exploited @xcite .
the main idea of our method is to consider events where both taus decay to selected final states .
one ( signal ) tau decays leptonically ( @xmath49 , @xmath50 ) while the opposite tau , which decays via @xmath51 , serves as a spin analyser .
we choose the @xmath52 decay mode because it has the largest branching fraction as well as properly studied dynamics @xcite . to write the total differential cross section for @xmath53 ( or , briefly , @xmath54 ) events , we follow the approach developed in refs .
the differential cross section of the @xmath55 reaction in the center - of - mass system ( c.m.s . ) is given by their formula @xcite : @xmath56 where @xmath57 , @xmath58 is the spin - spin correlation tensor , and @xmath59 is the polarisation vector of the @xmath60 in the @xmath60 rest frame ( unit vector along the @xmath60 spin direction ) .
the asterisk denotes a parameter measured in the associated @xmath0 rest frame .
the differential decay width of the signal is written in the form ( with the total normalization constant @xmath61 that is unimportant in this context ) : @xmath62 @xmath63 where the form factors @xmath64 and @xmath65 can be extracted from eq .
[ eqlepdec ] .
the @xmath66 decay width reads ( with the total normalization constant @xmath67 ) : @xmath68 @xmath69 @xmath70 @xmath71 where @xmath72 and @xmath73 are the momentum and solid angle of the @xmath4 meson in the @xmath0 rest frame ; @xmath74 and @xmath75 are the momentum and solid angle of the charged pion in the @xmath4 rest frame ; and @xmath76 is the pion form factor taken from ref .
the total differential cross section for @xmath54 events is : @xmath77 experimentally , we measure particle parameters in the c.m.s .
; hence , the visible differential cross section is given by @xcite : @xmath78 where the integration is performed over the unknown @xmath0 direction , which is constrained by the @xmath79 arc .
both @xmath80 and @xmath81 are calculated using parameters measured in the experiment .
the differential cross section is used to construct the probability density function ( p.d.f . ) for the measurement vector @xmath82 : @xmath83 @xmath84 @xmath85 where the form factors @xmath86 are calculated for each event and the five normalisation constants @xmath87 are evaluated using a monte carlo ( mc ) simulated sample .
there are several corrections that must be incorporated in the procedure to take into account the real experimental situation .
physics corrections include electroweak higher - order corrections to the @xmath88 cross section @xcite , the effect of the radiative leptonic decay @xmath89 @xcite , and the effect of the radiative hadronic decay @xmath90 @xcite .
apparatus corrections include the effect of the finite detection efficiency and resolution , the effect of the external bremsstrahlung for @xmath91 events , and the @xmath45 beam energy spread .
the method described is used for a precise measurement of michel parameters in @xmath54 events .
this analysis is based on a @xmath92 data sample that contains 446 @xmath93 pairs , collected with the belle detector at the kekb energy - asymmetric @xmath44 ( 3.5 on 8 gev ) collider @xcite operating at the @xmath94 resonance .
the belle detector is a large - solid - angle magnetic spectrometer that consists of a silicon vertex detector ( svd ) , a 50-layer central drift chamber ( cdc ) , an array of aerogel threshold cherenkov counters ( acc ) , a barrel - like arrangement of time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter ( ecl ) comprised of csi(tl ) crystals located inside a superconducting solenoid coil that provides a 1.5 t magnetic field .
an iron flux - return located outside the coil is instrumented to detect @xmath95 mesons and to identify muons ( klm ) .
two inner detector configurations are used in this analysis .
a beampipe with a radius of 2.0 cm and a 3-layer silicon vertex detector are used for the first sample of @xmath96 @xmath97 pairs , while a 1.5 cm beampipe , a 4-layer silicon detector and a small - cell inner drift chamber are used to record the remaining @xmath98 @xmath97 pairs @xcite .
the detector is described in detail elsewhere @xcite .
this analysis is based on events with one @xmath0 decaying to leptons @xmath99 and the other decaying via the hadronic channel @xmath51 . the selection process , which is designed to suppress background while retaining a high efficiency for the decays under study , proceeds in two stages .
\1 ) the first - stage criteria suppress beam background to a negligible level and reject most of the background from other physical processes : * there should be exactly two tracks extrapolated to the interaction point within @xmath100 cm in the transverse direction and @xmath101 cm along the beam and having a transverse momentum in the c.m.s .
@xmath102 gev/@xmath103 and a net charge of zero . * the sum of the track absolute momenta in the c.m.s . must satisfy @xmath104 gev/@xmath103 . *
the maximum value of the transverse momentum for all tracks in the laboratory frame should satisfy @xmath105 gev/@xmath103 .
* the maximum opening angle @xmath106 between tracks should exceed @xmath107 . *
the number of photons @xmath108 with c.m.s . energy @xmath109
mev should be five or fewer .
* the total ecl energy deposition in the laboratory frame should satisfy @xmath110 gev .
* the total energy of additional photons in the laboratory frame should be @xmath111 gev .
* the missing mass should lie in the range @xmath112 gev/@xmath113 gev/@xmath114 . *
the polar angle of the missing momentum in the c.m.s . should satisfy @xmath115 .
the last two criteria are especially effective in suppressing background from radiative bhabha , @xmath116 and two - photon processes .
\2 ) in the second stage , the @xmath117 ( @xmath118 , @xmath119 ) samples are selected from the remaining events . to select electrons , a likelihood ratio cut @xmath120 is applied , where the electron likelihood function @xmath121 and the non - electron function @xmath122 include information on the specific ionization ( @xmath123 ) measurement by the cdc , the ratio of the cluster energy in the ecl to the track momentum measured in the cdc , the transverse ecl shower shape and the light yield in the acc @xcite .
the efficiency of this cut for electrons is @xmath124 . to select muons ,
the likelihood ratio cut @xmath125 is applied .
it has an @xmath126 efficiency for muons .
each of the muon(@xmath127 ) , pion(@xmath128 ) and kaon(@xmath129 ) likelihood functions is evaluated from two variables : the difference between the range calculated from the momentum of the particle and the range measured by klm and the @xmath130 of the klm hits with respect to the extrapolated track @xcite . to separate pions from kaons ,
we determine for each track the pion ( @xmath128 ) and kaon ( @xmath129 ) likelihoods from the acc response , the @xmath123 measurement in the cdc and the tof flight - time measurement , and form a likelihood ratio @xmath131 to separate pions and kaons . for pions , we require @xmath132 , which provides a pion identification efficiency of about @xmath133 while keeping the pion fake rate at the @xmath134 level . finally , we select events with only one lepton @xmath135 ( @xmath136 ) , one charged pion @xmath137 and one @xmath138 candidate .
a @xmath138 meson is reconstructed from a pair of gammas with an energy in the laboratory frame of @xmath139 mev and the @xmath140 invariant mass in the range @xmath141 mev/@xmath142 mev/@xmath114 .
the absolute value of the @xmath138 momentum in the c.m.s . must satisfy @xmath143 gev/@xmath103 .
the invariant mass of the @xmath144 system must lie in the range @xmath145 gev/@xmath146 gev/@xmath114 .
the opening angles between a lepton and charged pion and between a lepton and @xmath138 should exceed @xmath147 . to avoid the uncertainty due to the simulation of low energy fake ecl clusters , we allow additional photons in an event with the total energy in the laboratory frame of @xmath148 gev . to evaluate the background and calculate efficiencies , a monte carlo sample of @xmath149 pairs is produced with the kkmc / tauola generators @xcite .
the detector response is simulated by a geant3-based program @xcite .
the detection efficiencies for signal events are @xmath150 and @xmath151 .
it is found that the dominant background comes from other @xmath0 decays ; a contribution from non-@xmath152 processes is very small less than @xmath153 .
the dominant background arises from @xmath154 ( or , briefly , @xmath155 ) events , where the second @xmath138 is lost .
its contribution is @xmath156% for the @xmath91 and @xmath157% for the @xmath158 events . for the @xmath158 events , an additional background at the level of @xmath159% originates from @xmath160 ( or , briefly , @xmath161 ) events , where a pion is misidentified as a muon .
the remaining background comes from other @xmath0 decays ; its contribution is @xmath162% for @xmath91 events and @xmath163% for @xmath158 events . the main background processes , @xmath155 and @xmath161 , are included in the p.d.f .
analytically while the remaining background is taken into account using the mc - based approach @xcite .
the total p.d.f .
is written as : @xmath164 @xmath165 where @xmath166 , @xmath167 and @xmath168 are the cross sections for the @xmath54 , @xmath155 and @xmath161 events , respectively ; @xmath169 is the detection efficiency for signal events in the nine - dimensional phase space ; and @xmath170 is the average signal detection efficiency .
after all selections , about 5.5 million events in all four configurations ( @xmath171 , @xmath172 , @xmath173 , @xmath174 ) are selected for the fit .
events : @xmath175 momentum ( upper left ) and polar angle ( upper right ) in the c.m.s . , @xmath176 invariant mass ( lower left ) , extra gamma energy in laboratory frame ( lower right ) .
open histograms - signal mc simulation , yellow shaded histograms - the main background components from the @xmath177 events , green shaded histograms - the remaining background , points with errors - experimental data .
blue arrows show applied selections .
mc and experimental histograms are normalized to the same number of events.,title="fig:",scaledwidth=48.0% ] events : @xmath175 momentum ( upper left ) and polar angle ( upper right ) in the c.m.s . , @xmath176 invariant mass ( lower left ) , extra gamma energy in laboratory frame ( lower right ) .
open histograms - signal mc simulation , yellow shaded histograms - the main background components from the @xmath177 events , green shaded histograms - the remaining background , points with errors - experimental data .
blue arrows show applied selections .
mc and experimental histograms are normalized to the same number of events.,title="fig:",scaledwidth=48.0% ] + events : @xmath175 momentum ( upper left ) and polar angle ( upper right ) in the c.m.s . , @xmath176 invariant mass ( lower left ) , extra gamma energy in laboratory frame ( lower right ) . open histograms - signal mc simulation , yellow shaded histograms - the main background components from the @xmath177 events , green shaded histograms - the remaining background , points with errors - experimental data .
blue arrows show applied selections .
mc and experimental histograms are normalized to the same number of events.,title="fig:",scaledwidth=48.0% ] events : @xmath175 momentum ( upper left ) and polar angle ( upper right ) in the c.m.s .
, @xmath176 invariant mass ( lower left ) , extra gamma energy in laboratory frame ( lower right ) . open histograms - signal mc simulation , yellow shaded histograms - the main background components from the @xmath177 events , green shaded histograms - the remaining background , points with errors - experimental data .
blue arrows show applied selections .
mc and experimental histograms are normalized to the same number of events.,title="fig:",scaledwidth=48.0% ] + figure [ selpic ] shows the distributions of selected kinematical parameters for @xmath171 events .
clearly , the experimental electron momentum spectrum is shifted markedly to higher momenta in comparison with the mc one .
this is an artifact of the strong nonuniformity of the experimental trigger efficiency , which is not properly simulated .
a special procedure has been developed to evaluate the trigger efficiency corrections , @xmath178 ; see fig .
[ trgeff ] .
data samples as a function of electron momentum in the laboratory frame - black points with errors .
it is fitted by an empirical function shown by the red solid line.,title="fig:",scaledwidth=48.0% ] data samples as a function of electron momentum in the laboratory frame - black points with errors .
it is fitted by an empirical function shown by the red solid line.,title="fig:",scaledwidth=48.0% ] + the trigger efficiency correction as well as the lepton identification efficiency correction , @xmath179 , are incorporated in the fitter by modifying the detection efficiency in eq .
[ allbgs ] : @xmath180 .
the result of the fit of the @xmath181 experimental data is illustrated in fig .
[ fitres ] . experimental events .
@xmath175 energy spectrum in cms ( upper left ) , relative difference between experimental energy spectrum and fit result ( lower left ) , likelihood per event ( upper right ) and distribution of @xmath0 helicity sensitive variable @xmath182 ( lower right ) .
points with errors show experimental data , histogram - result of the fit .
open histograms show signal events , shaded histograms - background contributions .
, title="fig:",scaledwidth=48.0% ] experimental events . @xmath175 energy spectrum in cms ( upper left ) , relative difference between experimental energy spectrum and fit result ( lower left ) , likelihood per event ( upper right ) and distribution of @xmath0 helicity sensitive variable @xmath182 ( lower right ) .
points with errors show experimental data , histogram - result of the fit .
open histograms show signal events , shaded histograms - background contributions .
, title="fig:",scaledwidth=48.0% ] experimental events .
@xmath175 energy spectrum in cms ( upper left ) , relative difference between experimental energy spectrum and fit result ( lower left ) , likelihood per event ( upper right ) and distribution of @xmath0 helicity sensitive variable @xmath182 ( lower right ) .
points with errors show experimental data , histogram - result of the fit .
open histograms show signal events , shaded histograms - background contributions .
, title="fig:",scaledwidth=48.0% ] experimental events . @xmath175 energy spectrum in cms ( upper left ) ,
relative difference between experimental energy spectrum and fit result ( lower left ) , likelihood per event ( upper right ) and distribution of @xmath0 helicity sensitive variable @xmath182 ( lower right ) .
points with errors show experimental data , histogram - result of the fit .
open histograms show signal events , shaded histograms - background contributions . , title="fig:",scaledwidth=48.0% ] + reasonable agreement can be observed for the whole energy range , although the relative difference between these spectra indicates a remaining systematic effect of about a few percent .
the distribution of the likelihood per event demonstrates the acceptable quality of the fit .
the distribution of the @xmath0 helicity sensitive variable @xmath182 @xcite is also shown in fig . [ fitres ] .
a spin - spin correlation of tau leptons is clearly demonstrated in fig .
[ spincor ] for @xmath181 events ; the @xmath175 energy spectrum shape changes notably as @xmath182 varies from @xmath183 to @xmath184 . [ cols="^,^,^ " , ] it is confirmed that the uncertainties arising from the physical and apparatus corrections to the p.d.f .
are well below 1% ; see table [ systab ] .
the statistical uncertainties of the normalisation coefficients are kept as small as possible .
the contribution to the systematic uncertainties of the michel parameters due to the finite accuracy of the normalisation coefficients shown in table [ systab ] are evaluated with the entire available generic @xmath97 mc sample ; they provide the dominant contributions .
we observe a correlation of about @xmath185% between the @xmath4 and @xmath5 parameters .
the slope of the corresponding error ellipse exhibits an approximate dependence of @xmath186 , which is incorporated as an inflated uncertainty of the @xmath5 parameter in table [ systab ] .
cc@ c@ c@ c source & @xmath187,% & @xmath188,% & @xmath189,% & @xmath190,% + + isr+@xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + @xmath196 & @xmath197 & @xmath192 & @xmath198 & @xmath199 + @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 + + resolution@xmath205brems .
& @xmath192 & @xmath206 & @xmath203 & @xmath207 + @xmath208 & @xmath209 & @xmath210 & @xmath197 & @xmath195 + + @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 + @xmath216 & @xmath217 & @xmath217 & @xmath218 & @xmath197 + @xmath219 & @xmath204 & @xmath197 & @xmath220 & @xmath221 + total & @xmath222 & @xmath223 & @xmath224 & @xmath225 + however , we still observe a systematic bias of the order of a few percent , especially in the @xmath2 and @xmath3 michel parameters .
this bias originates from the remaining inaccuracies in the description of the @xmath155 background .
we present a study of michel parameters in leptonic @xmath0 decays using a @xmath92 data sample collected at belle .
michel parameters are extracted in the unbinned maximum likelihood fit of the @xmath54 events in the full nine - dimensional phase space .
we exploit the spin - spin correlation of tau leptons to extract @xmath226 and @xmath227 in addition to the @xmath4 and @xmath5 michel parameters .
although systematic uncertainties coming from the physical and apparatus corrections as well as from the normalisation are below 1% , currently we still have a relatively large systematic bias in the @xmath2 and @xmath3 parameters , which originates from the inaccurate description of the dominant @xmath155 background .
we thank the kekb group for the excellent operation of the accelerator ; the kek cryogenics group for the efficient operation of the solenoid ; and the kek computer group , the national institute of informatics , and the pnnl / emsl computing group for valuable computing and sinet4 network support . we acknowledge support from the ministry of education , culture , sports , science , and technology ( mext ) of japan , the japan society for the promotion of science ( jsps ) , and the tau - lepton physics research center of nagoya university ; the australian research council and the australian department of industry , innovation , science and research ; austrian science fund under grant no .
p 22742-n16 and p 26794-n20 ; the national natural science foundation of china under contracts no .
10575109 , no .
10775142 , no . 10825524 , no .
10875115 , no . 10935008 and no . 11175187 ; the ministry of education , youth and sports of the czech republic under contract no .
lg14034 ; the carl zeiss foundation , the deutsche forschungsgemeinschaft and the volkswagenstiftung ; the department of science and technology of india ; the istituto nazionale di fisica nucleare of italy ; national research foundation of korea grants no .
2011 - 0029457 , no .
2012 - 0008143 , no . 2012r1a1a2008330 , no .
2013r1a1a3007772 , no .
2014r1a2a2a01005286 , no . 2014r1a2a2a01002734 , no .
2014r1a1a2006456 ; the brl program under nrf grant no .
krf-2011 - 0020333 , no .
krf-2011 - 0021196 , center for korean j - parc users , no .
nrf-2013k1a3a7a06056592 ; the bk21 plus program and the gsdc of the korea institute of science and technology information ; the polish ministry of science and higher education and the national science center ; the ministry of education and science of the russian federation and the russian federal agency for atomic energy ; the slovenian research agency ; the basque foundation for science ( ikerbasque ) and the upv / ehu under program ufi 11/55 ; the swiss national science foundation ; the national science council and the ministry of education of taiwan ; and the u.s.department of energy and the national science foundation .
this work is supported by a grant - in - aid from mext for science research in a priority area ( `` new development of flavor physics '' ) and from jsps for creative scientific research ( `` evolution of tau - lepton physics '' ) .
w. fetscher , h. j. gerber and k. f. johnson , phys .
b * 173 * ( 1986 ) 102 .
w. fetscher and h. j. gerber , adv .
high energy phys .
* 14 * ( 1995 ) 657 .
k. a. olive _ et al . _
( particle data group ) , chin .
c * 38 * ( 2014 ) 090001 .
a. b. arbuzov , g. v. fedotovich , e. a. kuraev , n. p. merenkov , v. d. rushai and l. trentadue , jhep * 9710 * ( 1997 ) 001 [ hep - ph/9702262 ] .
a. b. arbuzov , g. v. fedotovich , f. v. ignatov , e. a. kuraev , a. l. sibidanov , binp preprint 2004 - 70 .
+ http://www.inp.nsk.su/activity/preprints/files/2004_070.pdf a. b. arbuzov , g. v. fedotovich , f. v. ignatov , e. a. kuraev and a. l. sibidanov , eur .
j. c * 46 * ( 2006 ) 689 [ hep - ph/0504233 ] .
e. a. kuraev and v. s. fadin , sov .
j. nucl .
* 41 * ( 1985 ) 466 [ yad .
* 41 * ( 1985 ) 733 ] .
f. a. berends , r. kleiss , s. jadach and z. was , acta phys .
b * 14 * ( 1983 ) 413 .
s. jadach and z. was , acta phys .
b * 15 * ( 1984 ) 1151 [ erratum - ibid .
b * 16 * ( 1985 ) 483 ] .
s. jadach and z. was , comput .
commun .
* 36 * ( 1985 ) 191 .
a. b. arbuzov , phys .
b * 524 * ( 2002 ) 99 .
s. kurokawa and e. kikutani , nucl .
methods phys .
a * 499 * ( 2003 ) 1 , and other papers included in this volume ; t.abe _ et al . _ ,
prog . theor .
* 2013 * ( 2013 ) 03a001 and following articles up to 03a011 . | we present a study of michel parameters in leptonic @xmath0 decays using experimental information collected at the belle detector .
michel parameters are extracted in the unbinned maximum likelihood fit of the @xmath1 events in the full nine - dimensional phase space .
we exploit the spin - spin correlation of tau leptons to extract @xmath2 and @xmath3 in addition to the @xmath4 and @xmath5 parameters . |
depending on the type and concentration of surfactant molecules and added salt , solutions of surfactant wormlike micelles have shear thinning or thickening behavior under shear flow . unlike most fluids , wormlike micelles often have non - analytic flow curves with sharply - selected plateaus along which strain rate or stress may change discontinuously . in the well documented case of shear thinning solutions the usual explanation of the constant stress plateau
is shear banding @xcite , _
i.e. _ a separation of the material into bands of different viscosities , triggered by a constitutive instability ( such as an isotropic - to - nematic transition @xcite ) . as shown recently @xcite , the stress selection and history independence of shear banding can be explained using the inhomogeneities of the relevant mesoscopic order parameter ( polymer stress ) , _ i.e. _ by incorporating `` diffusive '' terms in the constitutive equations .
order parameter diffusion was introduced long time ago by van der waals in the so - called `` gradient theory '' of the gas - liquid interface @xcite , and is obligatory in phase field models for pattern formation .
notwithstanding a few attempts to deal with inhomogeneous stresses @xcite the same concept has not obtained full acceptance in the rheological community . while one might argue that diffusion terms are negligibly small , these _ non - perturbative _ terms resolve stress selection even for infinitesimal values @xcite . however , a small diffusion coefficient should also imply a slow approach to steady state ; the main purpose of this letter is to demonstrate these long time scales experimentally and relate them to simple model diffusive behavior .
shear banding involves spatial inhomogeneity and several temporal stages .
light polarization probes the local micellar orientation , while rheology detects the molecular stress . in this work
we shall calculate the stress transients using a theoretical model .
this will be compared to birefringence measurements that are assumed to probe the state of molecular orientation and hence stress ( but see ref .
@xcite ) rather than the typical start - up transient experiment , we consider the simpler experiment of a step between two fixed values of the shear rate in the banded regime .
small steps should induce less drastic changes in the fluid while still remaining in the non - linear regime , and hopefully yield more controllable results .
the transient features will be shown to be intimately related to the dynamics of the interface between the bands .
ccccccccccc no . & salt & t [ @xmath0c ] & @xmath1 [ pa ] & @xmath2 [ s ] & @xmath3 [ s ] & @xmath4 [ s ] & @xmath5 [ s ] & @xmath6 & @xmath7 & @xmath8 + 1 & 1.79 m @xmath9 & 30 & 232 & 0.17 & 0.2 & 2.4 & 30.6 & 0.85 & 19 & 0.64 + 2 & 0.405 m @xmath9 & 30 & 238 & 0.17 & 0.2 & 1.8 & 25 & 1.27 & 19 & 0.66 + 3 & 0.3 m @xmath10 & 34 & 235 & 0.16 & 0.2 & 3.7 & 9.3 & 1.12 & 80 & 0.66 the surfactant solutions used are summarized in table [ t.1 ] .
the stress response was measured using an rfs rheometrics scientific controlled strain rate rheometer in couette ( radii @xmath11 ) and cone - plate geometries .
linear response is of the maxwell type with almost identical relaxation times @xmath2 and moduli @xmath1 for the three solutions . in the non - linear regime
the coexistence plateau width @xmath12 for solution @xmath13 is larger than for solutions @xmath14 and @xmath15 , while solutions @xmath14 and @xmath15 have a sloped stress plateau that roughly follows a power law @xmath16 , with @xmath17 ; a slope could indicate concentration differences between coexisting states @xcite . upon a step increase of the shear rate between banded states
the stress increases to a maximum , and then decreases monotonically ( or sometimes by a small undershoot and a monotonic increase ) .
three relaxation times ( @xmath18 , @xmath19 and @xmath20 ) follow successively after the overshoot until steady flow is reached ( fig .
[ f.1]a ) .
the three time scales are well separated and the result is reproducible for all solutions and both cone - plate and couette geometries .
we filmed the step shear rate experiment for the third solution in a couette geometry with a slightly larger gap ( radii @xmath21 ) between crossed polarizers to extract the average extinction angle @xmath22 .
the kinetics of @xmath22 resemble that of the shear stress , displaying time scales similar to the second and the third time scales found in the rheology ( fig .
[ f.1]c ) .
the first , unresolvable , time scale is shorter than the interval between successive video frames .
the second time scale is well resolved , although shorter by a factor of two than the equivalent time in the stress rheology measurements .
the third time scale is buried in noise and regression does not provide significant results .
the momentum balance is @xmath23 where @xmath24 is the fluid density and @xmath25 is the velocity field .
the stress tensor @xmath26 is given by @xmath27 , where the pressure @xmath28 is determined by incompressibility ( @xmath29 ) , @xmath30 is the `` solvent '' viscosity , @xmath31 is the `` polymer '' stress , and @xmath32 is the symmetric part of the velocity gradient tensor @xmath33 .
the non - newtonian `` polymer '' viscoelastic stress @xmath34 is assumed to obey the diffusive johnson - segalman ( d - js ) model @xcite , @xmath35 where @xmath36 is the anti - symmetric part of @xmath37 , @xmath38 is the `` polymer '' viscosity , @xmath2 is a relaxation time , and @xmath39 is the diffusion coefficient . the `` slip parameter '' @xmath40 ( describing the non - affinity of the deformation ) is necessary to reproduce a non - monotonic constitutive curve , and the added diffusion term was shown to resolve stress selection @xcite .
the initial dynamics is governed by inertia ; within a very short time @xmath41 ( @xmath42 for @xmath43 and a gap @xmath44 ) momentum diffuses across the gap and the momentum balance becomes @xmath45 .
the subsequent slower dynamics is controlled by the viscoelastic response of the fluid . in a planar geometry with @xmath46
, our constitutive model leads to a system of reaction - diffusion equations : @xmath47 where @xmath48 is the shear rate , @xmath49 , and @xmath50 is a combination of the polymer normal stresses , @xmath51 and @xmath52 .
@xmath53 are the order parameters of the transition ( @xmath54 is small in the nematic ( n ) band , and large in the isotropic ( i ) band ) .
they can diffuse across stream lines with diffusion coefficient @xmath55 , and relax in the linear regime within the linear ( maxwell ) time @xmath2 .
the non - linear reaction terms @xmath56 and @xmath57 can be straightforwardly derived from eq .
( [ eq : js ] ) @xcite .
where @xmath62 . system ( [ eq : cshear ] ) describes the dynamics along a streamline at prescribed shear rate ; in this case @xmath63 changes proportionately to @xmath54 according to eq .
( [ eq : stress ] ) .
system ( [ eq : cstress ] ) describes the dynamics along a streamline at constant total stress @xmath63 .
the two dynamical systems have the same fixed points ( since , for homogeneous steady flow , @xmath48 and @xmath63 are related by eq .
[ eq : stress ] ) : stable fixed points ( attractors ) representing the bands i and n , and an intermediate unstable saddle fixed point .
coexistence of bands at common total stress is possible only for @xmath64 $ ] . linearizing systems ( [ eq : cshear ] ) and ( [ eq : cstress ] ) about the fixed points yields the dominant relaxation times of the attractors , @xmath65 and @xmath66 .
these are different for the two dynamical systems , denoted at constant shear rate by @xmath67 and at constant stress by @xmath68 . for the js model @xmath69 for all @xmath70 .
@xmath71 is close to @xmath2 , while @xmath72 is larger than @xmath71 and diverges as @xmath73 ( fig .
[ f.2]a ) .
this divergence is consistent with ref .
@xcite : controlled stress experiments have much longer relaxation times on the metastable extension of the high viscosity branch above the constant stress plateau .
consider an initial banded steady state , with average shear rate @xmath74 . suddenly increasing the average shear rate to @xmath75 produces a stress overshoot because the amount of high viscosity i band is too large ; both i and n bands then become unstable , and the stress can decrease by producing more of the low viscosity n material .
numerical simulation ( figs.[f.1 ] , [ f.3 ] ) shows that this occurs in three stages : _ 1 .
band i destabilization_ during this stage the i band tries a direct passage toward the nematic band n , fig.[f.3]a ) .
complete transformation is forbidden by the average shear rate constraint and the i band stops before reaching the basin of attraction of the steady n band .
a representation of the subsequent kinetics in the @xmath76 plane ( fig . [ f.2]b ) shows that _ the n band almost follows the steady flow curve , while the i band evolves at constant shear rate_. thus , the characteristic time @xmath77 is controlled by the i band dynamics .
the total shear stress at the end of this stage depends on the final position of the interface and is sensitive to the details of the constitutive model .
if this value is below the plateau one finds an undershoot ( as in the numerical simulation fig .
[ f.1]b ) , and otherwise the subsequent stress evolution is monotonic ( fig .
[ f.1]a ) .
although we did not succeed in reproducing monotonic evolution using the js model , as long the i band evolves at a constant shear rate , the magnitude of the characteristic time should not be affected by the presence or absence of an undershoot .
interface reconstruction_ at the end of stage 1 the interface separates an unsteady i band close to the unstable saddle point from a nearly stable n band . the part of the profile closer to the i attractor will evolve toward this one , while the other part approaches the n attractor .
this reconstructs the interface in a more advanced position , stabilizes the bands , and increases the contrast between them .
interestingly , there is a spatial position at which @xmath54 and @xmath50 practically remain fixed at their saddle fixed point values . during this stage
@xmath74 is constant ( because it is imposed ) and @xmath78 is almost constant ( because of the compensating evolutions of the two bands ) , so according to eq .
( [ eq : stress ] ) the total stress variation is small , and the characteristic time ( controlled by the i band ) is @xmath79 .
this time exceeds the linear viscoelastic time @xmath2 ( see fig .
[ f.2 ] ) , and depends on how close the stress at the end of stage 1 is to the spinodal limit @xmath80 , and on the quantitative details of the curve in fig .
[ f.2 ] , all of which are sensitive to the constitutive model . the analysis suggested by the numerical experiment is confirmed by the birefringence measurements .
the sequence of images in fig .
[ f.4 ] show the gap of the couette cell filmed between crossed polarizers during stage 2 . although we can not quantitatively compare figs . [ f.3]b ) and [ f.4 ] ( the relation between the transmitted intensity and the order parameter is unknown and sure to be non - linear ) , the sharpening of the contrast corresponding to the interface reconstruction is visible .
the difference between the characteristic times for the extinction angle and rheology ( fig.[f.1 ] ) could be due to the different couette cell gap widths .
_ 3 . interface travel_ the instability and reconstruction of the interface in the first two stages is ensured by the reaction terms of the eq .
( [ eq : system ] ) , ending when a sharp interface between stable bands is fully formed .
this interface could have a non - zero velocity if it forms in at a position corresponding to a stress value above or below the plateau stress @xmath81 .
`` front propagation '' over the small distance toward the final equilibrium position is then controlled by @xmath55 ( this distance is too small to observe by birefringence ) . because of the undershoot in the numerical simulation the sign of the displacement during stage 3 is opposite to the one in the first stages ( fig.[f.3]c ) .
the characteristic time @xmath5 for this stage follows from the velocity @xmath82 of the sharp interface close to steady state which is independent of the presence or absence of the undershoot @xcite ( see below ) .
jump : birefringence images and profiles ( averaged between the two vertical lines on the film ) corresponding to the second time scale , showing the interface reconstruction .
the moving ( inner ) cylinder is at @xmath83 and the fixed ( outer ) cylinder is at @xmath84 . ]
let us consider a single sharp interface , at a position @xmath85 inside the gap .
at imposed shear rate , the lever rule @xmath86 relates @xmath63 and @xmath85 and leads to @xmath87 } , \label{eq : drivingf}\ ] ] where @xmath88 is the width of the plateau , and @xmath89 .
we consider @xmath90 , @xmath91 ( as in the experiments ) so that @xmath92 ( true for piecewise linear flow curves and obeyed well by the js model ) .
we showed previously that the velocity @xmath82 of the interface is a function of the total shear stress @xmath63 only , and that @xmath93 when @xmath94 @xcite which via the lever rule corresponds to a unique stable interface position @xmath95 .
thus , close to this position @xmath96 the equation of motion of the interface is : using eqs .
( [ eq : drivingf],[eq:2 ] ) , and the derivative @xmath98 , we find the solution @xmath99 , with characteristic time @xmath100/\dot\gamma_i}{\delta \dot\gamma } , \label{eq : time3}\ ] ] where the dimensionless constant @xmath101 depends on the particular constitutive model .
( [ eq : time3 ] ) implies that a fully formed interface equilibrates faster in systems with larger plateaus @xmath102 , such as ctab / kbr .
in such cases , eq .
( [ eq : drivingf ] ) implies larger stress variations and thus larger interface accelerations for the same position variation .
this is compatible with the shorter @xmath5 in table [ t.1 ] . for simplicity ,
( [ eq : drivingf ] ) was for a planar geometry ; in cylindrical couette flow a slight correction ( negligible for the thin gaps we consider ) leads to a smaller @xmath5 . using the experimental value of @xmath5 and eq .
[ eq : time3 ] , we can estimate @xmath55 . in order to do this
we need the value of @xmath101 . in the js model , while @xmath1 and @xmath2 are measurable , the two free parameters @xmath103 and @xmath40 determine the function @xmath101 .
nevertheless , @xmath104 , @xmath105 and @xmath106 are functions of @xmath103 only , given to a good approximation ( @xmath107 ) by @xmath108 , @xmath109 , @xmath110 . from either @xmath111 , or @xmath112 and experimental data
, one can estimate @xmath103 ( the average of the two values is given in table [ t.2 ] ) and then @xmath113 gives k ( table [ t.2 ] ) .
thus we do not need the value of @xmath40 .
microscopically , we expect @xmath114 where @xmath115 is the stress correlation length . in dilute solutions this should be the micelle gyration radius @xcite , while in concentrated solutions a reasonable candidate is the mesh size @xmath116 , which can be estimated from @xmath117 .
the results are presented in table [ t.2 ] .
the stress correlation length is of order the mesh size , which is reasonable ; however there is still no theory for such a diffusive term in concentrated solutions .
to conclude , a general dynamical systems analysis of the d - js model provides plausible explanations for the observed time scales , and consistent estimates of the stress diffusion coefficient @xmath55 .
nevertheless , neither the d - js model , nor reptation - retraction - reaction models @xcite can provide perfect fits of transient stress curves in the nonlinear regime . also , it it conceivable that concentration differences between the bands could influence the observed time scales .
the difference of the values of @xmath118 between samples 1,2 and 3 ( table [ t.2 ] ) could be a concentration effect , consistent with the different slopes of the flow curves plateaus ( tilted for 1,2 , almost horizontal for 3 ) .
ccccccccc sample & @xmath119 [ mm ] & @xmath120 & @xmath103 & @xmath101 & d [ @xmath121 & @xmath115 [ nm ] & @xmath116 [ nm ] + 1 & 0.3 & 4 & @xmath122 & @xmath123 & @xmath124 & 44 & 26 + 2 & 0.3 & 2.33 & @xmath125 & @xmath126 & @xmath124 & 43 & 26 + 3 & 0.3 & 1.85 & @xmath127 & @xmath128 & @xmath129 & 22 & 26 we thank j .-
f.tassin for his kind assistance and for allowing our access to the rheometer facilities of universit du maine .
acknowledges funding from eu cost action p1 on soft condensed matter and from epsrc ( gr / l70455 ) at the beginning of this work . | we show the existence of three well defined time scales in the dynamics of wormlike micelles after a step between two shear rates on the stress plateau .
these time scales are compatible with the presence of a structured interface between bands of different viscosities and correspond to the isotropic band destabilization during the stress overshoot , reconstruction of the interface after the overshoot and travel of a fully formed interface .
the last stage can be used to estimate a stress diffusion coefficient . |
[ [ section ] ] + the currently accepted standard model of elementary particles and their interactions is based on higgs mechanism of spontaneous symmetry breaking of the local electroweak gauge symmetry that leaves the renormalizability of the model intact @xcite . in this mechanism masses of the quarks ,
the charged leptons , the weak gauge bosons and the neutral scalar higgs of the model are generated from the asymmetry of the vacuum .
[ [ section-1 ] ] + a convenient tool to investigate the nature of the vacuum structure of a renormalizable quantum field theory and stability of the theory with spontaneous symmetry breaking at zero and finite temperature is the effective potential , which contains all quantum radiative corrections to the tree classical potential .
the formalism of the effective potential as an essential tool for the investigation of the spontaneous symmetry breaking was first considered by goldstone and et al @xcite and jona - lasinio @xcite , ( for review of early works see for example ref.@xcite ) .
the physical interpretation of the effective potential as a means to explore the vacuum structure of a quantum field theory has been elaborated by symanzik @xcite , ( for review see for example ref.@xcite and ref.@xcite ) .
the vacuum structure of a quantum field theory involves several quantitative calculable quantities such as symmetry and asymmetry of the vacuum , stability and instability of the vacuum , symmetry breaking patterns , restoration of symmetry , induced asymmetry and vacuum energy .
[ [ section-2 ] ] + the effective potential at higher order is used to investigate the vacuum structure of spontaneous symmetry breaking through coleman - weinberg mechanism @xcite for massless standard model ( sm ) .
starting with a theory massless in the tree - level approximation is an attractive idea where it avoids the problems associated with the negative mass square term .
[ [ section-3 ] ] + the standard model contains some of free parameters , one of these parameters is the mass of the higgs boson , as a tree level higgs scalar coupling @xmath5 so the mass of the higgs boson has a priori no constraints from sm theory , and its determination self - consistently poses a challenging problem . in the present work , an attempt is made to determine the higgs boson mass self - consistently from the renormalization group equations and the effective potential of the model .
investigation the stability of the vacuum leads to determine the higgs boson mass from the effective potential for each order by using the renormalization condition .
this paper is organized as follows , in sec .
ii , we apply the renormalization group equation method with use corresponding renormalization group functions to calculate the effective potential up to three - loop order for the massless sm . in sec .
iii , we drive a relation that allows to calculate the higgs boson mass and we discuss the results and investigate the stability of the vacuum structure under radiative quantum corrections up to three - loop order in different mass scale . in addition , we compare our result with some results of other authors employing different approaches . in sec .
iv , we present our conclusion .
[ [ section-4 ] ] + in the mass - independent thooft - weinberg scheme @xcite the full effective potential @xmath14 satisfies renormalization group equation rge .
thus , in general the rge for @xmath14 : @xmath15 where @xmath16 + here @xmath17,@xmath18,@xmath19 and @xmath20 are the renormalization group functions for scalar , mass and gauge couplings , respectively and @xmath20 is the anomalous dimension , which usually calculated perturbatively .
let us now rewrite the rge as a short form by defining the parameters of the theory , @xmath21 to get : @xmath22v(\varphi)=0 \label{eq:2_3}\ ] ] + where @xmath23 equals @xmath24 for mass coupling and 1 otherwise .
also @xmath25 represents all the gauge couplings @xmath26,@xmath27,and @xmath28 of u(1 ) , su(2 ) and su(3 ) respectively , in addition to the top quark coupling @xmath29 .
the renormalization group function rgf @xmath30 , which called the beta function is given by : @xmath31 with @xmath32 + in the loop expansion , the effective potential has the expression@xcite : @xmath33 likewise , the renormalization group functions have the expansion : @xmath34 where the @xmath35 is the n - loop contribution to @xmath36 . then following the ref.@xcite
, we can write the following n - loop of @xmath37-order as : @xmath38 where @xmath39 is the differential operator .
hence , one can find the n - loop contribution to the effective potential by using the recursion formula eq.([eq:2_7 ] ) in terms of rgfs and the tree level potential .
[ [ section-5 ] ] + the tree level ( zero - loop ) effective potential for massless sm is : @xmath40 one obtains : @xmath41 + now , applying eq.([eq:2_7 ] ) to eq.([eq:2_9 ] ) using eqs.([eq:2_6 ] ) and ( [ eq:2_8 ] ) one gets for one - loop contribution to the effective potential : @xmath42 where @xmath43 integration of eq.([eq:2_10 ] ) yields : @xmath44 + the two - loop contribution to the effective potential : @xmath45 with @xmath46 + @xmath47 , \label{eq:2_15}\ ] ] + @xmath48 . \end{aligned } \label{eq:2_16}\ ] ] + for the three - loop contribution to the effective potential , one gets : @xmath49-\frac{k_{3}}{1152}\varphi^{4}(\ln\frac{\mu^{2}}{\mu^{2}_{0}})^{3}+c_{3}(\lambda , g,\varphi,\mu_{0 } ) , \end{aligned } \label{eq:2_17}\ ] ] where @xmath50 , \label{eq:2_18}\ ] ] with @xmath51 + @xmath52 , \end{aligned}\label{eq:2_20}\ ] ] with @xmath53 , \end{aligned}\label{eq:2_21}\ ] ] + @xmath54 + @xmath55 , \label{eq:2_23}\ ] ] with @xmath56 [ [ section-6 ] ] + it is worth noting that integrating renormalization group equation for the effective potential in the rge method a constant @xmath57 that depending on @xmath58 and @xmath59 as it is clear in eqs.([eq:2_12 ] ) , ( [ eq:2_13 ] ) and ( [ eq:2_17 ] ) should be fixed .
in general , to determine this constant one needs to specify renormalization conditions on the effective potential .
there are several renormalization conditions .
for example , @xmath60 here the constants @xmath61 and @xmath62 are determined by using the condition : @xmath63 + therefore , the full effective potential approximated to one - loop , two - loop and three - loop respectively becomes : @xmath64 , \label{eq:2_26}\ ] ] + @xmath65+\frac{b}{192}\varphi^{4}[(\ln\frac{\varphi^{2}}{\mu^{2}})^{2}-\frac{35}{3 } ] , \end{aligned}\label{eq:2_27}\ ] ] + @xmath66\\+\frac{i_{2}}{192}\varphi^{4}[(\ln\frac{\varphi^{2}}{\mu^{2}})^{2}-\frac{35}{3}]+\frac{i_{3}}{1152}\varphi^{4}[(\ln\frac{\varphi^{2}}{\mu^{2}})^{3}-20 ] , \end{aligned}\label{eq:2_28}\ ] ] with @xmath67 + now by using renormalization group functions in the standard model one can obtain the effective potential of the massless sm for one - loop eq.([eq:2_26 ] ) , two - loop order eq.([eq:2_27 ] ) and for three - loop order eq.([eq:2_28 ] ) , these rgfs are given in the appendix a for one , two and three level @xcite .
[ [ section-7 ] ] + in the massless case the scalar - field self - interacting coupling constant @xmath5 , is determined from the minimum of the potential by using the tadpole condition : @xmath68 + by applying eq.([eq:3_1 ] ) to one - loop eq.([eq:2_26 ] ) , two - loop eq.([eq:2_27 ] ) and three - loop eq.([eq:2_28 ] ) order effective potential we get : @xmath69 @xmath70 @xmath71=0 , \end{aligned}\label{eq:3_3}\ ] ] @xmath72 @xmath73+\frac{36i_{3}}{1152}(\ln\frac{v^{2}_{3}}{\mu^{2}})^{2}+\frac{24 i_{3}}{1152}[(\ln\frac{v^{2}_{3}}{\mu^{2}})^{3}-20]=0 . \end{aligned}\label{eq:3_4}\ ] ] + these equations allows us to obtain the scalar higgs coupling @xmath5 for each order if the gauge and yukawa couplings at the renormalization scale are known , where @xmath74 is the vacuum expectation value .
the vacuum expectation values have the expansions up to one - loop , two - loop and three - loop respectively : @xmath75 @xmath76 @xmath77 + in perturbative theory the order n - contribution should be less than @xmath78-order to make the perturbative series convergent .
so , @xmath79 and @xmath80 contributions are too small compared to the one - loop contribution value , so we consider the approximation @xmath81 experiment value ( 246.2 gev ) . + +
now the higgs boson mass is calculated from @xcite : @xmath82 and satisfies the condition of eq.([eq:3_1 ] ) .
+ + at one - loop order , using eq.([eq:2_26 ] ) , eq.([eq:3_2 ] ) and eq.([eq:3_5 ] ) : @xmath83 + for two - loop order , using eq.([eq:2_27 ] ) , eq.([eq:3_3 ] ) and eq.([eq:3_5 ] ) : @xmath84 + and finally at three - loop order , using eq.([eq:2_28 ] ) , eq.([eq:3_4 ] ) and eq.([eq:3_5 ] ) : @xmath85 + equations ( [ eq:3_6 ] ) , ( [ eq:3_7 ] ) and ( [ eq:3_8 ] ) give the higgs boson mass in one - loop , two - loop and three - loop approximations , which are not the physical masses but the running masses .
[ [ section-8 ] ] + in the massless case , we can simplify the condition in eq.([eq:3_1 ] ) if we recall that @xmath6 is an arbitrary parameter , we are allowed to choose @xmath6 to be location of the minimum of the effective potential as has been done in ref.@xcite .
equation ( [ eq:3_2 ] ) , eq.([eq:3_3 ] ) , and eq.([eq:3_4 ] ) become : @xmath86 @xmath87 @xmath88 + note that , eq.([eq:3_9 ] ) can be viewed as a generalization of the condition that relates the scalar coupling to the gauge coupling obtained in ref.@xcite for the case of massless scalar quantum electrodynamics .
+ + the higgs boson mass for @xmath89 is described by : @xmath90 @xmath91 @xmath92 in present work , our initially input parameters are taken from ref.@xcite at the neutral weak scale @xmath93 .
these parameters can be computed at any scale for each loop by using the following perturbative expression : @xmath94 @xmath95 for the strong coupling @xmath96 and the top quark mass @xmath97 we have taken the range as in @xcite,@xcite : @xmath98 + * rusults and discussion * [ [ section-9 ] ] + the computations for the one - loop , two - loop and three - loop order within the @xmath99range @xmath100 are presented in table.([tab:3_1 ] ) for @xmath1 gev , @xmath101 . in this table and
the other tables we have used the reduced scalar coupling @xmath102 , which corresponds to gauge couplings @xmath103 , @xmath104 for u(1 ) , su(2 ) and su(3 ) , and appears as perturbative expansion parameter in the loop expansion of the effective potential . for the case
@xmath105 gev , there is a minimum of the one - loop , two - loop and three - loop effective potential ( see fig.([fig:3_1 ] ) ) with too large values for higgs scalar coupling and mass , so this result may be ruled out since it lies outside the range of validity of the approximation .
it turns out that , increasing the value of @xmath6 improves the results . by considering the neutral weak mass @xmath93 as a mass scale ,
the higgs boson mass value is @xmath106 gev at one - loop order .
this result is consistent with the recent lhc discovery @xcite .
however for the value , @xmath107 gev , we get @xmath108 gev , @xmath3 gev and @xmath109 gev . at @xmath110 gev ,
the scalar coupling @xmath5 at two - loop becomes negative and leads to metastable vacuum while the three - loop order is stable even at high scale energy @xmath7 gev . for the larger @xmath6-range @xmath8 at @xmath1 gev ,
@xmath101 spontaneous symmetry breaking for one - loop and three - loop occurs ( see fig.([fig:3_3 ] ) at approximately the same scalar coupling values as shown in table.([tab:3_2 ] ) .
in addition , we get @xmath9 gev at @xmath10 gev while for three - loop @xmath11 gev at @xmath12 gev .
[ [ section-10 ] ] + we present our results in table.([tab:3_3 ] ) for the scalar higgs coupling , obtained from eqs.([eq:3_9]),([eq:3_10 ] ) and eq.([eq:3_11 ] ) , and higgs boson mass in the ranges of @xmath111 and @xmath112 for each loop order with consideration that the mass scale equals to vacuum expectation value @xmath113 . thus the higgs boson mass for each loop is : @xmath114 + at first glance to the numerical results of one - loop order , one can find that higgs boson mass @xmath0 is large for all the range under our study because the values of the scalar higgs coupling @xmath5 may be too large to justify the neglect of higher order contributions .
however , the two - loop and three - loop results are not appreciably different from each other .
the two - loop and three - loop radiative corrections have led to an improvement of the higgs boson mass and the value of the scalar coupling .
[ [ section-11 ] ] + it is interesting to note that in massless theory the absence of tree level mass term leads to large value for the scalar coupling to the limit , where the perturbation method becomes invalid , since as known the perturbation parameters should be small in order to become successively smaller at higher - order terms . in order to overcome this problem one can choose a suitable mass scale with the inclusion of higher - order contributions . [ [ section-12 ] ] +
the recent discovery of higgs - like particle does not confirm or exclude the possibility of many higgs particles with masses larger than the one discovered in cern .
extensions of the standard model to for example supersymmetry ( susy ) models predict the existence of families of higgs bosons , rather than the one higgs particle of the standard model . in all of them ,
there is one neutral higgs boson with properties similar to those of the standard model higgs boson in addition to neutral and charged higgs bosons .
for example in the minimal supersymmetric standard model ( mssm ) , which the simplest model of the susy , the higgs mechanism yields to two higgs doublets , leading to the existence of a quintet of scalar particles . in this model , the lightest higgs boson mass that represent the standard model - like higgs ,
is predicted to be @xmath115 gev .
conversely , larger values of the standard model - like higgs mass up to @xmath116 gev can be obtained with the non - minimal supersymmetric standard model ( nmssm ) that can avoid the problems of the mssm with larger particle content @xcite,@xcite .
[ [ section-13 ] ] + recently , late last year ( 2015 ) , a heavy higgs - like particle is not exclude from the large hadron collider lhc where atlas and cms at lhc have reported that proton - to - proton collisions had led to create more photon pairs with energies around 750 gev than was expected .
this excess in the diphoton mass distribution ( @xmath117 gev ) , if confirmed , would indicate the presence of unexpected new elementary particles @xcite .
a large number of theoretical interpretations have appeared following lhc discovery @xcite .
most of these papers explain the excess through some new boson with mass-750 gev .
for example in ref.@xcite , a heavy higgs - like boson which is six times heavier than the standard model - higgs particle is expected to couple with new types of fermions .
[ [ section-14 ] ] + in addition , there is an agreement between our result eq.([eq:3_144 ] ) and some pervious theoretical results of other authors employing different methods @xcite .
[ [ section-15 ] ] + a comparison between the one , two and three - loop effective potential at lowest values @xmath1 gev and @xmath118 and largest values @xmath119 gev and @xmath120 are shown in figs.([fig:3_5 ] ) and ( [ fig:3_8 ] ) .
the conclusion that can be inferred from these figures is the curves of the two - loop and three - loop effective potential are almost similar and the position of the vacuum for the three - loop order is slightly deeper than the two - loop .
the one - loop effective potential is deeper at the vacuum position although it has the same trend .
furthermore , it must be noted that all these figures with different values range of @xmath97 and @xmath96 are almost identical and show similar trend .
.values of the higgs scalar coupling , @xmath121 , and higgs boson mass for one - loop , two - loop and three - loop order at @xmath1 gev and @xmath101 for the mass scale - range @xmath122.[tab:3_1 ] [ cols="^,^,^,^,^,^,^,^,^,^ " , ] scale = 100 gev . ]
[ [ section-16 ] ] scale @xmath123 gev . ]
scale = @xmath124 . ]
scale = @xmath124 . ]
[ [ section-17 ] ] + in this paper , we have used the renormalization group method to calculate the effective potential for massless standard model up to three - loop order .
we have also investigated the stability of the effective potential .
it is shown that while the vacuum structure for two - loop order is metastable , it is stable for three - loop order even at high energy scale .
the higgs boson mass is calculated at each loop order for a range of energy scale and a range of the top quark mass and the strong coupling .
the obtained values of the higgs boson mass are in the experimental range for the scale energy of order @xmath125 gev .
99 g. thooft , nucl .
( 1971 ) 167 .
j. goldstone , a. salam and s. weinberg , phys .
( 1962 ) 965 .
g. jona - lasinio , nuovo cimento ( 1964 ) 1790
. b. zumino , ( 1970 ) in lectures in__elementary particles and quantum field theory _ _ , edit by s. deser et al , mit cambridge , mass .
k. symanzik , commun .
( 1970 ) 48 .
r. h. brandenberger , rev .
( 1985 ) 1 .
m. sher , phys .
( 1989 ) 274 .
s. coleman and e. weinberg , phys .
( 1973 ) 1888 .
g.thooft , nucl .
( 1973 ) 455 ; s. weinberg , phys .
( 1973 ) 3497 .
h. alhendi , phys . rev . ( 1988 ) 3749 ; ibdi ( 1989 ) 683 e. m. machacek and m. t. vaughn , nucl
( 1985 ) 70 . c. ford , i. jack and d. r. t. jones , nucl . phys .
( 1992 ) 373 ; ( 1997 ) 551(e ) .
m. luo and y. xiao , phys .
( 2003 ) 011601 .
h. a. alhendi , t. barakat and i. gh .
loqman , phys . rev .
( 2010 ) 053008 .
a. v. bednyakov , a. f. pikelner and v. n. velizhanin , nucl .
( 2013 ) 552 ; phys .
( 2013 ) 336 .
k. g. chetyrkin and m. f. zoller , jhep .
( 2013 ) 091 .
i. masina , phys .
( 2013 ) 053001 .
c. amsler et al , `` _ _ particle data group collaboration _ _ '' , phys .
( 2008 ) 1 .
r. garisto , phys . rev . lett . ( 2016 ) 150001 .
olive et al , ( particle data group ) , chin .
( 2014 ) 090001 , and 2015 update .
n. christensen , t. han and s. su , phys .
( 2012 ) 115018 ; u. ellwanger , c. hugonie and a. m. teixeira , phys .
( 2010 ) 1 .
y. nakiai , r. sato and k. tobioka , phys .
( 2016 ) 151802 ; c. petersson and r. torre , phys .
( 2016 ) 151804 ; w. cho et al , phys .
( 2016 ) 151805 .
g. li et al , phys .
( 2016 ) 151803 . v. elias , r. b. mann ,
d. g. c. mckeon and t. g. steele , phys .
( 2003 ) 251601 .
f. a. chishtie , v. elias and t. g. steele , int .
( 2005 ) 6241 .
p. p. pal and s. chakrabarty , acta .
phys . pol .
( 2009 ) 1645 .
p. kielanowski , s. r. juarez w and h. g. solis - rodriguez , phys .
( 2005 ) 096003 .
m. d. scadron , r. delbourgo and g. rupp , j. phys .
( 2006 ) 735 .
v. a. bednyakov , n. d. giokaris and a. v. bednyakov , phys .
( 2008 ) 13 . | the effective potential of massless standard model ( sm ) is calculated up to three - loop order .
the stability of the effective potential and the higgs boson mass are investigated up to three - loop order .
we found that , higgs boson mass @xmath0 of one - loop order is large .
the two - loop and three - loop results are not appreciably different from each other .
the two - loop and three - loop radiative corrections have led to an improvement of higgs boson mass and the value of the scalar coupling . for the value @xmath1 gev at the energy scale @xmath2gev , we get @xmath3 gev . at the energy scale @xmath4 ,
the scalar coupling @xmath5 at two - loop becomes negative and leads to metastable vacuum while the three - loop level is stable even at high @xmath6 @xmath7 gev .
for the larger @xmath6-range @xmath8 spontaneous symmetry breaking for one - loop and three - loop occurs at approximately the same scalar coupling values .
we get @xmath9 gev at @xmath10 gev while for three - loop @xmath11 gev at @xmath12 gev .
keywords : effective potential , renormalization group equation , spontaneous symmetry breaking , higgs boson ( mass ) + ( 0,0 ) ( 4,0 ) ; + e - mail : [email protected] , [email protected] , [email protected] |
absorption line spectroscopy is a powerful tool for studying the structure of the distant universe . by observing the absorption features imprinted in the spectra of background qsos
, we can study otherwise invisible gaseous structures to redshift as high as background qsos can be . the sloan digital sky survey
( sdss ; york et al .
2000 ) has yielded optical spectra of @xmath4k qsos at @xmath5 ( schneider et al .
this unprecedentedly large sample of qso spectra offers a rich resource for studying the distant universe using absorption spectroscopy .
for example , mgii@xmath6 2796,2803 absorption doublets are commonly seen in qso spectra obtained using ground - based spectrographs .
they provide a uniform probe of intervening gas over a broad redshift range from @xmath7 to @xmath8 .
several groups have conducted systematic searches of mgii absorption features in sdss qso spectra , producing a large sample of these absorbers for constraining their statistical properties ( e.g. bouch et al .
2004 ; nestor et al.2005 ; prochter et al . 2006 ; york et al .
2006 ; quider et al .
2011 ) .
mgii absorbers are routinely found in the vicinities ( @xmath9 kpc ) of distant galaxies and provide a convenient probe of galactic halos at high redshifts ( e.g. bergeron 1986 ; steidel et al .
1994 ; kacprzak et al .
2008 ; chen & tinker 2008 ; gauthier & chen 2011 ) .
however , their physical origin , whether the absorbers arise in infalling clouds , outflows from nearby star - forming regions , or a combination thereof , is unclear .
the ubiquitous presence of blueshifted mgii absorption features along the sightlines into the star - forming regions of @xmath10 galaxies indicates that starburst driven outflows are common in distant galaxies and that outflows contribute to some fraction of mgii absorbers uncovered along random qso sightlines ( e.g. weiner et al .
2009 ; rubin et al .
although the extent of galactic - scale winds around these galaxies is unknown , the observed self - absorption of mgii doublets in distant star - forming galaxies has motivated recent works that attribute all strong mgii absorbers ( rest - frame absorption equivalent width @xmath11 ) along qso sightlines to outflows ( e.g. mnard & chelouche 2009 ; chelouche & bowen 2010 ; mnard et al .
2011 ) . using a sample of 8500 mgii absorbers at @xmath12 from sdss dr4 quasar spectra , mnard et al .
( 2011 ; hereafter m11 ) observed a strong correlation between @xmath13 and their associated _ median _ [ oii ] luminosity per unit area , @xmath14}}$ ] , in stacked qso spectra .
the observed vs. @xmath15}}$ ] correlation is characterzed by @xmath16}}\right\rangle_{\rm med}=a\left[\frac{{\mbox{$w_r(2796)$}}}{1\,{\rm \aa}}\right]^b\ ] ] where @xmath17 and @xmath18 .
this observed correlation applies to mgii absorbers of @xmath19 .
a similar trend has also been mentioned in noterdaeme et al .
( 2010 ) , but these authors did not find a correlation between @xmath13 and @xmath1}$ ] in [ oii]-emission selected mgii absorbers . in addition , noterdaeme et al .
( 2010 ) pointed out that part of their strong mgii absorbers arise in low @xmath1}$ ] galaxies . because [ oii ] luminosity @xmath1}$ ] provides a measure of current star formation rate ( e.g. kennicutt 1998 ) , m11 interpreted the observed strong correlation as mgii absorbers tracing on - going star formation .
in addition , m11 showed that the frequency distribution function of mgii absorbers and the [ oii ] luminosity function share similar shape and amplitude , and that the number density evolution of mgii absorbers follows the cosmic star formation history . combining these empirical correlations
, the authors argue that outflows are the mechanism responsible for the observed mgii absorption in qso spectra .
the conclusion of outflows being responsible for the observed mgii absorbers appears to be discrepant from previous findings that correlates more strongly with galaxy mass and weakly with galaxy colors or recent star formation history ( e.g. steidel et al .
1994 ; chen et al .
2010a , b ) . such conclusion also makes it difficult to understand the identifications of strong mgii absorbers in the vicinities of quiescent galaxies ( e.g. gauthier et al .
2010 ; bowen & chelouche 2011 ; gauthier & chen 2011 ) .
while m11 presented a clever approach to estimate the co - moving star formation rate density @xmath20 based on the observations of mgii absorbers in sdss qso spectra , it is important to understand the underlying factors that shape the observed strong vs.@xmath14}}$ ] correlation in sdss data .
we note that the sdss fibers have a finite size of @xmath3 diameter on the sky , which corresponds to projected physical separations of @xmath21 kpc at @xmath12 .
the observed @xmath0 vs. @xmath2 anti - correlation ( e.g. chen et al .
2010a ) implies that galaxies at larger projected distances produce on average weaker mgii absorbers and lower ( or zero ) @xmath1}$ ] ( when the star - forming disks occur at angular distances @xmath22 , or @xmath23 kpc , of the qso sightlines ; see figure 1 ) .
this selection bias strenthens the apparent correlation of equation ( 1)}}=l_{\rm [ o\,ii]}(z ) / a_{\rm fiber}(z)$ ] , where @xmath24 represents the corresponding physical area of sdss fiber at redshift @xmath25 .
however , as illustrated in figure 1 , the reduction from total observed fluxes to flux surface density does not correct for the differential aperture loss of galaxy fluxes in sdss qso spectra . ] .
a similar point on the potential missed galaxy light in the sdss fibers has also been made in noterdaeme et al .
( 2010 ) .
emission features in the qso spectrum .
galaxies with gaseous halos intercepting the qso sightline are expected to imprint mgii absorption features in the qso spectrum . in this picture , the optical disks of galaxies @xmath26 and @xmath27 are covered by the fiber ( in part or in full ) , galaxies @xmath28 and @xmath29 are not . only @xmath26 and @xmath27
are expected to contribute to the [ oii ] emission line seen in the stacked qso spectrum . on the other hand , the halos of galaxies @xmath29 , @xmath26 , and @xmath27 all intercept the qso sightline , while @xmath28 does not
. only @xmath29 , @xmath26 , and @xmath27 are therefore expected to contribute to the mgii absorption statistics .
but because the qso sightline impacts the outer part of galactic halo @xmath29 , the pathlength through the halo is shorter and the mgii absorbing strength is expected to be weaker ( e.g. chen et al .
2010a ) . ] to investigate the selection bias as a result of differential aperture loss , we have carried out a monte carlo simulation study . adopting a empirical model for describing gaseous clouds around galaxies from chen et al .
( 2010a ) and empirical relations for describing the luminosity and size distributions of the general galaxy population , we demonstrate in this paper that the observed vs. @xmath30}$ ] correlation can be well re - produced without any fine - tuning after accounting for the differential aperture loss of galaxy fluxes in the sdss fiber . while under the simple halo model the majority of mgii absorbers do not directly probe star - forming disks like high column density damped absorbers ( e.g. wild et al .
2007 ) , they do probe photo - ionized clouds around distant galaxies relation of chen et al .
( 2010a ) was established based on a sample of galaxies at close projected distances to a qso sightline , including those that may produce a damped absorption feature in the qso spectrum .
we therefore expect that absorber samples generated based on the mean @xmath2 relation and the observed scatter include contributions from star - forming disks
. based on the observed number densities of damped absorbers ( rao et al .
2006 ) and mgii absorbers of @xmath31 ( nestor et al .
2005 ) at @xmath32 , we estimate that no more than 20% of these strong mgii absorbers have contributions from star - forming disks ] if more massive galaxies are surrounded by more extended mgii absorbing gas , then the number density evolution of mgii absorbers naturally follows the cosmic star formation history .
because models which do not require outflows can also reproduce the empirical correlations between absorber abundances and star formation properties , we caution drawing conclusions in favor of an outflow origin for qso absorbers based on these simple correlations . throughout this paper , we adopt a @xmath33cdm cosmology with @xmath34 and @xmath35 and a dimensionless hubble constant of @xmath36 .
to simulate the sdss observations , we first generate a mock catalog of 200,000 galaxies distributed uniformly within @xmath37 angular radius of a qso sightline at redshift between @xmath7 and @xmath38 .
the large number is necessary to provide a representative sampling of a wide range of galaxy properties ( such as size and luminosity ) and to minimize statistical noise in our simulations .
the redshift range is selected to match the study of m11 . over the redshift range of @xmath12 ,
the angular radius of @xmath39 corresponds roughly to @xmath40 kpc .
the working hypothesis of this exercise is that all galaxies are surrounded by extended gaseous halos .
this hypothesis is supported by empirical observations that reveal a high gas covering fraction around galaxies of a wide range of luminosity and color ( e.g. chen et al.2010a ) .
the gaseous halos are expected to produce mgii absorption features in the spectrum of a background qso when intercepting the qso sightline . for every mgii absorber in the mock catalog ,
the location and intrinsic properties of the absorbing galaxy are known .
it is therefore possible to compute the emission fluxes of the absorbing galaxies recorded in qso spectra .
we first generate random galaxies following a schechter probability function of their rest - frame absolute @xmath29-band magnitude @xmath41 , @xmath42 where @xmath43 is the faint - end slope of the galaxy luminosity function .
we adopt @xmath44 for the faint - end slope ( e.g. faber et al .
2007 ) and @xmath45 following chen et al.(2010a ) .
the mock galaxy sample spans a lumionsity range between @xmath46 and @xmath47 . as discusse below , we also repeat the monte carlo simulations with different faint - end slop values , @xmath48 and @xmath49 and our findings remain the same .
next , we determine the optical size of each galaxy in the mock catalog according to the luminosity - size relation from cameron & driver ( 2007 ) , @xmath50 where @xmath51 is the half - light radius , the radius within which the galaxy emits half of its total flux . for a galaxy of @xmath41 ,
the half - light radius is drawn from a gaussian distribution about the mean luminosity - size relation with a 1-@xmath52 width of 0.4 dex in the @xmath53 space . to include the majority of the galaxy light , the radius that contains 90% of the total flux @xmath54 is more relevant in our study .
we therefore convert @xmath51 to @xmath54 , assuming an exponential surface brightness profile .
next , we determine the [ oii ] luminosity @xmath1}$ ] for each galaxy assuming that the rest - frame @xmath29-band fluxes trace the mean profile of [ oii ] emission .
the expected @xmath1}$ ] is then calculated according to the correlation between @xmath41 and @xmath30}$ ] found in deep survey data .
based on the study of zhu et al .
( 2007 ) , we find that the @xmath41 vs. @xmath55}$ ] correlation can be characterized by @xmath56}/({\rm erg}\,{\rm s}^{-1})]=35 - 0.3\,m_b \ ] ] with a 1-@xmath52 scatter of @xmath57}}=0.3 $ ] dex .
note that equation ( 4 ) is an empirical relation between observed quantities .
it does not include dust extinction corrections .
the inferred @xmath1}$ ] can therefore be directly compared to what is observed in the stacked sdss qso spectra . for a galaxy of @xmath41 , @xmath1}$ ]
is drawn from a gaussian distribution about the mean @xmath41-@xmath1}$ ] relation with a 1-@xmath52 width of 0.3 dex in the @xmath58}$ ] space . to examine whether or not our mock galaxy sample is representative of the field galaxy population
, we present the [ oii ] luminosity distribution of our mock galaxy sample in figure 2 along with the observed [ oii ] luminosity function of @xmath10 galaxies from different deep surveys ( takahashi et al . 2007 ;
zhu et al . 2009 ) .
all the empirical measurements have been converted to have the same cosmological parameters adopted in our analysis .
figure 2 demonstrates that we reproduce the observed [ oii ] luminosity function with the mock galaxy sample for the luminous galaxy population . at the faint - end ,
the observations suffer from survey incompletness ( e.g. zhu et al .
2009 ) and therefore the observed space density represents a lower limit to the underlying faint galaxy population .
luminosity distribution of the mock galaxy sample described in section 2 .
the histograms display the simulated data for different faint - end slope of @xmath48 ( dotted histogram ) @xmath44 ( solid histogram ) , and @xmath49 ( dashed histogram ) .
the [ oii ] luminosity function measured for @xmath10 galaxies in deep2 ( closed circles ; zhu et al .
2009 ) and for @xmath59 galaxies in the cosmos field ( closed squares ; takahashi et al .
2007 ) and the sdf ( closed triangles ; takahashi et al .
2007 ) are also shown for comparison . ]
next , we model each galaxy in the mock catalog as a round disk and randomly place the galaxy within @xmath37 of a qso sightline and at redshift between @xmath7 and @xmath38 .
we compute the corresponding physical projected distance of the galaxy based on the redshift and angular distance to the qso . for each disk
, we also assign a random inclination angle with respect to the observer and a random position angle of the major axis of the inclined disk with respect to the line connecting the galaxy and the qso . to determine the fraction of the luminous disk that overlaps the sdss fiber centered at the qso
, we adopt the size of each disk galaxy @xmath54 and its relative distance and orientation to the qso .
we de - project the inclined disk and determine the galactocentric distance @xmath60 of each point ( @xmath61 , @xmath62 ) in the disk according to @xmath63\ ] ] where @xmath64 is the projected radius on the plane of the sky , @xmath43 is the position angle of the major axis of the inclined disk , @xmath65 is the azimuthal angle of point ( @xmath61 , @xmath62 ) from the major axis , and @xmath66 is the inclination angle of the disk .
the expected [ oii ] emission from the galaxy in the qso spectral data is then computed by integrating all the light within @xmath54 of the disk that falls in the @xmath3 diameter fiber .
we consider two different surface brightness profiles : ( 1 ) a flat , top - hat profile and ( 2 ) an exponential profile for this calculation .
if the luminous disk does not overlap the @xmath3 diameter fiber , then we set @xmath1}=0 $ ] .
finally , we divide the computed @xmath1}$ ] by the physical area of a @xmath3 diameter fiber at the redshift of the mock galaxy in order to calculate its [ oii ] luminosity surface density @xmath15}}$ ] that would be recorded in an sdss qso fiber . the mock mgii catalog is formed by calculating the expected mgii absorption strength in the spectrum of the background qso for every galaxy in the mock galaxy catalog . to determine the associated mgii absorption strength of a galaxy in the qso spectrum
, we adopt the uniform gaseous halo model of chen et al .
( 2010a ) . under this model ,
the mgii absorber strength for a galaxy at projected distance @xmath2 is characterized by a mean relation of @xmath67 and dispersion @xmath68 , where @xmath69 is the gaseous radius of the mgii halo , @xmath70 is the core radius and is @xmath71 and @xmath72 . the gaseous radius is determined based on the galaxy @xmath29-band luminosity following a power - law model , @xmath73 for which chen et al .
( 2010a ) found a best - fit characteristic radius of @xmath74 kpc for an @xmath75 galaxy and a scaling index of @xmath76 .
the scaling relation between gaseous radius and galaxy @xmath29-band luminosity is understood as more massive galaxies are surrounded by more extended halos ( e.g. tinker & chen 2008 ; chen et al .
in addition to the spatial profile of the extended mgii gas , chen et al .
( 2010a ) also measured a high gas covering fraction @xmath77 within @xmath69 .
specifically , they found a mean covering fraction of @xmath78% for absorbers of @xmath79 at @xmath80 and @xmath81% at @xmath82 .
the gas covering fraction increases with decreasing threshold and decreases with @xmath2 ( see figure 10 in chen et al .
2010a ) . for each simulated galaxy
, we compute the expected by randomly sampling the mean relation equations ( 6 ) within the observed scatter @xmath83 . for simulated dwarf galaxies with @xmath84 , we set @xmath85 . including the large scatter ( @xmath83 ) in the computation of allows the possibility of galaxies at small @xmath2 producing mgii absorbers that are much weaker than the mean .
although every galaxy at @xmath86 in the mock sample is expected to produce a mgii absorber , the gas covering fraction measured at a given threshold is not 100% . for a mock sample of 200,000 galaxies ,
the procedure described above produces a mock sample of @xmath87 mgii absorbers of @xmath88 at @xmath12 .
the frequency distribution function of the mock mgii absorber sample is presented in figure 3 . for comparison
, we also include in figure 3 the model from prochter et al .
( 2006 ) that best describe the incidence of mgii absorbers identified at @xmath89 in their blind survey . aside from the offset in the normalization ,
figure 3 shows that the frequency distribution function of the mock mgii absorber sample follows the shape of the best - fit model of prochter et al .
the difference in the incidence of mock mgii absorbers ( up to 20% ) for different adopted faint - end slope of the galaxy luminosity function is understood by the relatively steep dependence of halo gas cross section on galaxy luminosity , @xmath90 from equation ( 7 ) .
adopting known galaxy luminosity functions at @xmath32 from faber et al .
( 2007 ) and the scaling relation of equation ( 7 ) , we expect to find ( following equation 8 below ) a number density of @xmath91 per line of sight for mgii absorbers of @xmath92 .
our model therefore well reproduces the observe absorber statistics at @xmath32 ( e.g. nestor et al .
2005 ; prochter et al .
2006 ) . , @xmath93 , and @xmath94 , respectively . for comparison
, we also include the observed absorber frequency distribution function from prochter et al .
( 2006 ) for mgii absorbers found at @xmath89 . ]
using the mock galaxy and mgii absorber catalogs , we proceed to examine the relation between and @xmath1}$ ] of mgii absorbers . for each mgii absorber in the mock catalog , we know from our simulation the location and luminosity of the absorbing galaxy .
we have also calculated according to the procedures described in section 2.1 the fraction of galaxy light that would be recorded in the sdss qso fibers .
we can therefore directly compare our simulation data with the observations of m11 .
we present in figure 4 the distribution of and @xmath15}}$ ] of the mock mgii absorber sample .
the top panel shows the @xmath14}}$ ] distribution of individual mgii absorbers , including detections ( dots ) and non - detections ( indicated by the arrows ) .
the bottom panel shows the fraction of mgii absorbers for which the absorbing galaxies are not expected to overlap the fiber and therefore have @xmath14}}=0 $ ] in the qso spectra .
it is clear that a growing fraction of galaxy light is missed in the qso spectra for absorbers of decreasing strength due to the vs. @xmath2 anti - correlation ( e.g. chen et al .
2010 ) . to reproduce the observations in sdss qso spectra
, we divide the mock catalog of @xmath95 mgii absorbers into subsamples according to their absorption strengths . following the proceduce described in m11 , who formed a median qso spectrum at the rest - frame of the mgii absorbers and measured the associated [ oii ] emission line flux , we compute the median value of [ oii ] luminosity surface density @xmath96}}\rangle_{\rm med}$ ] for all mgii absorbers in each bin , including those with @xmath97}}=0 $ ] . to estimate the scatter , we repeat the monte carlo simulation 100 times to generate 100 mock samples of galaxies and mgii absorbers .
we measure the 1-@xmath52 dispersion in @xmath96}}\rangle_{\rm med}$ ] among the 100 mock samples .
the stars and the associated errorbars in figure 4 represent the computed @xmath96}}\rangle_{\rm med}$ ] and the associated dispersion in each bin . for comparison ,
the best - fit power - law model of the observed @xmath96}}\rangle_{\rm med}$ ] correlation ( equation 1 ) from m11 is included in figure 4 as the dash - dotted line .
the m11 model is found to match well with the simulated data .
however , both fall below the locus defined by individual dots because of non - detections quantified in the bottom panel .
we also calculate the mean [ oii ] flux @xmath98}}$ ] expected in the stacked sdss qso spectra , averaged over all mgii absorbers in each bin including non - detections .
the results are shown in solid circles in figure 4 . } } $ ] distribution of the mock galaxy and absorber samples , in comparison to observations of m11 .
the dots in the top panel represent individual mgii absorbers with detectable emission fluxes in the qso spectra from their absorbing galaxies . recall that these are calculated by dividing the expected [ oii ] line flux in the fiber with the corresponding physical area of the @xmath3 fiber at the redshift of the absorber .
the arrows at the bottom row indicate the presence of absorbing galaxies with @xmath14}}=0 $ ] ( non - detections ) , because their optical disks do not overlap with the fiber . for completeness
, the bottom panel shows the fraction of mgii absorbers that have @xmath14}}=0 $ ] .
the ( red ) stars in the top panel represent the _ median _ values of @xmath99}}$ ] for _ all _ mgii absorbers ( including detections and non - detections ) in each bin with the 1-@xmath52 dispersion indicated by the errorbars .
the dash - dotted line shows the best - fit power - law model ( equation 1 ) of the observed @xmath96}}\rangle_{\rm med}$ ] correlation from m11 .
the solid ( blue ) circles represent the _ mean _ @xmath14}}$ ] averaged over all mgii absorbers in each @xmath0 bin .
note that the stars fall below the locus defined by individual dots because of non - detections quantified in the top panel . ]
we have carried out a monte carlo study to investigate the effect of differential aperture loss of extended emission from intervening galaxies in sdss fibers .
we generate a mock galaxy sample based on known empirical relations that describe the luminosity and size distribution of the general galaxy population uncovered in deep surveys .
the mock galaxy sample is accompanied by a mock mgii absorber sample that is generated based on a simple assumption that extended gaseous halos are a common and generic feature of distant galaxies and the gaseous extent scales proportionally with galaxy mass . for each absorber in the mock sample ,
the luminosity and projected distance of the absorbing galaxy are known , allowing us to make predictions for the observed relation between @xmath0 and the associated @xmath1}$ ] , in sdss qso spectra .
our study shows that combining the known vs. @xmath2 anti - correlation of mgii galaxy pairs ( e.g. chen et al .
2010a ) and differential fiber selection of mgii absorbing galaxies ( figure 1 ) reproduces the observed vs. @xmath1}$ ] correlation in the sdss qso spectra without additional fine - tuning or scaling .
the results of our study indicate that the observed vs. @xmath30}$ ] correlation of mgii absorbers in sdss data is likely due to a differential fiber selection bias and does not provide a physical understanding of the origin of the mgii absorber population . on the basis of
the the observed vs. @xmath1}$ ] correlation , m11 further attempted to draw connections between the number density evolution of mgii absorbers @xmath100 and the cosmic star formation history of the universe as characterized by the comoving [ oii ] luminosity density @xmath101 . given a good agreement between @xmath100 and @xmath101 , the authors argue that outflows are the mechanism responsible for mgii absorption .
the monte carlo simulation presented in section 2 also allows us to address the agreements in these measurements under a simple , generic halo model .
it is clear from figures 3 & 4 show that both the frequency distribution function of mgii absorbers and the galaxy [ oii ] luminosity function are well reproduced in our mock galaxy and absorber samples with no preference in starburst systems .
although under the simple halo model mgii absorbers do not directly probe star - forming disks like high column density damped absorbers ( e.g. wild et al .
2007 ) , they trace the halo gas of distant galaxies .
if more massive galaxies are surrounded by more extended mgii absorbing gas ( equation 7 ) , then @xmath100 is calculated according to @xmath102 where @xmath103 is the speed of light , @xmath104 is the galaxy luminosity function , @xmath105 is the incidence of extended gas that is the product of halo gas covering fraction and the fraction of galaxies with extended gaseous halos , and @xmath106 is the cross section of the gaseous halo . as mentioned in 2.2 ,
the mean covering fraction of mgii absorbing gas is found to be high , roughly 100% for absorbers of @xmath79 at @xmath80 ( chen et al .
2010a ) . such high covering fraction is consistent with the analysis presented in kacprzak et al .
( 2008 ) for a scaling index of @xmath76 .
combining equations ( 4 ) and ( 7 ) yields @xmath107}^{0.93}*}\,l_{\rm [ o\,ii]}^{0.93 } \approx \frac{\kappa_{\rm
gas}\,(\pi\,r_{\rm gas}^2*)}{l_{\rm [ o\,ii]}*}\,l_{\rm [ o\,ii]}.\ ] ] substituting equation ( 9 ) into equation ( 8) leads to @xmath108}*}\int l\,\phi(l , z)\,d\,l \nonumber \\ & = & \frac{\kappa_{\rm gas}\,(\pi\,r_{\rm gas}^2*)}{\varepsilon(z)\,l_{\rm [ o\,ii]}*}\,\ell_{\rm [ o\,ii]}(z),\end{aligned}\ ] ] where @xmath109 is defined so that @xmath110 . the same redshift dependent factor is defined in equation ( 7 ) of m11 .
equation ( 10 ) shows that the number density evolution of absorbers naturally follows the comoving [ oii ] luminosity density , @xmath111}(z)$ ] , if extended gaseous halos are a common and generic feature of field galaxies .
one can therefore apply @xmath100 for constraining the cosmic star formation history , or apply known co - moving luminosity density for constraining the fraction of absorbers originating in galactic halos ( chen et al .
2000 ) . in practice ,
as illustrated in m11 the sdss @xmath3 fibers define a survey volume of [ oii ] emitting galaxies .
the comoving [ oii ] luminosity density , @xmath111}(z)$ ] , can be estimated according to @xmath112}(z ) & \equiv & \frac{\delta\,l_{\rm [ o\,ii]}}{\delta\,v_c } \nonumber \\ & = & \frac{1}{(1+z)^2}\frac{1}{d_a^2(z)\,d\,\omega}\frac{\delta\,l_{\rm [ o\,ii]}}{\delta\,n}\frac{\delta\,n}{\delta\,r } \nonumber \\ & = & \frac{\delta\,z}{(1+z)^2\,\delta\,r}\frac{\bar{l}_{\rm [ o\,ii]}}{d_a^2(z)\,d\,\omega}\frac{\delta\,n}{\delta\,z } \nonumber \\ & = & \varepsilon(z)\,\bar{\sum}_{l_{\rm [ o\,ii]}}\,n(z),\end{aligned}\ ] ] where @xmath113}$ ] is the total [ oii ] luminosity observed in a coadded qso spectrum , @xmath114 is the total number of mgii absorbers over the survey redshift pathlength @xmath115 , and @xmath116}$ ] is the mean [ oii ] luminosity averaged over the entire galaxy ( absorber ) population .
it is clear that equation ( 11 ) resembles equation ( 10 ) and that @xmath111}(z)$ ] can be estimated based on the product of @xmath100 and the mean [ oii ] luminosity surface density averaged over _ all _ mgii absorbers .
we present in figure 4 the mean [ oii ] luminosity surface density @xmath98}}$ ] measured in the @xmath3 sdss fibers and averaged over the number mgii absorbers in each bin ( solid circles ) .
the mean values can be characterized by @xmath117}}=a_{3''}\left[\frac{{\mbox{$w_r(2796)$}}}{1\,{\rm \aa}}\right]^{b_{3''}}\ ] ] where @xmath118 and @xmath119 ( solid line in figure 4 ) .
this correlation applies to a @xmath3 survey cylinder at @xmath12 . note that
because of the extreme bimodal distribution of @xmath15}}$ ] in which a large fraction of absorbers have @xmath14}}=0 $ ] , the median value @xmath120}}\rangle_{\rm med}$ ] is expected to be significantly smaller than the mean @xmath98}}$ ] . while equation ( 11 ) provides a useful tool for estimating the co - moving star formation rate density @xmath20 based on the observations of mgii absorbers in sdss qso spectra ( as shown cleverly in m11 ) , we note that the observed strong vs.@xmath14}}$ ] correlation in sdss data is shaped primarily by a differential fiber loss of the observed [ oii ] flux and therefore unphysical .
to illustrate the fiber selection bias in driving the apparent correlation in figure 4 , we repeat the monte carlo analysis described in 2 for @xmath121 diameter fibers .
this is five times the aperture size of sdss spectra . at @xmath122 ,
such a large aperture covers an area of roughly @xmath123 kpc projected radius from qso lines of sight , sufficient to cover the majority of the light from galaxies producing strong mgii absorbers ( @xmath124 ) .
figure 5 shows the distribution of and @xmath15}}$ ] measured over @xmath121 diameter fibers for the mock mgii absorber sample .
the mean [ oii ] flux @xmath125}}$ ] share a similar slope as the median [ oii ] flux @xmath96}}\rangle_{\rm med}$ ] within the @xmath121 diameter aperture and the slope is consistent with @xmath126 ( solid line in figure 5 ) .
fibers in the monte carlo simulation described in 2 . both @xmath127}}\rangle_{\rm med}$ ]
( red stars ) and @xmath128}}$ ] ( solid blue circles ) show a lack of correlation with .
the bottom panel shows that with @xmath129 larger fibers , differential aperture loss has minimal impact on the survey of mgii selected galaxies along qso lines of sight . ] in summary , we have demonstrated that ( 1 ) because of the observed @xmath0 vs. @xmath2 anti - correlation ( e.g. chen et al .
2010a ) , galaxies at larger projected distances produce on average weaker mgii absorbers and weaker ( or zero ) @xmath1}$ ] in sdss qso spectra and that ( 2 ) because of the extreme bimodal distribution of @xmath14}}$ ] in which a large fraction of absorbers have @xmath14}}=0 $ ] , the median value @xmath120}}\rangle_{\rm med}$ ] can be significantly smaller than the mean @xmath98}}$ ] .
together these effects strenthen the apparent @xmath0@xmath96}}\rangle_{\rm med}$ ] correlation of m11 .
consequently , empirical correlations between star - forming properties of galaxies and statistical properties of qso absorbers do not provide the unambiguous evidence necessary to discriminate between infalling clouds and outflows as the mechanism for producing mgii absorbing clouds at @xmath130 kpc of a galaxy .
while outflows are a natural product of starbursts , gas accretion provides the fuels for star formation in galaxies . as discussed by previous authors ,
mass is likely a more fundamental factor that determines the properties of gaseous halos around galaxies ( e.g. steidel et al .
1994 ; ledoux et al .
2006 ; chen et al .
2010b ) and more massive galaxies can support more extended gaseous halos .
given that models that do not require outflows can also reproduce the empirical correlations between absorber and star formation properties , we caution drawing conclusions in favor of outflows based on simple correlations between absorber and star formation properties .
we thank jean - ren gauthier , michael rauch , rob simcoe , ben weiner , vivienne wild , and art wolfe for helpful discussions and comments . g.l .
acknowledges support from the physics reu program at the university of chicago . | this paper investigates the effect of differential aperture loss with sdss fibers and examines whether such selection bias would result in the observed correlation between rest - frame absorption equivalent width of mgii absorbers , @xmath0 , and mean associated [ oii ] luminosity , @xmath1}$ ] , in sdss qso spectra .
we demonstrate based on a monte carlo simulation that the observed vs. @xmath1}$ ] correlation of mgii absorbers can be well - reproduced , if all galaxies found in deep surveys possess extended mgii halos and if the extent of mgii halos scales proportionally with galaxy mass as shown in previous studies .
the observed correlation can be explained by a combination of ( 1 ) the known vs. @xmath2 anti - correlation in galaxy and mgii absorber pairs and ( 2 ) an increasing aperture loss in the @xmath3 diameter sdss fiber for galaxies at larger @xmath2 .
galaxies at larger projected distances produce on average weaker mgii absorbers and weaker ( or zero ) @xmath1}$ ] in sdss qso spectra .
we show that such correlation diminishes when larger fibers are adopted and is therefore not physical . while under a simple halo model the majority of mgii absorbers do not directly probe star - forming disks , they trace photo - ionized halo gas associated with galaxies .
we show that because of the scaling relation between extended gas cross - section and galaxy mass , the number density evolution of the mgii absorber population as a whole provides a good measure of the cosmic star formation history .
galaxies : halos galaxy : star formation quasars : absorption lines |
[ intro ] heavy quarks ( charm and beauty ) are essential probes of the properties of the medium created in heavy - ion collisions , since they are produced in the early stage of hadronic collisions via scattering processes with large momentum transfer .
heavy - flavour production has been measured via semi - electronic and semi - muonic decays , as well as fully reconstructed d mesons , in pp , p pb and pb pb collisions with alice @xcite at the lhc . in central pb
pb collisions at @xmath1 = 2.76 tev , a strong suppression of high transverse momentum ( @xmath4 ) d mesons and electrons ( muons ) from heavy - flavour hadron decays was observed at mid- ( forward ) rapidity @xcite .
this suppression is interpreted as an effect of parton energy loss in the medium created in heavy - ion collisions .
however , a quantitative understanding of the pb - pb results requires cold nuclear matter ( cnm ) effects to be taken into account , which can be accessed by studying p pb collisions @xcite assuming that a hot and dense extended system is not formed in such collisions .
cnm effects are studied using the nuclear modification factor @xmath0 , defined as @xmath5 where @xmath6 is the average nuclear overlap function estimated through the glauber model , which gives @xmath7 @xcite ; @xmath8 ( @xmath9 ) is the @xmath4-differential yield ( cross section ) in p
pb collisions and @xmath10 is the @xmath4-differential cross section in pp collisions ; a is the mass number of the pb nucleus .
the value of @xmath0 is unity in absence of nuclear effects .
alice is the dedicated heavy - ion experiment at the lhc .
it allows to investigate heavy - flavour production in several decay channels over a wide rapidity and transverse momentum range .
the alice detector consists of a set of central barrel detectors ( @xmath11 ) , a muon spectrometer ( @xmath12 ) and global detectors for triggering and event characterization purposes . at mid - rapidity ,
the inner tracking system ( its ) and the time projection chamber ( tpc ) provide track reconstruction down to very low transverse momentum ( @xmath13100 mev/@xmath3 ) with a momentum resolution better than 4@xmath14 for @xmath15 gev/@xmath3 , as well as good impact parameter ( distance of closest approach of the track to the primary interaction vertex ) resolution @xcite .
d mesons are reconstructed via their hadronic decay channels and electrons from semileptonic decays of charm and@xmath16or beauty hadrons are measured . at forward rapidity , muons from heavy - flavour hadron decays can be measured , since the alice muon spectrometer allows to identify muons by requiring that a track reconstructed in the tracking system is matched with the corresponding track candidate in the muon trigger system .
d mesons are reconstructed from their hadronic decay channels @xmath17 @xmath18 @xmath19 ( br = 3.88@xmath14 ) , @xmath20 @xmath18 @xmath21 ( br = 9.13@xmath14 ) , @xmath22 @xmath18 @xmath23 @xmath18 @xmath21 ( br = 67.7@xmath14 ) and @xmath24 @xmath18 @xmath25 @xmath18 @xmath26 ( br = 2.28@xmath14 ) . the selection of d meson candidates against the large combinatorial background is based on the reconstruction of decay vertices displaced by a few hundred @xmath27 from the interaction vertex , exploiting the @xmath28 of @xmath17 , @xmath20 and @xmath24 , which is about @xmath29 depending on the d - meson species . in order to further enhance the ratio between the d - meson signal and the combinatorial background , the measurements of the particle time - of - flight from the collision point to the time of flight ( tof ) detector and of the specific energy loss in the tpc gas are used to identify @xmath30 and @xmath31 @xcite .
the measurement is performed in the rapidity interval @xmath32 ( which is specific for p - pb collisions ) over a wide transverse momentum range .
the electrons are identified with tpc and tof at low @xmath4 , as well as the tpc and electromagnetic calorimeter ( emcal ) at high @xmath4 .
the measurement of electrons from heavy - flavour hadron decays at mid - rapidity ( @xmath33 ) requires the subtraction of contributions from several background sources from the inclusive electron distribution .
the dominant contribution is from photon conversions in the detector material and dalitz decays of light neutral mesons ( @xmath34 and @xmath35 , mainly ) .
these contributions are statistically subtracted by using the cocktail ( i.e. a calculation of the background based mainly on measured @xmath4-differential cross sections of the main electron sources ) and invariant mass ( i.e. the measurement of electrons from photon conversions and dalitz decays via low - mass electron - positron pairs ) methods @xcite . with the high spatial resolution of the track impact parameter measurement
, one can isolate the beauty - decay contribution thanks to the much larger lifetime of beauty hadrons ( @xmath36 ) compared to that of charm hadrons and other background sources .
the electrons from their semileptonic decays have , consequently , a larger average impact parameter with respect to the primary interaction vertex .
muons from heavy - flavour hadron decays have been measured at forward ( i.e. proton - beam direction , @xmath37 ) and backward ( i.e. pb - beam direction , @xmath38 ) rapidity , respectively , by analyzing data collected with different beam configurations .
the main source of background in the @xmath4-differential inclusive spectrum consists of muons from light hadron decays ( @xmath31 and @xmath30 , mainly ) .
this contribution dominates the muon yield for @xmath39 gev/@xmath3 and prevents a measurement of muons from heavy - flavour decays at low @xmath4 . at larger @xmath4 the contamination
is estimated via monte - carlo simulations , as well as with a data - driven method based on the extrapolation of charged hadron yields @xcite measured at mid - rapidity with alice to forward rapidity .
the nuclear modification factors , @xmath0 , of @xmath17 , @xmath20 , @xmath22 and @xmath24 mesons are consistent with each other , and they are compatible with unity within uncertainties @xcite .
the pp reference cross section at @xmath40 = 5.02 tev is obtained by a pqcd - based energy scaling @xcite of the @xmath4-differential cross sections measured at @xmath40 = 7 tev @xcite .
figure [ fig : dmeson ] ( left ) presents the average @xmath0 of prompt @xmath17 , @xmath20 and @xmath22 in @xmath41 gev/@xmath3 .
the @xmath0 can be described by means of next - to - leading - order ( nlo ) pqcd @xcite calculations , including the eps09 @xcite nuclear modification of the cteq6 m @xcite parton distribution functions ( pdf ) and calculations based on the color glass condensate ( cgc @xcite ) initial - state prescription .
the data are also well described by calculations which include energy loss in cold nuclear matter , nuclear shadowing and @xmath42-broadening @xcite .
as demonstrated in figure [ fig : dmeson ] ( right ) , cold nuclear matter effects are small for @xmath433 gev/@xmath3 .
this indicates that the suppression of d mesons observed in the 20@xmath14 most central pb
pb collisions at @xmath1 = 2.76 tev @xcite results from an effect due to the presence of the hot and dense medium .
the @xmath0 of electrons from heavy - flavour hadron decays in p
pb collisions at @xmath1 = 5.02 tev is shown in figure [ fig : electron ] ( left ) .
the pp reference is obtained following a strategy similar to the one used for the d - meson analysis .
the @xmath0 is compatible with unity , showing that cold nuclear matter effects are small .
the measurement is compared with a pqcd calculation with the eps09 @xcite nlo parameterization of nuclear pdfs . within uncertainties ,
the data can be described by the model predictions .
figure [ fig : electron ] ( right ) displays the measured @xmath0 of electrons from heavy - flavour hadron decays together with that of electrons from beauty - hadron decays .
the @xmath0 of electrons from beauty - hadron decays is consistent with unity within uncertainties , and it is also compatible with that of electrons from heavy - flavour hadron decays .
the data indicate that the suppression of electrons from heavy - flavour hadron decays observed in the 10@xmath14 most central pb
pb collisions at @xmath1 = 2.76 tev @xcite results from a hot medium effect . of
prompt @xmath17 , @xmath20 and @xmath22 mesons as a function of @xmath4 compared to model calculations @xcite ; right : average @xmath0 of prompt @xmath17 , @xmath20 and @xmath22 mesons as a function of @xmath4 compared to the prompt d - meson @xmath44 in central ( 0 - 20@xmath14 ) and semi - peripheral ( 40 - 80@xmath14 ) pb pb collisions at @xmath1 = 2.76 tev @xcite.,title="fig:",scaledwidth=38.0% ] of prompt @xmath17 , @xmath20 and @xmath22 mesons as a function of @xmath4 compared to model calculations @xcite ; right : average @xmath0 of prompt @xmath17 , @xmath20 and @xmath22 mesons as a function of @xmath4 compared to the prompt d - meson @xmath44 in central ( 0 - 20@xmath14 ) and semi - peripheral ( 40 - 80@xmath14 ) pb pb collisions at @xmath1 = 2.76 tev @xcite.,title="fig:",scaledwidth=38.0% ] of electrons from heavy - flavour hadron decays as a function of @xmath4 compared to model calculations @xcite ; right : @xmath0 of electrons from heavy - flavour hadron decays compared to @xmath0 of electrons from beauty - hadron decays.,title="fig:",scaledwidth=40.0% ] of electrons from heavy - flavour hadron decays as a function of @xmath4 compared to model calculations @xcite ; right : @xmath0 of electrons from heavy - flavour hadron decays compared to @xmath0 of electrons from beauty - hadron decays.,title="fig:",scaledwidth=40.0% ] the nuclear modification factor of muons from heavy - flavour hadron decays , shown in figure [ fig : muon ] ( left ) , is measured in the center of mass rapidity domains @xmath37 ( forward rapidity ) and @xmath38 ( backward rapidity ) .
the pp reference cross section of heavy - flavour hadron decay muons at @xmath40 = 5.02 tev is obtained via a pqcd - based energy scaling procedure @xcite ( i.e. calculated taking as input the @xmath4-differential cross sections measured at @xmath40 = 7 tev @xcite ) , performed at forward and backward rapidity , respectively .
this procedure is used to obtain the @xmath4-differential cross section up to 12 gev/@xmath3 . in order to measure the @xmath0 in a wider transverse momentum range ,
the @xmath4-differential cross section in pp collisions at @xmath40 = 5.02 tev was extrapolated to higher @xmath4 using the spectrum predicted by fonll scaled to match the obtained pp reference in the range @xmath45 gev/@xmath3 .
the @xmath0 measured at forward rapidity is consistent with unity within uncertainties .
@xmath0 at backward rapidity is slightly larger than unity in the range @xmath2 gev/@xmath3 and close to unity at higher @xmath4 .
this supports the previous observation that the suppression measured in the 10@xmath14 most central pb
pb collisions at @xmath1 = 2.76 tev @xcite results from hot and dense medium effect .
another way to quantify cold nuclear matter effects is to study the forward - to - backward ratio defined as the cross section measured at forward rapidity with respect to the one at backward rapidity , the main advantage of using such a ratio is that the pp reference and the nuclear overlap function cancel .
the drawback of this approach is the limited statistics , because the common @xmath46 interval covered at both forward and backward rapidity is small ( @xmath47 units ) .
@xmath48 the measured @xmath49 is presented in figure [ fig : muon ] ( right ) .
it is systematically smaller than unity in the range @xmath2 gev/@xmath3 and close to unity at higher @xmath4 .
this is well reproduced by a pqcd calculation including the eps09 @xcite nlo parameterization of nuclear pdfs @xcite in the range @xmath50 gev/@xmath3 .
the @xmath0 measured at backward rapidity is slightly underestimated by such theoretical calculations at low @xmath4
. of muons from heavy - flavour hadron decays as a function of @xmath4 at forward rapidity ( @xmath37 ) and backward rapidity ( @xmath38 ) , compared to the @xmath44 for muons from heavy - flavour hadron decays in the 10@xmath14 most central pb
pb collisions at @xmath1 = 2.76 tev @xcite ; right : forward - to - backward ratio @xmath49 of muons from heavy - flavour hadron decays compared to a model calculation @xcite.,title="fig:",scaledwidth=45.0% ] of muons from heavy - flavour hadron decays as a function of @xmath4 at forward rapidity ( @xmath37 ) and backward rapidity ( @xmath38 ) , compared to the @xmath44 for muons from heavy - flavour hadron decays in the 10@xmath14 most central pb
pb collisions at @xmath1 = 2.76 tev @xcite ; right : forward - to - backward ratio @xmath49 of muons from heavy - flavour hadron decays compared to a model calculation @xcite.,title="fig:",scaledwidth=45.0% ]
the nuclear modification factor of heavy - flavour production has been measured in a wide rapidity and transverse momentum range , as well as in several decay channels , in p
pb collisions at @xmath1 = 5.02 tev with alice .
the @xmath0 results provide evidence that cold nuclear matter effects are small , showing that the strong suppression observed in central pb
pb collisions at @xmath1 = 2.76 tev is a final state effect due to in - medium parton energy loss .
00 k. aamodt _ et al . _
[ alice collaboration ] , jinst * 3 * , s08002 ( 2008 ) b. abelev _ et al . _
[ alice collaboration ] , jhep * 09 * , 112 ( 2012 ) e. pereira de oliveira filho for the alice collaboration , arxiv:1404.3983 b. abelev _ et al . _
[ alice collaboration ] , phys .
lett . * 109 * , 112301 ( 2012 ) b. abelev _ et al . _
[ alice collaboration ] , arxiv:1405.3452 b. abelev _ et al . _
[ alice collaboration ] , jhep * 02 * , 073 ( 2014 ) b. abelev _ et al . _
[ alice collaboration ] , phys .
lett . * 110 * , 082302 ( 2013 ) b. abelev _ et al . _
[ alice collaboration ] , j.phys .
g * 30 * , 1517 ( 2004 ) , j.phys .
g * 32 * , 1295 ( 2006 ) a. rossi for the alice collaboration , to be published in nuclear physics : proceedings of hard probe conference a. o. velasquez for the alice collaboration , to be published in nuclear physics : proceedings of hard probe conference r. averbeck , n. bastid , z. conesa del valle , p. crochet , a. dainese , x. zhang , arxiv:1107.3243 b. abelev _ et al . _
[ alice collaboration ] , jhep * 01 * , 112 ( 2012 ) m. cacciari _ et al .
_ , jhep * 10 * , 137 ( 2012 ) m. l. mangano _ et al .
_ , nucl .
phys . b * 373 * , 295 ( 1992 ) k. eskola _ et al . _
jhep * 0904 * , 065 ( 2012 ) d. stump , j. huston , j. pumplin , w. k. tung , h. l. lai , s. kuhlmann and j. f. owens , jhep * 03 * , 046 ( 2003 ) h. fujii and k. watanabe , nucl .
a * 920 * , 78 ( 2013 ) and nucl .
phys . a * 915 * , 1 ( 2013 ) r. sharma _ et al .
_ , phys .
rev . c * 80 * , 054902 ( 2009 ) b. abelev _ et al . _
[ alice collaboration ] , phys .
b * 708 * , 265 ( 2012 ) | the heavy - flavour nuclear modification factor @xmath0 has been measured with the alice detector in p
pb collisions at the nucleon nucleon center of mass energy @xmath1 = 5.02 tev at the cern lhc in a wide rapidity and transverse momentum range , as well as in several decay channels .
@xmath0 is consistent with unity within uncertainties at mid - rapidity and forward rapidity . in the backward region
a slight enhancement of the yield of heavy - flavour decay muons is found in the region @xmath2 gev/@xmath3 .
the results are described within uncertainties by theoretical calculations that include initial - state effects .
the measurements add experimental evidence that the suppression of heavy - flavour production observed at high @xmath4 in central pb
pb collisions with respect to pp collisions is due to a medium effect induced by the interaction of heavy quarks with the partonic matter .
alice , heavy - flavour production , in - medium effect , nuclear modification factor , cold nuclear matter effects |
matsumoto @xcite proved that every lefschetz fibration of genus one over a closed oriented surface is isomorphic to a fiber sum of copies of a holomorphic elliptic fibration on @xmath1 and a trivial torus bundle over the surface if it has at least one critical point .
this result played a crucial role in completing the classification of diffeomorphism types of elliptic surfaces ( see gompf and stipsicz ( * ? ? ?
* section 8.3 ) ) .
although such a classification has not been established for lefschetz fibrations of higher genus , auroux @xcite proved a stabilization theorem for lefschetz fibrations of genus two , which states that every lefschetz fibration of genus two over the @xmath0sphere becomes isomorphic to a fiber sum of copies of three typical fibrations after fiber summing with a holomorphic fibration on @xmath2 .
auroux @xcite gave a generalization of this theorem for lefschetz fibrations of higher genus , which states that two lefschetz fibrations of the same genus over the @xmath0sphere which have the same signature , the same numbers of singular fibers of each type , and admit sections of the same self - intersection number become isomorphic after fiber summing the same number of copies of a ` universal ' lefschetz fibration .
kamada @xcite introduced charts , which are labeled finite graphs in a disk , to describe monodromies of surface braids ( see also a textbook @xcite of kamada ) .
kamada , matsumoto , matumoto , and waki @xcite considered a variant of chart for lefschetz fibrations of genus one to give a remarkably simple proof of the above result of matsumoto .
furthermore kamada @xcite , and endo and kamada @xcite made use of generalized charts to give a simple proof of the above theorem of auroux for lefschetz fibrations of genus two , and to investigate a stabilization theorem and an invariant for hyperelliptic lefschetz fibrations of arbitrary genus .
see also baykur and kamada @xcite , and hayano @xcite for applications of charts to broken lefschetz fibrations . in this paper
we introduce a chart description for lefschetz fibrations of genus greater than two over closed oriented surfaces of arbitrary genus to show a signature formula and two theorems on stabilization for such fibrations . in section 2
we introduce charts and chart moves with respect to wajnryb s presentation of mapping class groups to examine monodromies of lefschetz fibrations . after a short survey of meyer s signature cocycle
, we generalize a signature formula @xcite of endo and nagami for lefschetz fibrations over the @xmath0sphere to that for lefschetz fibrations over a closed oriented surface of arbitrary genus in section 3 .
section 4 is devoted to proofs of two theorems on stabilization of lefschetz fibrations under fiber summing with copies of a ` universal ' lefschetz fibration .
in particular the first of our stabilization theorems is a generalization of the theorem of auroux @xcite .
we make several comments on variations of chart description and propose some possible directions for future research in section 5 .
in this section we review a definition and properties of lefschetz fibrations and introduce a chart description for lefschetz fibrations of genus greater than two . in this subsection
we review a precise definition and basic properties of lefschetz fibrations .
more details can be found in matsumoto @xcite and gompf and stipsicz @xcite .
let @xmath3 be a connected closed oriented surface of genus @xmath4 .
[ lf ] let @xmath5 and @xmath6 be connected closed oriented smooth @xmath7manifold and @xmath0manifold , respectively .
a smooth map @xmath8 is called a _ lefschetz fibration _ of genus @xmath4 if it satisfies the following conditions : \(i ) the set @xmath9 of critical values of @xmath10 is finite and @xmath10 is a smooth fiber bundle over @xmath11 with fiber @xmath3 ; \(ii ) for each @xmath12 , there exists a unique critical point @xmath13 in the _ singular fiber _ @xmath14 such that @xmath10 is locally written as @xmath15 or @xmath16 with respect to some local complex coordinates around @xmath13 and @xmath17 which are compatible with orientations of @xmath5 and @xmath6 ; \(iii ) no fiber contains a @xmath18sphere .
we call @xmath5 the _ total space _ ,
@xmath6 the _ base space _ , and @xmath10 the _
projection_. we call @xmath13 a critical point of _ positive type _ ( resp . of _
negative type _ ) and @xmath19 a singular fiber of _ positive type _ ( resp . of _ negative type _ ) if @xmath10 is locally written as @xmath15 ( resp .
@xmath20 ) in ( ii ) . for a regular value @xmath21 of @xmath10 , @xmath22
is often called a _ general fiber_. a lefschetz fibration in this paper
is called an _
achiral _ lefschetz fibration in many other papers .
let @xmath8 and @xmath23 be lefschetz fibrations of genus @xmath4 over the same base space @xmath6 .
we say that @xmath10 is _ isomorphic _ to @xmath24 if there exist orientation preserving diffeomorphisms @xmath25 and @xmath26 which satisfy @xmath27 .
if we can choose such an @xmath28 isotopic to the identity relative to a given base point @xmath29 , we say that @xmath10 is _ strictly isomorphic _ to @xmath24 .
let @xmath30 be the mapping class group of @xmath3 , namely the group of all isotopy classes of orientation preserving diffeomorphisms of @xmath3 .
we assume that @xmath30 acts on the _ right _ : the symbol @xmath31 means that we apply @xmath32 first and then @xmath33 for @xmath34 .
we denote the mapping class group of @xmath3 acting on the left by @xmath35 .
hence the identity map @xmath36 is an anti - isomorphism .
let @xmath8 be a lefschetz fibration of genus @xmath4 as in definition [ lf ] .
take a base point @xmath29 and an orientation preserving diffeomorphism @xmath37 .
since @xmath10 restricted over @xmath11 is a smooth fiber bundle with fiber @xmath3 , we can define a homomorphism @xmath38 called the _
monodromy representation _ of @xmath10 with respect to @xmath39 .
let @xmath40 be the loop based at @xmath41 consisting of the boundary circle of a small disk neighborhood of @xmath12 oriented counterclockwise and a simple path connecting a point on the circle to @xmath41 in @xmath11 .
it is known that @xmath42)$ ] is a dehn twist along some essential simple closed curve @xmath43 on @xmath3 , which is called the _ vanishing cycle _ of the critical point @xmath13 on @xmath22 . if @xmath13 is of positive type ( resp .
of negative type ) , then the dehn twist is right - handed ( resp .
left - handed ) .
a singular fiber is said to be of _ type i _ if the vanishing cycle is non - separating and of _ type @xmath44 _ for @xmath45 $ ] if the vanishing cycle is separating and it bounds a genus@xmath28 subsurface of @xmath3 .
a singular fiber is said to be of _ type @xmath46 _ ( resp .
_ type @xmath47 _ and _ type @xmath48 _ , _ type @xmath49 _ ) if it is of type i and of positive type ( resp . of type i and of negative type , of type @xmath44 and of positive type , of type @xmath44 and of negative type ) .
we denote by @xmath50 , @xmath51 , @xmath52 , and @xmath53 , the numbers of singular fibers of @xmath10 of type @xmath46 , @xmath47 , @xmath48 , and @xmath49 , respectively .
a lefschetz fibration is called _ irreducible _ if every singular fiber is of type i. a lefschetz fibration is called _ chiral _ if every singular fiber is of positive type .
suppose that the cardinality of @xmath54 is equal to @xmath55 . a system @xmath56 of arcs on @xmath6
is called a _
hurwitz arc system _ for @xmath54 with base point @xmath41 if each @xmath57 is an embedded arc connecting @xmath41 with a point of @xmath54 in @xmath6 such that @xmath58 for @xmath59 , and they appear in this order around @xmath41 ( see kamada @xcite ) .
when @xmath6 is a @xmath0-sphere , the system @xmath60 determines a system of generators of @xmath61 , say @xmath62 .
we call @xmath63 a _ hurwitz system _ of @xmath10 . in this subsection
we introduce a chart description for lefschetz fibrations of genus greater than two by employing wajnryb s finite presentation @xcite of mapping class groups .
general theories of charts for presentations of groups were developed independently by kamada @xcite and hasegawa @xcite .
we use the terminology of chart description in kamada @xcite .
we first review a finite presentation of the mapping class group of a closed oriented surface due to wajnryb . for @xmath64 ,
let @xmath65 be a right - handed dehn twist along the simple closed curve @xmath66 on @xmath3 depicted in figure [ curves ] .
@xmath67 [ r ] at 0 155 @xmath68 [ t ] at 45 139 @xmath69 [ t ] at 70 146 @xmath70 [ t ] at 98 139 @xmath71 [ r ] at 88 182 @xmath72 [ t ] at 242 139 @xmath68 [ t ] at 44 26 @xmath73 [ t ] at 105 26 @xmath74 [ t ] at 182 26 @xmath72 [ t ] at 243 26 @xmath75 [ r ] at 135 72 , title="fig : " ] [ pres ] suppose that @xmath4 is greater than two .
the mapping class group @xmath30 is generated by elements @xmath76 and has defining relations : @xmath77 @xmath78 @xmath79 @xmath80 @xmath81 @xmath82 @xmath83 for @xmath84 .
we make use of the presentation above to introduce a notion of chart which gives a graphic description of monodromy representations of lefschetz fibrations .
we set @xmath85\ } , \end{aligned}\ ] ] for @xmath86 , where @xmath87 ) , \end{aligned}\ ] ] and @xmath88 are defined as in theorem [ pres ] .
let @xmath6 be a connected closed oriented surface and @xmath89 a finite graph in @xmath6 such that each edge of @xmath89 is oriented and labeled with an element of @xmath90 .
we denote the label @xmath65 by @xmath91 for short .
choose a simple path @xmath40 which intersects with edges of @xmath89 transversely and does not intersect with vertices of @xmath89 .
for such a path @xmath40 , we obtain a word @xmath92 in @xmath93 by reading off the labels of intersecting edges along @xmath40 with exponents as in figure [ intersection ] ( a ) .
we call the word @xmath92 the _ intersection word _ of @xmath40 with respect to @xmath89 .
conversely , we can specify the number , orientations , and labels of consecutive edges in @xmath89 by indicating a ( dashed ) arrow intersecting the edges transversely together with the intersection word of the arrow with respect to @xmath89 ( see figure [ intersection ] ( b ) and ( c ) ) .
@xmath40 [ r ] at 0 18 @xmath94 [ t ] at 30 0 @xmath0 [ t ] at 59 0 @xmath94 [ t ] at 87 0 @xmath95 [ t ] at 116 0 @xmath0 [ t ] at 143 0 [ t ] at 85 -15 @xmath96 [ l ] at 255 18 [ t ] at 235 -15 @xmath96 [ b ] at 313 33 [ t ] at 313 -15 , title="fig : " ] for a vertex @xmath97 of @xmath89 , a small simple closed curve surrounding @xmath97 in the counterclockwise direction is called a _ meridian loop _ of @xmath97 and denoted by @xmath98 .
the vertex @xmath97 is said to be _ marked _ if one of the regions around @xmath97 is specified by an asterisk .
if @xmath97 is marked , the intersection word @xmath99 of @xmath98 with respect to @xmath89 is well - defined .
if not , it is determined up to cyclic permutation .
see kamada @xcite for details .
[ def : chart ] a _ chart _ in @xmath6 is a finite graph @xmath89 in @xmath6 ( possibly being empty or having _ hoops _ that are closed edges without vertices ) whose edges are labeled with an element of @xmath90 , and oriented so that the following conditions are satisfied ( see figure [ verticesa ] , figure [ verticesb ] , and figure [ verticesc ] ) : * the vertices of @xmath89 are classified into two families : _ white vertices _ and _ black vertices _ ; * if @xmath97 is a white vertex ( resp .
a black vertex ) , the word @xmath99 is a cyclic permutation of an element of @xmath100 ( resp .
of @xmath101 ) .
a white vertex @xmath97 is said to be of _ type _
@xmath102 ( resp . of _ type _
@xmath103 ) if @xmath104 is a cyclic permutation of @xmath105 ( resp . of @xmath106 ) .
a black vertex @xmath97 is said to be of _ type _
@xmath107 if @xmath99 is a cyclic permutation of @xmath108 .
a chart @xmath89 is said to be _ marked _ if each white vertex ( resp .
black vertex ) @xmath97 is marked and @xmath99 is exactly an element of @xmath100 ( resp . of @xmath101 ) .
if a base point @xmath41 of @xmath6 is specified , we always assume that a chart @xmath89 is disjoint from @xmath41 . a chart consisting of two black vertices and one edge connecting them
is called a _ free edge_. @xmath91 [ tr ] at 1 21 @xmath109 [ br ] at 1 52 @xmath91 [ bl ] at 31 52 @xmath109 [ tl ] at 31 21 @xmath91 [ tr ] at 95 18 @xmath110 [ r ] at 88 35 @xmath91 [ br ] at 96 54 @xmath110 [ bl ] at 133 54 @xmath91 [ l ] at 139 35 @xmath110 [ tl ] at 132 18 @xmath95 [ b ] at 222 73 @xmath0 [ b ] at 231 73 @xmath94 [ b ] at 239 73 @xmath95 [ b ] at 248 73 @xmath0 [ b ] at 256 73 @xmath94 [ b ] at 265 73 @xmath95 [ b ] at 273 73 @xmath0 [ b ] at 282 73 @xmath94 [ b ] at 291 73 @xmath95 [ b ] at 300 73 @xmath0 [ b ] at 309 73 @xmath94 [ b ] at 317 73 @xmath7 [ t ] at 346 0 @xmath95 [ t ] at 337 0 @xmath0 [ t ] at 330 0 @xmath94 [ t ] at 321 0 @xmath94 [ t ] at 312 0 @xmath0 [ t ] at 303 0 @xmath95 [ t ] at 295 0 @xmath7 [ t ] at 287 0 @xmath111 [ t ] at 278 0 @xmath7 [ t ] at 270 0 @xmath95 [ t ] at 261 0 @xmath0 [ t ] at 252 0 @xmath94 [ t ] at 243 0 @xmath94 [ t ] at 235 0 @xmath0 [ t ] at 227 0 @xmath95 [ t ] at 219 0 @xmath7 [ t ] at 210 0 @xmath111 [ t ] at 201 0 , @xmath112 , @xmath113,title="fig : " ] @xmath114 [ b ] at 29 90 @xmath94 [ b ] at 47 84 @xmath95 [ b ] at 56 84 @xmath115 [ b ] at 64 84 @xmath116 [ b ] at 81 90 @xmath117 [ b ] at 107 90 @xmath111 [ t ] at 128 9 @xmath118 [ t ] at 111 4 @xmath119 [ t ] at 85 4 @xmath117 [ t ] at 62 4 @xmath111 [ t ] at 42 9 @xmath118 [ t ] at 26 4 @xmath111 [ t ] at 8 9 @xmath120 [ b ] at 218 94 @xmath94 [ b ] at 244 94 @xmath94 [ b ] at 253 94 @xmath120 [ b ] at 278 94 @xmath121 [ l ] at 321 47 @xmath120 [ t ] at 278 0 @xmath94 [ t ] at 253 0 @xmath94 [ t ] at 244 0 @xmath120 [ t ] at 218 0 @xmath122 [ r ] at 178 47 and @xmath123,title="fig : " ] @xmath91 [ l ] at 46 34 @xmath91 [ l ] at 46 8 @xmath94 [ b ] at 223 45 @xmath124 [ b ] at 202 45 @xmath94 [ b ] at 193 45 @xmath124 [ b ] at 174 45 @xmath94 [ b ] at 139 45 @xmath124 [ b ] at 118 45 and @xmath125,title="fig : " ] it would be worth noting that the intersection word of a ` clockwise ' meridian of a white vertex of type @xmath102 is equal to @xmath102 , while that of a ` counterclockwise ' meridian of a black vertex of type @xmath107 is equal to @xmath107 in this paper .
this notation is different from those of kamada @xcite and hasegawa @xcite , who always consider ` counterclockwise ' meridians for both white and black vertices .
we next introduce several moves for charts .
let @xmath89 and @xmath126 be two charts on @xmath6 and @xmath41 a base point of @xmath6 .
let @xmath127 be a disk embedded in @xmath128 .
suppose that the boundary @xmath129 of @xmath127 intersects @xmath89 and @xmath126 transversely .
we say that @xmath126 is obtained from @xmath89 by a _ chart move of type w _
if @xmath130 and that both @xmath131 and @xmath132 have no black vertices .
we call chart moves of type w shown in figure [ movesa ] ( a ) , ( b ) , and ( c ) , a _ channel change _ , a _ birth / death of a hoop _ , and a _ birth / death of a pair of white vertices _ , respectively .
@xmath91 [ tr ] at 0 66 @xmath91 [ bl ] at 41 106 @xmath91 [ tr ] at 80 66 @xmath91 [ bl ] at 121 106 ( a ) [ t ] at 61 84 @xmath91 [ tl ] at 184 76 empty at 258 86 ( b ) [ t ] at 209 84 @xmath102 at 57 20 @xmath133 at 87 23 ( c ) [ t ] at 132 19 let @xmath107 and @xmath134 be elements of @xmath101 .
suppose that there exists a word @xmath96 in @xmath93 such that two words @xmath134 and @xmath135 determine the same element of @xmath30 .
if a chart @xmath89 contains a black vertex of type @xmath107 , then we can change a part of @xmath89 near the vertex by using a local replacement depicted in figure [ movesb ] to obtain another chart @xmath126 .
we say that @xmath126 is obtained from @xmath89 by a _
chart move of transition_. note that the blank labeled with @xmath136 can be filled only with edges and white vertices .
@xmath107 [ b ] at 35 81 @xmath134 [ b ] at 205 81 @xmath96 [ bl ] at 245 86 @xmath96 [ tl ] at 245 40 @xmath107 [ bl ] at 257 81 t at 232 64 we say that @xmath126 is obtained from @xmath89 by a
_ chart move of conjugacy type _ if @xmath126 is obtained from @xmath89 by a local replacement depicted in figure [ movesc ] .
@xmath41 [ l ] at 7 22 @xmath41 [ l ] at 117 22 @xmath91 [ l ] at 135 22 @xmath41 [ l ] at 197 22 @xmath41 [ l ] at 304 22 @xmath91 [ l ] at 323 22 let @xmath89 be a chart in @xmath6 with base point @xmath41 and @xmath137 the set of black vertices of @xmath89 .
for a loop @xmath40 in @xmath138 based at @xmath41 , the element of @xmath30 determined by the intersection word @xmath92 of @xmath40 with respect to @xmath89 does not depend on a choice of representative of the homotopy class of @xmath40 .
thus we obtain a homomorphism @xmath139 , which is called the _
homomorphism determined by @xmath89_. we now state a classification of lefschetz fibrations in terms of charts and chart moves .
let @xmath6 be a connected closed oriented surface . [
correspondence ] suppose that @xmath4 is greater than two .
( 1 ) let @xmath10 be a lefschetz fibration of genus @xmath4 over @xmath6 and @xmath140 a monodromy representation of @xmath10 .
then there exists a chart @xmath89 in @xmath6 such that the homomorphism @xmath141 determined by @xmath89 is equal to @xmath140 .
( 2 ) for every chart @xmath89 in @xmath6 , there exists a lefschetz fibration @xmath10 of genus @xmath4 over @xmath6 such that a monodromy representation of @xmath10 is equal to the homomorphism @xmath141 determined by @xmath89 .
we call such @xmath89 as in proposition [ correspondence ] ( 1 ) a chart _ corresponding to @xmath10 _ , and such @xmath10 as in proposition [ correspondence ] ( 2 ) a lefschetz fibration _ described by @xmath89_. instead of giving a proof of proposition [ correspondence ] , we show an example of a chart and describe the correspondence of the chart to a hurwitz system of a lefschetz fibration .
let @xmath6 be a @xmath0sphere .
we consider a chart @xmath89 in @xmath6 with base point @xmath41 and a system @xmath142 of loops based at @xmath41 , which is determined by a hurwitz arc system @xmath143 for the set @xmath137 of black vertices of @xmath89 , as in figure [ examplea ] .
the intersection words of the loops with respect to @xmath89 are @xmath144 each of which represents the image @xmath145 of the homotopy class @xmath146 of @xmath147 under the homomorphism @xmath148 .
since the group @xmath149 has a presentation @xmath150 , @xmath141 is determined by the system @xmath151 , which is a hurwitz system of a certain lefschetz fibration of genus @xmath4 over @xmath6 because each @xmath145 is a dehn twist .
note that the product @xmath152 of the intersection words represents the identity of @xmath30 .
@xmath41 [ t ] at 175 184 @xmath94 [ b ] at 30 156 @xmath0 [ b ] at 46 95 @xmath94 [ b ] at 112 77 @xmath0 [ b ] at 186 77 @xmath94 [ r ] at 154 103 @xmath95 [ t ] at 137 129 @xmath95 [ b ] at 191 115 @xmath95 [ r ] at 208 25 @xmath0 [ r ] at 243 50 @xmath0 [ t ] at 283 114 [ r ] at 70 135 [ t ] at 152 166 [ t ] at 233 142 [ b ] at 278 162 , title="fig : " ] [ classification ] suppose that @xmath4 is greater than two .
let @xmath10 and @xmath24 be lefschetz fibrations of genus @xmath4 over @xmath6 , and @xmath89 and @xmath126 charts corresponding to @xmath10 and @xmath24 , respectively .
then @xmath10 is strictly isomorphic to @xmath24 if and only if @xmath89 is transformed to @xmath126 by a finite sequence of chart moves of type w , chart moves of transitions , chart moves of conjugacy type , and ambient isotopies of @xmath6 relative to @xmath41 .
proposition [ correspondence ] and theorem [ classification ] follow from a classification theorem of lefschetz fibrations due to kas @xcite and matsumoto @xcite together with fundamental theorems on charts and chart moves by kamada ( * ? ? ?
* sections 48 ) .
we end this subsection with a definition and chart description of fiber sums of lefschetz fibrations .
let @xmath8 and @xmath153 be lefschetz fibrations of genus @xmath4 . take regular values @xmath29 and @xmath154 of @xmath10 and @xmath24 , and small disks @xmath155 and @xmath156 near @xmath41 and @xmath157 , respectively .
consider general fibers @xmath158 and @xmath159 and orientation preserving diffeomorphisms @xmath160 and @xmath161 , respectively .
let @xmath162 be an orientation preserving diffeomorphism and @xmath163 an orientation reversing diffeomorphism .
the new manifold @xmath164 obtained by glueing @xmath165 and @xmath166 by @xmath167 admits a lefschetz fibration @xmath168 of genus @xmath4 .
we call @xmath169 the _ fiber sum _ of @xmath10 and @xmath24 with respect to @xmath170 . although the diffeomorphim type of @xmath164 and the isomorphism type of @xmath169 depend on a choice of the diffeomorphism @xmath170 in general , we often abbreviate @xmath169 as @xmath171 .
let @xmath89 and @xmath126 be charts corresponding to @xmath10 and @xmath24 , and @xmath172 and @xmath173 small disks near @xmath41 and @xmath174 disjoint from @xmath89 and @xmath126 , respectively . connecting @xmath175 with @xmath176 by a tube , we have a connected sum @xmath177 of @xmath6 and @xmath178 .
let @xmath96 be a word in @xmath93 which represents the mapping class of @xmath170 in @xmath30 .
let @xmath179 be the union of @xmath89 , @xmath126 , and hoops on the tube representing @xmath96 ( see figure [ fibersum ] ) .
then the fiber sum @xmath169 is described by this new chart @xmath179 in @xmath177 with base point @xmath41 .
if the word @xmath96 is trivial , then the chart @xmath179 is denoted also by @xmath180 , which is called a _
product _ of @xmath89 and @xmath126 .
@xmath41 [ r ] at 128 38 @xmath157 [ l ] at 210 38 @xmath6 [ br ] at 3 57 @xmath178 [ bl ] at 330 59 @xmath96 [ t ] at 170 13 @xmath89 at 71 40 @xmath126 at 270 40 in @xmath177,title="fig : " ]
in this section we review the signature cocycle discovered by meyer and prove a signature theorem for lefschetz fibrations . in this subsection
we give a brief survey on meyer s signature cocycle .
we begin with the definition of the signature cocycle .
let @xmath4 be a positive integer .
[ cocycle ] for @xmath181 , we consider the vector space @xmath182 and the bilinear form @xmath183 defined by @xmath184 where @xmath185 is the standard inner product of @xmath186 and @xmath187 . since @xmath188 is symmetric , we can define an integer @xmath189 to be the signature of @xmath190 .
the map @xmath191 is called the _
signature cocycle_. let @xmath192 be a compact connected oriented surface of genus @xmath111 with three boundary components and @xmath193 a fiber bundle over @xmath192 with fiber @xmath3 and structure group @xmath194 .
the fundamental group @xmath195 of @xmath192 with base point @xmath196 is a free group generated by two loops @xmath197 and @xmath17 depicted in figure [ pants ] . if we take an orientation preserving diffeomorphism @xmath198 , we obtain the monodromy representation @xmath199 which sends @xmath197 to @xmath200 and @xmath17 to @xmath201 . since @xmath35 acts on @xmath202 preserving the intersection form , we have a representation @xmath203 by fixing a symplectic basis on @xmath204 .
let @xmath205 and @xmath6 denote matrices corresponding to @xmath200 and @xmath201 , respectively .
@xmath206 [ br ] at 1 56 @xmath207 [ bl ] at 117 53 @xmath196 [ b ] at 58 59 @xmath197 [ tl ] at 41 16 @xmath17 [ tr ] at 75 16 , title="fig : " ] meyer closely studied the signature of the total space @xmath208 to obtain the following theorem . [ meyer ] the signature @xmath209 of @xmath208 is equal to @xmath210 .
theorem [ meyer ] and novikov s additivity implies that @xmath211 is a @xmath0cocycle of @xmath212 .
we recall a maslov index for a triple of lagrangian subspaces and wall s non - additivity theorem , which are used in the proof of theorem [ meyer ] .
let @xmath213 be a real vector space of dimension @xmath214 , @xmath215 a symplectic form on @xmath213 , and @xmath216 the lagrangian grassmannian of @xmath217 , which is the set of lagrangian subspaces of @xmath217 . for @xmath218 ,
the bilinear form @xmath219 is symmetric .
we define an integer @xmath220 to be the signature of @xmath221 , which is called the _ ternary maslov index _ of the triple @xmath222 .
let @xmath223 be compact oriented smooth @xmath7manifolds , @xmath224 compact oriented smooth @xmath95manifolds , and @xmath225 a closed oriented smooth @xmath0manifold .
we assume that @xmath226 , and the orientations of these manifolds satisfy @xmath227=[m_1]+[m_2 ] , \ ; \partial_*[m_1]=[x_2]-[x_1 ] , \ ; \partial_*[m_2]=[x_3]-[x_2 ] , \\
\partial_*[x_1 ] = \partial_*[x_2]=\partial_*[x_3]=[\sigma].\end{gathered}\ ] ] let @xmath228 be the intersection form on @xmath229 and @xmath230 the kernel of the homomorphism @xmath231 induced by the inclusion @xmath232 for @xmath233 .
since @xmath234 for @xmath233 , we can define the ternary maslov index @xmath220 of the triple @xmath222 .
[ wall ] @xmath235 .
gambaudo and ghys @xcite ( and independently the first author ) made use of theorem [ wall ] to give the following proof of theorem [ meyer ] .
see also gilmer and masbaum @xcite .
consider @xmath192 to be a boundary sum of two annuli @xmath206 and @xmath207 ( see figure [ pants ] ) .
we set @xmath236 , @xmath237 , @xmath238 , @xmath239 , @xmath240 , and @xmath241 .
applying theorem [ wall ] to these manifolds , we have @xmath242 because each of @xmath243 and @xmath244 is a product of a mapping torus with an interval , which has signature zero . since the bordered component of @xmath245 is diffeomorphic to @xmath246 for @xmath233 , we put @xmath247 , @xmath248 , and obtain @xmath249 where @xmath204 is the first homology @xmath250 of @xmath3 and @xmath251 is the intersection form of @xmath3 .
it is easily seen that the subspace @xmath252 is written as @xmath253 and the symmetric bilinear form @xmath170 on @xmath252 is written as @xmath254 we consider the vector space @xmath255 and the symmetric bilinear form @xmath256 on @xmath257 defined by @xmath258 since the linear map @xmath259 is compatible with the bilinear forms , the signature of @xmath260 is equal to that of @xmath261 , which is isomorphic to @xmath262 under a choice of a symplectic basis of @xmath204 .
therefore we conclude that @xmath263 .
it is known that @xmath211 is a normalized , symmetric @xmath0cocycle of @xmath212 and invariant under conjugation .
the cohomology class @xmath264\in h^2({\rm sp}(2g,\bbb{z});\bbb{z})$ ] corresponds to @xmath265 under homomorphisms : @xmath266 for more details see meyer @xcite , turaev @xcite , barge and ghys @xcite , and kuno @xcite . in this subsection
we describe the signature of a lefschetz fibration of genus greater than two in terms of charts .
let @xmath4 be an integer greater than two .
let @xmath6 be a connected closed oriented surface and @xmath89 a chart in @xmath6 .
we denote the number of white vertices of type @xmath267 ( resp .
@xmath268 ) minus the number of white vertices of type @xmath269 ( resp .
@xmath270 ) included in @xmath89 by @xmath271 ( resp .
@xmath272 ) .
similarly , we denote the number of black vertices of type @xmath273 ( resp .
@xmath274 ) included in @xmath89 by @xmath275 ( resp .
@xmath276 ) , and set @xmath277 ( resp .
@xmath278 ) and @xmath279 .
[ sigma ] the number @xmath280}(4h(h+1)-1)\ , n_h(\gamma)\ ] ] is called the _ signature _ of @xmath89 .
let @xmath8 be a lefschetz fibration of genus @xmath4 and @xmath89 a chart in @xmath6 corresponding to @xmath10 .
the purpose of this subsection is to show the following theorem .
[ signature ] the signature @xmath281 of @xmath5 is equal to @xmath282
. it immediately follows from theorem [ signature ] that @xmath282 is invariant under chart moves of type w and chart moves of transition .
although any combinatorial proof of this fact does not seem to be known , hasegawa @xcite proved that @xmath282 is invariant under chart moves of transitions by a purely combinatorial method on the assumption that it is invariant under chart moves of type w. let @xmath283 be the set of right - handed dehn twists along simple closed curves in @xmath3 and @xmath284 the set of words in @xmath285 representing an element of the kernel of the natural epimorphism from the free group generated by @xmath283 to @xmath30 .
for a word @xmath286 , we define an integer @xmath287 where @xmath211 is the signature cocycle ( definition [ cocycle ] ) , @xmath288 is the image of @xmath289 under the composition of the natural map @xmath290 and a natural epimorphism @xmath203 , and @xmath291 is the number of dehn twists along separating simple closed curves included in @xmath96 .
suppose that @xmath6 is a @xmath0sphere .
if we choose a monodromy representation @xmath140 and a hurwitz arc system @xmath143 for @xmath54 with base point @xmath41 , we have a hurwitz system @xmath292 of @xmath10 .
since @xmath293 are dehn twists and @xmath294 in @xmath30 , we think @xmath295 as a word @xmath296 in @xmath284 .
theorem [ meyer ] and novikov s additivity for signature imply the next theorem .
[ nagami ] the signature @xmath281 of @xmath5 is equal to @xmath297 .
we are now ready to prove theorem [ signature ] .
choose a base point @xmath298 and a disk @xmath127 in @xmath299 centered at @xmath41 .
we denote the set of edges of @xmath89 by @xmath300 .
for each @xmath301 , we choose a point @xmath302 in a region of @xmath299 adjacent to @xmath303 , and a simple path @xmath304 from @xmath302 to @xmath41 which intersects with edges of @xmath89 transversely and does not intersect with vertices of @xmath89 .
let @xmath305 be the intersection word of @xmath304 with respect to @xmath89 and @xmath306 the label of @xmath303 .
we choose a family @xmath307 of mutually disjoint disks included in @xmath127 and put the chart @xmath308 depicted in figure [ charte ] in @xmath309 for each @xmath303 .
@xmath310 [ r ] at 0 34 @xmath311 [ t ] at 62 31 , title="fig : " ] taking the union of @xmath89 with @xmath308 for all @xmath301 , we obtain a new chart @xmath312 in @xmath6 , which describes a fiber sum @xmath313 of @xmath10 with lefschetz fibrations over @xmath314 described by a free edge . for each @xmath301
, we apply channel changes as in figure [ channelc ] to let a free edge pass through the edges intersecting with @xmath304 .
we then apply a channel change as in figure [ channeld ] to ` cut ' @xmath303 into two edges .
thus we obtain a new chart @xmath315 in @xmath6 .
@xmath303 [ b ] at 1 199 @xmath302 [ b ] at 16 158 @xmath311 [ l ] at 7 128 @xmath304 [ b ] at 40 160 @xmath305 [ tl ] at 85 131 @xmath316 [ b ] at 111 157 @xmath41 [ b ] at 90 158 @xmath311 [ t ] at 172 152 @xmath308 [ l ] at 230 156 @xmath303 [ b ] at 1 84 @xmath311 [ l ] at 7 13 @xmath305 [ tl ] at 84 16 @xmath311 [ t ] at 172 38 @xmath316 [ b ] at 172 85 @xmath303 [ b ] at 3 86 @xmath311 [ l ] at 7 13 @xmath311 [ l ] at 70 41 @xmath311 [ l ] at 176 13 @xmath311 [ l ] at 176 72 since each component of @xmath315 is a tree , a lefschetz fibration @xmath317 corresponding to @xmath315 is a fiber sum of a lefschetz fibration @xmath318 with a trivial @xmath3bundle over @xmath6 .
drawing a copy of @xmath315 in @xmath314 , we have a chart @xmath319 corresponding to @xmath320 . the signature of a lefschetz fibration over @xmath314 described by a free edge is equal to zero because @xmath321 for any @xmath322 ( see meyer ( * ? ? ?
* section 2 ) ) .
hence we have @xmath323 by theorem [ classification ] and novikov s additivity . since we did not change the numbers of white vertices and black vertices of type @xmath324 to make @xmath319 from @xmath89 , we see @xmath325 .
hence we only have to show @xmath326 in order to conclude @xmath327 .
applying chart moves of transition to each component of @xmath319 as in figure [ transition ] , we remove white vertices of type @xmath328 to obtain a union of copies of @xmath329 , where @xmath330 are charts depicted in figure [ chartb ] and figure [ chartf ] , and @xmath331 ( resp .
@xmath332 , @xmath333 ) is the mirror image of @xmath334 ( resp .
@xmath335 , @xmath336 ) with edges orientation reversed .
for the proof of @xmath326 , it is enough to show that the signature of a lefschetz fibration described by each of these charts coincides with the signature of the chart .
@xmath109 [ br ] at 34 192 @xmath91 [ bl ] at 64 192 @xmath91 [ tr ] at 34 161 @xmath109 [ tl ] at 64 161 @xmath109 [ br ] at 103 192 @xmath91 [ bl ] at 134 192 @xmath109 [ tl ] at 134 161 @xmath91 [ br ] at 172 195 @xmath110 [ bl ] at 209 195 @xmath110 [ r ] at 165 177 @xmath91 [ l ] at 216 177 @xmath91 [ tr ] at 171 157 @xmath110 [ tl ] at 209 157 @xmath91 [ br ] at 273 195 @xmath110 [ bl ] at 309 195 @xmath110 [ t ] at 267 175 @xmath91 [ l ] at 317 177 @xmath110 [ tl ] at 310 157 @xmath120 [ b ] at 43 125 @xmath94 [ b ] at 69 125 @xmath94 [ b ] at 78 125 @xmath120 [ b ] at 103 125 @xmath337 [ l ] at 144 81 @xmath94 [ l ] at 142 62 @xmath338 [ l ] at 144 44 @xmath120 [ t ] at 103 0 @xmath94 [ t ] at 78 0 @xmath94 [ t ] at 69 0 @xmath120 [ t ] at 43 0 @xmath337 [ r ] at 2 44 @xmath94 [ r ] at 6 63 @xmath338 [ r ] at 2 80 @xmath94 [ b ] at 210 64 @xmath338 [ b ] at 238 119 @xmath120 [ b ] at 255 114 @xmath94 [ b ] at 281 114 @xmath94 [ b ] at 290 114 @xmath120 [ b ] at 316 114 @xmath337 [ b ] at 332 119 @xmath91 [ b ] at 25 28 @xmath94 [ b ] at 218 50 @xmath124 [ b ] at 199 50 @xmath94 [ b ] at 188 50 @xmath124 [ b ] at 169 50 @xmath94 [ b ] at 135 50 @xmath124 [ b ] at 113 50 and @xmath334,title="fig : " ] @xmath94 [ b ] at 124 83 @xmath0 [ b ] at 115 83 @xmath95 [ b ] at 107 83 @xmath94 [ b ] at 98 83 @xmath0 [ b ] at 91 83 @xmath95 [ b ] at 82 83 @xmath94 [ b ] at 73 83 @xmath0 [ b ] at 64 83 @xmath95 [ b ] at 56 83 @xmath94 [ b ] at 48 83 @xmath0 [ b ] at 39 83 @xmath95 [ b ] at 31 83 @xmath111 [ t ] at 9 5 @xmath7 [ t ] at 18 5 @xmath95 [ t ] at 27 5 @xmath0 [ t ] at 35 5 @xmath94 [ t ] at 43 5 @xmath94 [ t ] at 52 5 @xmath0 [ t ] at 61 5 @xmath95 [ t ] at 69 5 @xmath7 [ t ] at 78 5 @xmath111 [ t ] at 86 5 @xmath7 [ t ] at 94 5 @xmath95 [ t ] at 103 5 @xmath0 [ t ] at 112 5 @xmath94 [ t ] at 120 5 @xmath94 [ t ] at 129 5 @xmath0 [ t ] at 138 5 @xmath95 [ t ] at 146 5 @xmath7 [ t ] at 154 5 @xmath114 [ b ] at 216 88 @xmath94 [ b ] at 234 83 @xmath95 [ b ] at 243 83 @xmath115 [ b ] at 251 83 @xmath116 [ b ] at 267 88 @xmath117 [ b ] at 293 88 @xmath111 [ t ] at 315 4 @xmath118 [ t ] at 297 1 @xmath119 [ t ] at 273 1 @xmath117 [ t ] at 248 1 @xmath111 [ t ] at 230 4 @xmath118 [ t ] at 214 1 @xmath111 [ t ] at 196 4 and @xmath336,title="fig : " ] let @xmath339 be one of @xmath340 drawn in @xmath314 and @xmath341 a lefschetz fibration described by @xmath339 .
if @xmath339 is equal to @xmath342 , it is easily seen that @xmath343 .
if @xmath339 is equal to @xmath334 , the word @xmath344 corresponds to a hurwitz system of @xmath345 ( see figure [ chartb ] ) , where @xmath346 is a right - handed dehn twist along the curve @xmath347 depicted in figure [ curves ] .
thus we have @xmath348 from definition [ sigma ] , theorem [ nagami ] , and explicit computations for @xmath349 due to endo and nagami ( * ? ? ?
* lemma 3.5 , proposition 3.9 ) . if @xmath339 is equal to @xmath335 ( resp .
@xmath336 ) , the word @xmath113 ( resp .
@xmath350 ) corresponds to a hurwitz system of @xmath345 ( see figure [ chartf ] ) .
thus we have @xmath351 from definition [ sigma ] , theorem [ nagami ] , and formulae of endo and nagami ( * ? ? ?
* lemma 3.5 , remark 3.7 , propositions 3.9 and 3.10 ) .
suppose that @xmath339 is equal to one of @xmath352 .
the mirror image @xmath353 of @xmath339 with edges orientation reversed corresponds to the lefschetz fibration @xmath354 with total space orientation reversed .
hence we have @xmath355 because we have already shown that @xmath343 is valid for @xmath356 .
this completes the proof of theorem [ signature ] .
in this section we prove two theorems on stabilization of lefschetz fibrations under taking fiber sums with copies of a fixed lefschetz fibration .
following auroux @xcite , we first introduce a notion of universality for lefschetz fibrations .
suppose that @xmath4 is greater than two .
a lefschetz fibration of genus @xmath4 over @xmath314 is called _ universal _ if it is irreducible , chiral , and it contains @xmath357 singular fibers of type @xmath46 whose vanishing cycles @xmath358 satisfies the following conditions : ( i ) @xmath146 and @xmath359 intersect transversely at one point for every @xmath360 ; ( ii ) @xmath361 and @xmath362 intersect transversely at one point ; ( iii ) @xmath146 and @xmath363 does not intersect for other pairs @xmath364 . a lefschetz fibration over @xmath314 is universal if and only if it is described by a chart @xmath365 depicted in figure [ universal ] by virtue of proposition [ correspondence ] , where the blank labeled with @xmath366 is filled only with edges , white vertices , and black vertices of type @xmath367 .
@xmath111 [ b ] at 16 57 @xmath94 [ b ] at 25 57 @xmath120 [ b ] at 53 57 @xmath368 at 36 13 [ l ] at 75 45 , title="fig : " ] a universal lefschetz fibration does exist for every @xmath4 greater than two .
for example , lefschetz fibrations @xmath369 constructed by auroux @xcite are universal except @xmath370 for @xmath371
. there would be many universal lefschetz fibrations of genus @xmath4 for a fixed @xmath4 .
we now state the first of our main theorems .
let @xmath6 be a connected closed oriented surface and @xmath372 a universal lefschetz fibration of genus @xmath4 .
[ main1 ] let @xmath8 and @xmath23 be lefschetz fibrations of genus @xmath4 .
there exists a non - negative integer @xmath373 such that @xmath374 is isomorphic to @xmath375 if and only if the following conditions hold : ( i ) @xmath376 ; ( ii ) @xmath377 for every @xmath378 $ ] ; ( iii ) @xmath379 .
auroux @xcite proved the ` if ' part of theorem [ main1 ] for chiral lefschetz fibrations over @xmath314 under the assumption that @xmath10 and @xmath24 have sections with the same self - intersection number .
hasegawa @xcite gave another proof of auroux s theorem by using chart description .
moreover he removed the assumption about existence and self - intersection number of sections in auroux s theorem .
the isomorphism class of a fiber sum @xmath380 of a lefschetz fibration @xmath10 with a universal lefschetz fibration @xmath381 does not depend on a choice of an orientation preserving diffeomorphism @xmath170 ( see proof of theorem [ main1 ] ) .
we first prove the ` if ' part .
assume that @xmath10 and @xmath24 satisfy the conditions ( i ) , ( ii ) , and ( iii ) .
let @xmath89 and @xmath126 be charts in @xmath6 corresponding to @xmath10 and @xmath24 , respectively .
we suppose that @xmath381 is described by a chart @xmath365 depicted in figure [ universal ] .
since every edge has two adjacent vertices , the sum of the signed numbers of adjacent edges for all vertices of @xmath89 is equal to zero : @xmath382}h(2h+1)\cdot n_h(\gamma)=0.\ ] ] a similar equality for @xmath126 also holds . interpreting the conditions ( i ) and ( ii ) as conditions on @xmath89 and @xmath126
, we have @xmath383 and @xmath384 for @xmath378 $ ] .
thus we obtain @xmath385 on the other hand , we have @xmath386 by the condition ( iii ) , theorem [ signature ] , and @xmath384 for @xmath378 $ ] .
hence @xmath387 and @xmath388 .
let @xmath373 be an integer larger than both of the number of edges of @xmath89 and that of @xmath126 . choose a base point @xmath389 .
the fiber sum @xmath374 is described by a chart @xmath390 for some words @xmath391 in @xmath93 . since
hoops surrounding @xmath365 can be removed by use of the edges of @xmath365 as in figure [ removea ] , the chart is transformed into a product @xmath392 by channel changes .
similarly , the fiber sum @xmath375 is described by a product @xmath393 .
@xmath91 [ b ] at 41 71 @xmath111 [ b ] at 23 51 @xmath91 [ b ] at 43 52 @xmath120 [ b ] at 61 51 @xmath368 at 43 11 @xmath91 [ b ] at 188 72 @xmath111 [ b ] at 142 51 @xmath120 [ b ] at 179 51 @xmath368 at 161 11 @xmath111 [ b ] at 257 51 @xmath91 [ b ] at 275 52 @xmath120 [ b ] at 293 51 @xmath368 at 276 11 we choose and fix @xmath357 edges of @xmath365 which are labeled with @xmath394 and adjacent to black vertices .
we apply chart moves only to these edges in the following .
since @xmath365 can pass through any edge of @xmath89 as shown in figure [ pass ] , we can move @xmath365 to any region of @xmath299 by channel changes .
@xmath91 [ b ] at 35 69 @xmath111 [ b ] at 18 51 @xmath91 [ b ] at 36 52 @xmath120 [ b ] at 55 51 @xmath368 at 35 11 [ t ] at 37 0 @xmath91 [ b ] at 106 72 @xmath91 [ b ] at 160 72 @xmath111 [ b ] at 115 52 @xmath120 [ b ] at 151 52 @xmath368 at 134 11 [ t ] at 134 0 @xmath91 [ t ] at 200 66 @xmath91 [ t ] at 255 66 @xmath120 [ t ] at 210 86 @xmath111 [ t ] at 246 86 @xmath368 at 228 126 [ t ] at 228 0 @xmath91 [ t ] at 322 66 @xmath120 [ t ] at 303 86 @xmath91 [ t ] at 321 85 @xmath111 [ t ] at 341 86 @xmath368 at 322 126 [ t ] at 321 0 for each edge of @xmath89 , we move a copy of @xmath365 to a region adjacent to the edge and apply a channel change to the edge and @xmath365 as in figure [ pass ] ( a ) and ( b ) .
applying chart moves of transition to each component of the chart as in figure [ transition ] , we remove white vertices of type @xmath328 to obtain a union of copies of @xmath395 shown in figures [ charth ] , [ chartc ] , [ chartd ] , where we use a simplification of diagrams as in figure [ symbol ] .
@xmath94 [ b ] at 113 50 @xmath124 [ b ] at 92 50 @xmath94 [ b ] at 82 50 @xmath124 [ b ] at 64 50 @xmath94 [ b ] at 30 50 @xmath124 [ b ] at 8 50 , title="fig : " ] @xmath94 [ b ] at 124 78 @xmath0 [ b ] at 115 78 @xmath95 [ b ] at 107 78 @xmath94 [ b ] at 98 78 @xmath0 [ b ] at 91 78 @xmath95 [ b ] at 82 78 @xmath94 [ b ] at 73 78 @xmath0 [ b ] at 64 78 @xmath95 [ b ] at 56 78 @xmath94 [ b ] at 48 78 @xmath0 [ b ] at 39 78 @xmath95 [ b ] at 31 78 @xmath111 [ t ] at 9 0 @xmath7 [ t ] at 18 0 @xmath95 [ t ] at 27 0 @xmath0 [ t ] at 35 0 @xmath94 [ t ] at 43 0 @xmath94 [ t ] at 52 0 @xmath0 [ t ] at 61 0 @xmath95 [ t ] at 69 0 @xmath7 [ t ] at 78 0 @xmath111 [ t ] at 86 0 @xmath7 [ t ] at 94 0 @xmath95 [ t ] at 103 0 @xmath0 [ t ] at 112 0 @xmath94 [ t ] at 120 0 @xmath94 [ t ] at 129 0 @xmath0 [ t ] at 138 0 @xmath95 [ t ] at 146 0 @xmath7 [ t ] at 154 0 @xmath94 [ b ] at 226 79 @xmath0 [ b ] at 235 79 @xmath95 [ b ] at 244 79 @xmath94 [ b ] at 252 79 @xmath0 [ b ] at 260 79 @xmath95 [ b ] at 269 79 @xmath94 [ b ] at 278 79 @xmath0 [ b ] at 286 79 @xmath95 [ b ] at 295 79 @xmath94 [ b ] at 302 79 @xmath0 [ b ] at 311 79 @xmath95 [ b ] at 320 79 @xmath111 [ t ] at 341 0 @xmath7 [ t ] at 333 0 @xmath95 [ t ] at 324 0 @xmath0 [ t ] at 316 0 @xmath94 [ t ] at 308 0 @xmath94 [ t ] at 299 0 @xmath0 [ t ] at 291 0 @xmath95 [ t ] at 281 0 @xmath7 [ t ] at 274 0 @xmath111 [ t ] at 265 0 @xmath7 [ t ] at 256 0 @xmath95 [ t ] at 247 0 @xmath0 [ t ] at 239 0 @xmath94 [ t ] at 231 0 @xmath94 [ t ] at 222 0 @xmath0 [ t ] at 214 0 @xmath95 [ t ] at 206 0 @xmath7 [ t ] at 197 0 and @xmath396,title="fig : " ] @xmath114 [ b ] at 29 88 @xmath94 [ b ] at 46 83 @xmath95 [ b ] at 55 83 @xmath115 [ b ] at 64 83 @xmath116 [ b ] at 80 88 @xmath117 [ b ] at 105 88 @xmath111 [ t ] at 128 3 @xmath118 [ t ] at 110 0 @xmath119 [ t ] at 85 0 @xmath117 [ t ] at 60 0 @xmath111 [ t ] at 43 3 @xmath118 [ t ] at 26 0 @xmath111 [ t ] at 9 3 @xmath111 [ t ] at 296 3 @xmath117 [ t ] at 279 0 @xmath111 [ t ] at 262 3 @xmath118 [ t ] at 245 0 @xmath116 [ t ] at 220 0 @xmath117 [ t ] at 195 0 @xmath111 [ t ] at 177 3 @xmath118 [ b ] at 197 88 @xmath119 [ b ] at 223 88 @xmath115 [ b ] at 240 85 @xmath95 [ b ] at 249 85 @xmath94 [ b ] at 258 85 @xmath397 [ b ] at 273 88 and @xmath398,title="fig : " ] @xmath91 [ b ] at 3 45 @xmath111 [ b ] at 89 57 @xmath91 [ b ] at 107 71 @xmath120 [ b ] at 126 57 @xmath368 at 108 14 @xmath399 at 40 28 if there is a pair of @xmath400 and @xmath396 , we remove them by a death of a pair of white vertices to obtain many copies of @xmath365 .
similarly , we remove a pair of @xmath401 and @xmath398 . since there is at least one @xmath365 , any copy of @xmath342 can be transformed into @xmath402 as in figure [ label ] .
@xmath109 [ l ] at 28 152 @xmath111 [ b ] at 83 186 @xmath91 [ b ] at 84 201 @xmath120 [ b ] at 120 186 @xmath368 at 101 143 @xmath109 [ l ] at 199 152 @xmath91 [ l ] at 212 164 @xmath111 [ b ] at 255 186 @xmath109 [ b ] at 255 201 @xmath120 [ b ] at 292 186 @xmath368 at 274 143 @xmath109 [ l ] at 370 152 @xmath91 [ l ] at 382 150 @xmath109 [ l ] at 390 142 @xmath111 [ b ] at 426 186 @xmath109 [ b ] at 444 186 @xmath120 [ b ] at 463 186 @xmath368 at 444 143 @xmath91 [ b ] at 26 91 @xmath109 [ t ] at 26 46 @xmath91 [ b ] at 26 23 @xmath109 [ t ] at 26 13 @xmath111 [ b ] at 82 73 @xmath109 [ b ] at 100 73 @xmath120 [ b ] at 119 73 @xmath368 at 100 29 @xmath91 [ l ] at 202 82 @xmath109 [ tl ] at 216 8 @xmath91 [ b ] at 201 10 @xmath111 [ b ] at 254 73 @xmath109 [ b ] at 273 73 @xmath120 [ b ] at 291 73 @xmath368 at 273 29 @xmath91 [ l ] at 372 44 @xmath111 [ b ] at 425 73 @xmath109 [ b ] at 444 73 @xmath120 [ b ] at 463 73 @xmath368 at 443 29 or @xmath403),title="fig : " ] thus we have a union @xmath312 of @xmath404 copies of @xmath402 , @xmath405 copies of @xmath406 , @xmath407 copies of @xmath331 , @xmath408 copies of @xmath400 ( or @xmath396 ) , @xmath409 copies of @xmath401 ( or @xmath398 ) , and @xmath410 copies of @xmath365 for some @xmath410 .
a similar argument implies that @xmath393 is transformed into a union @xmath411 of @xmath412 copies of @xmath402 , @xmath413 copies of @xmath406 , @xmath414 copies of @xmath331 , @xmath415 copies of @xmath400 ( or @xmath396 ) , @xmath416 copies of @xmath401 ( or @xmath398 ) , and @xmath417 copies of @xmath365 for some @xmath417 by chart moves of type w and chart moves of transition . by virtue of the conditions ( i ) and ( ii ) together with @xmath387 , @xmath388 , @xmath418 , and @xmath419 , we conclude that @xmath420 because of @xmath421 .
hence @xmath312 is transformed into @xmath411 by an ambient isotopy of @xmath6 relative to @xmath41 , which means that @xmath392 is transformed into @xmath393 by chart moves of type w , chart moves of transition , and ambient isotopies of @xmath6 relative to @xmath41 .
therefore @xmath374 is ( strictly ) isomorphic to @xmath375 by theorem [ classification ] .
we next prove the ` only if ' part .
take a non - negative integer @xmath373 so that @xmath374 is isomorphic to @xmath375 .
since an isomorphism preserves numbers and types of vanishing cycles and signatures , we have @xmath422 , @xmath423 for every @xmath378 $ ] , and @xmath424 . the conditions ( i ) , ( ii ) , ( iii ) follows from additivity of @xmath425 under fiber sum .
a lefschetz fibration of genus @xmath4 over @xmath314 is called _ elementary _ if it contains exactly two singular fibers of type @xmath46 and of type @xmath47 which have the same vanishing cycles .
a chart @xmath342 in @xmath314 corresponds to an elementary lefschetz fibration .
two elementary lefschetz fibrations of genus @xmath4 are isomorphic to each other .
the total space of an elementary lefschetz fibration of genus @xmath4 is diffeomorphic to @xmath426 .
we state the second of our main theorems .
let @xmath6 be a connected closed oriented surface and @xmath427 an elementary lefschetz fibration of genus @xmath4 .
[ main2 ] let @xmath8 and @xmath23 be lefschetz fibrations of genus @xmath4 .
there exists a non - negative integer @xmath373 such that a fiber sum @xmath428 is isomorphic to a fiber sum @xmath429 if and only if the following conditions hold : ( i ) @xmath376 ; ( ii ) @xmath377 for every @xmath378 $ ] ; ( iii ) @xmath379 .
in contrast to theorem [ main1 ] , the isomorphism class of a fiber sum @xmath430 of a lefschetz fibration @xmath10 with an elementary lefschetz fibration @xmath431 depends on a choice of an orientation preserving diffeomorphism @xmath170 in general .
we only show the ` if ' part .
the ` only if ' part is the same as that of the proof of theorem [ main1 ] .
assume that @xmath10 and @xmath24 satisfy the conditions ( i ) , ( ii ) , and ( iii ) .
let @xmath89 and @xmath126 be charts in @xmath6 corresponding to @xmath10 and @xmath24 , respectively .
it follows from the same argument as in the proof of theorem [ main1 ] that @xmath387 and @xmath388 .
let @xmath373 be an integer larger than both of the number of edges of @xmath89 and that of @xmath126 .
we construct the chart @xmath308 in @xmath6 for each @xmath301 as in the proof of theorem [ signature ] . taking the union of @xmath89 with @xmath308 for all @xmath301 and with @xmath432 copies of @xmath433
, we obtain a new chart @xmath312 in @xmath6 , which describes a fiber sum @xmath428 . applying channel changes as in the proof of theorem [ signature ] and deaths of pairs of white vertices appropriately ,
we obtain a union @xmath315 of @xmath405 copies of @xmath334 , @xmath407 copies of @xmath331 , @xmath408 copies of @xmath335 ( or @xmath332 ) , @xmath409 copies of @xmath336 ( or @xmath333 ) , and @xmath434 copies of @xmath342 for some @xmath434 , similarly , @xmath126 is transformed into @xmath411 , which describes a fiber sum @xmath429 , and then a union @xmath435 of @xmath413 copies of @xmath334 , @xmath414 copies of @xmath331 , @xmath415 copies of @xmath335 ( or @xmath332 ) , @xmath416 copies of @xmath336 ( or @xmath333 ) , and @xmath436 copies of @xmath342 for some @xmath436 .
a similar argument on the number @xmath437 as in the proof of theorem [ main1 ] implies that @xmath438 . adding @xmath439 copies of @xmath342 to either @xmath315 or @xmath435 if necessary
, we may assume that @xmath440 for every @xmath441 .
hence @xmath315 is transformed into @xmath435 by an ambient isotopy of @xmath6 relative to @xmath41 , which means that @xmath428 is ( strictly ) isomorphic to @xmath429 by theorem [ classification ] .
let @xmath4 be an integer greater than two and @xmath442 connected closed oriented surfaces .
we consider a lefschetz fibration @xmath443 of genus @xmath4 for each @xmath444 , and a universal lefschetz fibration @xmath372 of genus @xmath4 .
[ sums ] for ( possibly different ) fiber sums @xmath10 and @xmath24 of @xmath445 , fiber sums @xmath446 and @xmath447 are isomorphic to each other .
let @xmath89 and @xmath126 be charts corresponding to @xmath10 and @xmath24 . since hoops surrounding a component of @xmath89 ( and @xmath126 ) can be removed by use of the edges of @xmath365 as in figure [ removea ] , @xmath448 and @xmath449 are transformed into the same chart . proposition [ sums ] implies that there are many examples of non - isomorphic lefschetz fibrations with the same base , the same fiber , and the same numbers of singular fibers of each type which become isomorphic after one stabilization .
for example , the lefschetz fibration on @xmath450 constructed by fintushel and stern ( * ? ? ?
* theorem 14 ) ( see also park and yun @xcite ) for a fibered knot @xmath451 becomes isomorphic to that on @xmath452 for another fibered knot @xmath453 of the same genus after one stabilization .
similar results hold for lefschetz fibrations on @xmath454 constructed by fintushel and stern @xcite ( see also park and yun @xcite ) as well as fiber sums of ( generalizations of ) matsumoto s fibration studied by ozbagci and stipsicz @xcite , korkmaz @xcite , and okamori @xcite .
in this section we discuss possible variations of chart description for lefschetz fibrations .
we first choose large @xmath456 , and @xmath101 .
let @xmath90 be the set of right - handed dehn twists along simple closed curves in @xmath3 and @xmath101 the set of dehn twists along non - trivial simple closed curves in @xmath3 .
by virtue of a theorem of luo @xcite , @xmath457 gives an infinite presentation of @xmath30 for the set @xmath458 of the following four kinds of words : ( 0 ) trivial relator @xmath459 , where @xmath197 is the dehn twist along a trivial simple closed curve on @xmath3 ; ( 1 ) primitive braid relator @xmath460 , where @xmath461 and the curve for @xmath43 is the image of the curve for @xmath197 by @xmath17 ; ( 2 ) @xmath0chain relator @xmath462 , where @xmath463 and the curves for @xmath67 and @xmath68 intersect transversely at one point and the curve for @xmath464 is the boundary curve of a regular neighborhood of the union of the curves for @xmath67 and @xmath68 ; ( 3 ) lantern relator @xmath465 , where @xmath466 and the curves for @xmath197 and @xmath17 intersect transversely at two points with algebraic intersection number zero , the curve for @xmath43 is obtained by resolving the intersections of these two curves , and the curves for @xmath467 are the boundary curves of a regular neighborhood of those for @xmath468 .
let @xmath6 be a connected closed oriented surface .
charts in @xmath6 for the triple @xmath455 defined above have white vertices of type @xmath469 ( see figure [ verticesd ] ) . for a chart @xmath89 in @xmath6
, we denote the number of white vertices of type @xmath470 minus the number of white vertices of type @xmath471 included in @xmath89 by @xmath472 , where @xmath473 .
@xmath197 [ b ] at 4 54 @xmath197 [ br ] at 40 50 @xmath17 [ bl ] at 70 50 @xmath17 [ tr ] at 40 24 @xmath43 [ tl ] at 70 24 @xmath68 [ b ] at 114 73 @xmath67 [ b ] at 123 73 @xmath68 [ b ] at 132 73 @xmath67 [ b ] at 141 73 @xmath68 [ b ] at 150 73 @xmath67 [ b ] at 159 73 @xmath68 [ b ] at 167 73 @xmath67 [ b ] at 176 73 @xmath68 [ b ] at 185 73 @xmath67 [ b ] at 194 73 @xmath68 [ b ] at 203 73 @xmath67 [ b ] at 212 73 @xmath464 [ t ] at 162 0 @xmath43 [ b ] at 263 73 @xmath17 [ b ] at 271 73 @xmath197 [ b ] at 281 73 @xmath474 [ t ] at 257 0 @xmath475 [ t ] at 267 0 @xmath476 [ t ] at 277 0 @xmath477 [ t ] at 287 0 , @xmath478 , @xmath113 , @xmath350,title="fig : " ] let @xmath6 be a connected closed oriented surface of genus @xmath0 and @xmath455 the triple defined above for @xmath371 .
let @xmath480 be right - handed dehn twists along simple closed curves of the same names on @xmath481 depicted in figure [ twists ] .
we present @xmath6 as an octagon with opposite sides identified and consider a chart @xmath89 and loops @xmath482 based at @xmath41 in @xmath6 as in figure [ exampleb ] .
we use a simplification of diagrams as in figure [ verticese ] ( a ) if the curves for @xmath483 intersect transversely at one point , and that as in figure [ verticese ] ( b ) if the curves for @xmath484 and @xmath485 are disjoint .
since the intersection words of @xmath482 with respect to @xmath89 are @xmath486 a lefschetz fibration @xmath487 of genus @xmath95 described by @xmath89 is isomorphic to the lefschetz fibration constructed by korkmaz and ozbagci ( * ? ? ?
* theorem 1.2 ) .
@xmath10 has only one singular fiber and it is of type @xmath46 . we can compute the signature @xmath281 of the total space @xmath5 by proposition [ signature2 ] : @xmath488 which coincides with the value computed in ( * ? ? ?
* proposition 14 ) .
@xmath474 [ b ] at 16 53 @xmath477 [ b ] at 70 53 @xmath476 [ b ] at 124 53 @xmath475 [ b ] at 177 53 @xmath197 [ r ] at 88 25 @xmath17 [ t ] at 132 17 @xmath67 [ b ] at 247 60 @xmath68 [ b ] at 302 60 @xmath69 [ b ] at 356 60 @xmath43 [ t ] at 284 16 , title="fig : " ] @xmath41 [ t ] at 234 450 [ b ] at 145 427 [ b ] at 23 305 [ t ] at 24 144 [ t ] at 146 26 [ t ] at 323 33 [ t ] at 430 142 [ b ] at 432 306 [ b ] at 310 427 [ r ] at 116 361 @xmath475 [ b ] at 188 400 @xmath474 [ t ] at 158 358 @xmath43 [ r ] at 153 304 @xmath17 [ r ] at 193 289 @xmath197 [ t ] at 228 328 @xmath477 [ b ] at 259 404 @xmath476 [ b ] at 230 384 @xmath69 [ b ] at 97 324 @xmath43 [ b ] at 73 287 @xmath475 [ b ] at 58 253 @xmath69 [ b ] at 41 224 @xmath69 [ r ] at 104 219 @xmath475 [ b ] at 59 158 @xmath69 [ l ] at 161 212 @xmath43 [ r ] at 135 193 @xmath475 [ l ] at 113 143 @xmath69 [ r ] at 88 113 @xmath67 [ b ] at 240 300 @xmath17 [ b ] at 275 295 @xmath68 [ r ] at 278 245 @xmath476 [ l ] at 298 239 @xmath197 [ r ] at 319 198 @xmath68 [ l ] at 342 187 @xmath477 [ t ] at 412 260 @xmath67 [ l ] at 424 223 @xmath197 [ t ] at 420 184 @xmath68 [ b ] at 334 315 @xmath476 [ t ] at 353 310 @xmath197 [ r ] at 370 289 @xmath68 [ t ] at 383 270 @xmath68 [ t ] at 343 277 @xmath67 [ t ] at 292 171 @xmath67 [ r ] at 205 183 @xmath17 [ l ] at 224 164 @xmath477 [ r ] at 341 102 @xmath67 [ l ] at 364 82 @xmath477 [ t ] at 388 120 [ b ] at 203 343 @xmath484 [ b ] at 4 56 @xmath485 [ b ] at 26 56 @xmath484 [ b ] at 49 56 @xmath485 [ t ] at 4 0 @xmath484 [ t ] at 26 0 @xmath485 [ t ] at 49 0 @xmath489 [ b ] at 70 22 @xmath484 [ b ] at 91 56 @xmath485 [ b ] at 113 56 @xmath484 [ b ] at 135 56 @xmath485 [ t ] at 91 0 @xmath484 [ t ] at 113 0 @xmath485 [ t ] at 135 0 @xmath490 [ b ] at 117 29 [ t ] at 70 -15 @xmath484 [ br ] at 180 44 @xmath485 [ bl ] at 210 44 @xmath485 [ tr ] at 180 12 @xmath484 [ tl ] at 210 12 @xmath489 [ b ] at 228 22 @xmath484 [ br ] at 243 44 @xmath485 [ bl ] at 273 44 @xmath485 [ tr ] at 243 12 @xmath484 [ tl ] at 273 12 [ t ] at 228 -15 we next mention chart description for lefschetz fibrations with bordered base and fiber .
kamada @xcite gave a general theory for charts in a compact oriented surface with boundary .
various presentations of mapping class groups of surfaces with boundary have been investigated by researchers including gervais @xcite , labrure and paris @xcite , margalit and mccammond @xcite .
combining these two kinds of studies , one can immediately obtain a chart description for lefschetz fibrations with bordered base and fiber .
it would be worth considering compositions of monodromy representations with appropriate homomorphisms and charts corresponding to the compositions .
for example , hasegawa @xcite adopted a homomorphism from the @xmath491string braid group @xmath492 to the semi - direct product @xmath493 , while endo and kamada @xcite used a standard epimorphism from the hyperelliptic mapping class group of a closed oriented surface of genus @xmath4 to the mapping class group of a sphere with @xmath494 marked points .
theorem [ main1 ] and theorem [ main2 ] tell us that the numbers of singular fibers of all types and the signature of the total space completely determine the stable isomorphism class of a lefschetz fibration with given base and fiber .
thus any numerical invariant of lefschetz fibrations which is additive under fiber sum is determined by these invariants in principle .
nosaka @xcite has recently defined an invariant which is not additive under fiber sum .
non - numerical invariants such as monodromy group would be also useful ( see matsumoto @xcite and park and yun @xcite ) .
* acknowledgements * the authors would like to thank the referees for their helpful suggestions and corrections .
the first author was partially supported by jsps kakenhi grant numbers 21540079 , 25400082 .
the third author was partially supported by jsps kakenhi grant numbers 21340015 , 26287013 .
the fourth author was partially supported by jsps kakenhi grant numbers 21740042 , 26400082 .
y. matsumoto , _ lefschetz fibrations of genus two a topological approach _
, in `` topology and teichmller spaces '' s. kojima et al , eds .
the 37-th taniguchi sympo . , pp .
123148 , world scientific publishing , river edge , nj , 1996 . | we employ a certain labeled finite graph , called a chart , in a closed oriented surface for describing the monodromy of a(n achiral ) lefschetz fibration over the surface .
applying charts and their moves with respect to wajnryb s presentation of mapping class groups , we first generalize a signature formula for lefschetz fibrations over the @xmath0sphere obtained by endo and nagami to that for lefschetz fibrations over arbitrary closed oriented surface .
we then show two theorems on stabilization of lefschetz fibrations under fiber summing with copies of a typical lefschetz fibration as generalizations of a theorem of auroux . |
the discovery of the current accelerated expansion of the universe @xcite has generated a lot of excitement in the last few years .
it turns out that this unexpected behaviour can be modeled in many ways , most simply through a cosmological constant . on the other hand
it has been argued that , since this behaviour appears on large scales , the acceleration of the universe is due to some modification of gravity on such scales .
there are quite a few models which can describe such modifications of gravity , such as scalar - tensor theories @xcite-@xcite , the so called @xmath0 theories @xcite-@xcite ( which are a subset of scalar tensor theories ) , brane world models of which randall - sundrum models @xcite and dgp models @xcite are special cases , and some more complicated function of curvature invariants , such as @xmath1 @xcite where @xmath2 represents the gauss - bonnet combination defined as @xmath3 .
modifying gravity in a consistent way is not an easy task , hence many of these models face quite stringent theoretical bounds which can reduce the parameter space of these theories . in the same way one can try to use data to restrict further the parameter space and
possibly rule out some of these models . in this paper
we study @xmath4 and @xmath5 models , most generally , without specifying any explicit form for these functions . in the @xmath4 case , we consider both the metric based , and the palatini formalisms .
we use the recent expansion history of the universe , reconstructed allowing the hubble parameter to be a free function in redshift bins , using type ia supernovae ( sne ia ) data together with relevant contraints from the cosmic microwave background ( cmb ) and galaxy surveys , and invert this to obtain constraints on the parameters of the @xmath4 models .
this perspective has not been adopted before .
although it is similar in nature to the reconstruction of the quintessence potential @xcite , here the scalar degrees of freedom come entirely from the gravity sector .
much work has been dedicated to the subject of constraining @xmath4 theories using solar system measurements @xcite-@xcite .
we will not study this here for several reasons .
first , as believed by some ( e.g. @xcite ) , the cosmological parameters may be different from those measured locally , because as the background changes from a locally spherically symmetric metric to a homogeneous and isotropic one , the behaviour of quantities such as the coefficients of a taylor - expansion of @xmath4 might be quite different .
second , for this theory locally ( at the solar system scale ) the weak field approximation may not hold , so that it can not be matched to a perturbation of gr ; i.e. , the usual constraints can not be trivially applied because these are found by assuming the metric to be a perturbation about the gr - schwarzschild solution @xcite .
furthermore we are restricting our study to the background evolution at low redshifts , assuming gr - like behaviour until then .
we are not taking into account perturbations and their evolution , nor considering the stability of modes in the early universe , in constraining the models .
these relevant issues have been studied in @xcite .
this paper is organized as follows . in section
ii we discuss the general metric based @xmath4 case and its solutions . in section
iii we discuss the palatini formalism and its solutions . in section iv we consider the generalization including the gauss - bonnet term .
we end with a discussion and conclusions section .
appendices follow .
let us begin with an @xmath0 theory with the following action s = d^4 x + s_m , [ staction ] where @xmath6 is the ricci scalar and @xmath7 is the action for the matter fields .
the equations of motion are @xmath8 where an underscore @xmath6 implies a partial derivative with respect to @xmath6 ( @xmath9 ) . in a frw background
then , @xmath10 where @xmath11 is the hubble parameter and a prime denotes differentiation with respect to @xmath12 .
evaluated today this equation becomes @xmath13 + \gamma\ , \frac{r_0'}{h_0 ^ 2 } = \frac{8 \pi g}{3h_0 ^ 2}\ , \rho_0\ , \label{fried2}\ ] ] where an underscore 0 implies present values , and @xmath14 , @xmath15 , and @xmath16 are the dimension - less parameters of this theory .
since we have three parameters to solve for , we need three equations .
these are obtained by differentiating the friedmann equation twice .
the equations then involve higher derivatives of @xmath11 evaluated today ( in this case upto four ) . assuming that @xmath17 can be taylor - expanded in @xmath6 about today and retaining only up to the second order terms ( this being the simplest non - trivial case ) , @xmath18 + \frac{1}{2 } \hspace{0.25em } f_{rr 0 } \hspace{0.25em } h_0 ^ 4 \left [ \frac{r - r_0}{h_0 ^ 2 } \right]^{\ ! 2}\ , .
\label{frtaylor}\ ] ] in order to use the equations of motion effectively one needs to truncate the taylor expansion at some order , otherwise one would need an infinite number of equations , obtained from repeatedly differentiating the friedmann equation , to specify an infinite number of parameters , the taylor coefficients .
we chose to truncate this expansion of @xmath4 at second order because stability constraints on these theories involve @xmath19 and @xmath20 terms ( see e.g. @xcite ) .
differentiating the friedmann equation once gives 6hh(1+f_r)+3(hh-h^2)f_r+3h^2f_r=8g .
[ dfr1b ] this relation together with the conservation of stress - energy ( which ignoring contribution from radiation leads to ) =-3(1+w)-3 .
[ frenergy ] is equivalent to the second einstein equation
. equations ( [ frtaylor ] ) , ( [ dfr1b ] ) and ( [ frenergy ] ) lead to 6(1+f_r0)+3(-1)f_rr0r_0 + 3f_rr0r_0=-9_m , [ fre2 ] where we have defined the ( present day ) derivatives of the ( normalized ) hubble parameter as = = . for a flat friedmann - robertson - walker background @xmath21 hence equation ( [ fre2 ] ) can also be written as = - [ betag ] differentiating the friedmann equation a second time , evaluating it today and using equation ( [ betag ] ) gives & = & 12_m(3+^2+)^-1 , [ gammag ] where .
we see that = ( , , , , _ m ) , and the same is true for @xmath22 and @xmath23 .
the relations are non - linear .
posterior distributions of @xmath24 , @xmath25 , @xmath26 , @xmath27 and @xmath28 are obtained from a likelihood analysis of data in the following way .
we use 182 type ia supernovae ( sne ia ) from the hst / goods program , together with first year snls and some nearby sne ia , as compiled by @xcite .
the @xmath29 combination , where @xmath6 and @xmath30 are cmb shift parameters@xcite , is used to account for relevant constraints from the cmb @xcite .
the sdss baryon acoustic oscillation ( bao ) scale measurement is also used @xcite . following the analysis method of @xcite-@xcite,@xcite , model
independent constraints are derived on the hubble parameter in linear redshift bins using a markov chain monte carlo ( mcmc ) algorithm .
the @xmath31 s of the mcmc chain elements are then converted into the derivatives of @xmath32 , as described in appendix a. the top panel of fig 1 shows the constraints thus derived on what we for convenience shall sometimes refer to as the `` initial parameters '' . from the mcmc chains now we can obtain constraints on @xmath33 , @xmath22 , and @xmath23 using the equations derived above .
gravity : constraints on @xmath34 , @xmath15 , and @xmath35 ( bottom panel ) , obtained from constraints on @xmath36 and the derivatives of @xmath32 ( top panel ) , using current cosmological data.,title="fig : " ] [ fig:1 ] as a function of @xmath26 and @xmath27 setting @xmath24 and @xmath25 to their best values.,title="fig : " ] [ fig2 ] equation ( [ gammag ] ) is highly non - linear in the initial parameters .
small changes in these parameters can have a large effect on the values of the @xmath4 parameters , so that a relatively small number of high likelihood combinations of the initial parameters can lead to a large range of values for the @xmath4 parameters .
this leads to flat 1d parameter likelihood curves .
this effect is suppressed in the mcmc posterior of the @xmath4 parameters , because in addition to the likelihood this takes into account the number of samples that fall into each bin . in other words
the discrepancy between these curves is due to there being a main posterior peak ( representing a concentration of high likelihood points ) together with isolated points in parameter space that are allowed by the likelihood .
in addition , some allowed values of the initial parameters lead to singularities ; these are the zero values of the @xmath24 s and the zeroes of denominator in the expression for @xmath23 .
given the discrete samples in the mcmc chain , the parameters @xmath33 , @xmath22 and @xmath23 do nt actually blow up , but instead as a result have a larger range .
these problems should be at least somewhat eased by better data . for now , in order to avoid the discrepancy between the likelihood and the posterior , and to discount the region around singular points , we proceed to find solutions by first linearizing the equations for the modified gravity parameters about the mean values of the initial parameters , supported also by the fact that the initial parameters have close to gaussian distributions .
this approach will be used in subsequent sections as well .
we linearize the equations for @xmath33 , @xmath22 and @xmath23 about their maximum likelihood ( in mcmc , the relevant quantity is the mean ) values , which are obtained in turn from the mean values of the initial parameters : & & + ( p_i - p_i)._p_j = p_j , [ linal ] + & & + ( p_i - p_i)._p_j = p_j , + & & + ( p_i - p_i)._p_j = p_j , [ lingam ] where @xmath37 and hats represent mean values .
we obtain & = & -29.7708 + 31.9923 - 0.652908 - 0.0628292 + & & + 0.0513295 + 63.1444
_ m , + & = & -1.89148 + 2.29135 - 0.0479787 - 0.00179758
+ [ betaeqn ] & & + 0.00402273 + 4.20931
_ m , + & = & -0.00179067 - 0.00887807 + 0.000625996 + 0.0000797733 + & & - 0.0000138531 - 0.02083 _ m .
even though the linearization is about the mean values of the initial parameters , we then use the mcmc chain to obtain the distributions of the modified gravity parameters .
these are shown in the bottom panel of fig 1 .
as described above these represent the main body of solutions of the @xmath4 theory allowed by the data ; given the non - linearity of the equations there are other solutions isolated in parameter space .
we see that while the order of magnitude of @xmath33 and @xmath22 is constrained , @xmath38 is found to be slightly negative today over most of its allowed range . even though we are not discussing stability issues here
, @xcite find that @xmath20 at high redshift needs to be positive in order to avoid instability and obey gr . therefore under this theory
@xmath23 would have to change sign at some @xmath6 before today .
this is interesting .
fig 2 shows the solutions for @xmath22 at the mean values of the better constrained derivatives @xmath24 and @xmath25 , and over the allowed ranges of the higher derivatives .
a range of values for @xmath22 are possible , including the special case of @xmath39 discussed below .
if we assume that in the solar system the real metric can be expanded about gr schwarzchild , then for the previous action , the effective newton s constant can be written as @xcite @xmath40 which can be recast as a constraint on @xmath22 today ( imposing @xmath41 ) @xmath42 imposing @xmath39 , equation ( [ fried2 ] ) and ( [ betag ] ) give @xmath43 and @xmath44 = - 3 \hspace{0.25em } \omega_{m } \label{eq : param2 } .\ ] ] the data define constraints on the @xmath24 s as discussed in the previous section .
equations ( [ eq : param1 ] ) and ( [ eq : param2 ] ) can then be used to solve for @xmath33 and @xmath23 .
the complications arising due to the equations being non - linear in the initial parameters , as discussed earlier , apply here as well .
hence as before we proceed to linearize the equations of motion about the mean values of the @xmath24 s and @xmath36 .
using equations ( [ linal ] ) and ( [ lingam ] ) together with equations ( [ eq : param1 ] ) and ( [ eq : param2 ] ) , we get & & -4.53219 + 12.1526 - 1.54509 - 0.974158
+ 20.5967 _ m , + & & -0.028308 - 0.0521268 + 0.0262386 + 0.00930701 - 0.139456 _ m .
these relations , used on the mcmc chains , give the results shown in fig 3 .
case of the general metric - based @xmath4 theories.,title="fig : " ] [ fig3 ]
so far we have considered theories for which the gravity variables were chosen to be the metric elements @xmath45 . on the other hand , after writing down the action , one is free to choose a different set of fields .
for example , one can choose to find the equations of motion by varying the action with respect to the following two tensorial quantities , @xmath46 and @xmath47 , which are the metric and the christoffel symbols perturbations respectively .
in standard gr , the approach of choosing two different field variables , leads to the same standard einstein equations of motion .
however , in @xmath4 theories , with @xmath48 , the equations of motion are indeed different . of course
, one may wonder which description of gravity is the correct one .
however , since gravity is the least well known force , many physicists have argued that the possibility of introducing extra fields should be considered . in this formalism ( refered to as palatini , also studied in @xcite ) , @xmath49 becomes a function only of the christoffel symbols , so that it can not be written as usual in terms of derivatives of the scale factor . in this case
it is customary to write the general action in the following way s = d^4x+s_m .
the einstein equations obtained by varying the action with respect to the metric elements are f_r r_-12fg_=8gt _ , [ pala00 ] the trace of which is f_rr-2 f=-8g_m .
[ pala0 ] taylor - expanding @xmath17 in terms of @xmath6 about today , as before , we have and @xmath50 is its value today .
therefore from equation ( [ pala0 ] ) |r_0=and in general |r=|r_0- , this relation is important as we know how @xmath6 varies close to today .
taking the derivative of @xmath6 with repect to time , plugging it into the 00 component of equation ( [ pala00 ] ) , we get the new friedmann equation = , [ palafr ] with @xmath51 is the hubble factor @xmath52 , and = 4[e^3n(^2 - 2)+3_m]^2 following @xcite .
the friedmann equation ( [ palafr ] ) together with its two @xmath53-derivatives give us three equations for three unknowns .
in contrast to the metric - based case , the friedmann equation here does not involve any derivatives of the hubble parameter , because @xmath6 is independent of @xmath11 .
we end up with three non - linear equations involving @xmath36 , @xmath24 , @xmath25 , to be solved simultaneously for @xmath33 , @xmath22 and @xmath23 .
this time though , the equations are non - linear even in @xmath33 , @xmath22 , and @xmath23 , so that there is nt a unique solution for these parameters .
in general there is no telling how many real solutions there are .
we solve the equations numerically in the range [ -10,10 ] for the parameters .
listed in appendix b are multiple solutions found at the mean values of the initial parameters .
next we consider one of these solutions .
the equations of motion can be written as @xmath54 , with @xmath55 , where @xmath56=0 corresponds to the friedmann equation , and @xmath57=0 , and @xmath58=0 to its two derivatives .
let @xmath59 and @xmath60 .
the solutions listed in appendix b are the @xmath61 which solve @xmath62 , where @xmath63 correspond to the mean values of the initial parameters .
about any one solution we may linearize the equations of motion ( for reasons explained in section ii ) : f_i(p_j , q_k)=0f_i(p_j , q_k)+._p , q(p_j - p_j ) + ._p , q(q_j - q_j ) . with @xmath64 and @xmath65
, one has p_j=-(a^-1b)_jkq_k+p_j+(a^-1b)_jkq_k .
[ mateqn ] as an example , consider the solution [ 0.222102 , 0.00488155 , 4.12207e-05 ] .
we can obtain the distributions of @xmath66 about this solution using the linearized equations together on the mcmc chains for the initial parameters .
results are shown in fig 4 , and for this solution the matrix @xmath67 is also given in appendix b. though the figure shows reasonably strong constraints on the modified gravity parameters , it should be remembered that in this formalism the data allow multiple such solutions .
one may try to distinguish between the solutions by imposing consistency checks derived from taking higher derivatives of the friedmann equation . in that case
we would use constraints on the hubble parameter in more redshift bins , which would allow more freedom and hence more solutions overall to distinguish between ( even though here we are only discussing solutions about one set of values for the initial parameters ) .
hence this endevour will not be fruitful at this time . with much better data
, the solutions may be tractable .
gravity in the palatini formalism .
the constraints in the bottom panel are about one solution of the three non - linear equations of motion .
there are other solutions , see text for further discussion.,title="fig : " ] [ fig : pal ] we also note in passing that when we constrain just two @xmath24 s , these turn out to be constrained to non - zero values ( fig 4 ) .
a negative @xmath24 implies that the universe is not super - accelerating today , and positive @xmath25 implies that @xmath24 is increasing ( approaching zero ) today .
this is as expected .
when more @xmath24 s are involved , as in the remaining sections , they of course become less constrained .
it would be interesting to look also at other generalizations which have been considered in the literature @xcite-@xcite .
these models have actions of the form s=116gd^4x[r+ f(r , ) ] + s_m , where @xmath69 is the gauss - bonnet combination .
the equations of motion which follow can be written as & & ( 1+f)g_-__f + g_f -12 g_(f - fr- ) + & & -2r__+2rg_-4r_-4r_(^ _ ) _ _ + & & -4g_r^__+8r_(^_)_=8gt _ , where @xmath70 and @xmath71 and the partial derivatives should be found treating @xmath6 and @xmath2 as independent variables . in a frw background this becomes 3h^2(1+f+f ) + 12(f - fr-)+12h^4=8 g , [ eq : gsc ] where = 24ah^2=24h^3(h+h ) , and @xmath6 has been already introduced in equations ( [ derivr ] ) .
the taylor - expansion of @xmath17 about today up to second order in the scalars takes the form fh_0 ^ 2 & = & + ( |r-|r_0 ) + 12_1(|r-|r_0)^2 + & & + _ 2(|r-|r_0)(- ) + 12_3(-)^2 , where , f_0 , _ 1h_0 ^ 2._0 , _
2h_0 ^ 4._0 , _
3h_0 ^ 6._0 , with @xmath72 and @xmath73 .
it should be noted that there is no linear term in the expansion for @xmath2 , because such a term would give no contribution to the equations of motion , and the constant term in @xmath74 is considered absorbed in @xmath33 .
then one has f&=&+_1(|r-|r_0)+_2(- ) , + h_0 ^ 2&=&_2(|r-|r_0)+_3(- ) .
today equation ( [ eq : gsc ] ) becomes & = & 6[-1 + ( 1+ ) -6_1(4 + ^2 + ) + & & -48_2 ( 4 + 2^2 + ) + & & -96_3 ( 4 + 3^2 + ) + _ m ] .
taking four derivatives of equation ( [ eq : gsc ] ) allows us to solve for the five parameters in term of @xmath36 and the the derivatives of @xmath32 ( up to the 6th derivative ) .
one derivative gives & = & -1 + ( 2)^-1[6(4 - 7^2 - ^3 - 3 - 4 - )_1 ] + & & + ( 2)^-1[48(4 - 14^2 - 6^3 - 3 - 8 - ) _ 2 ] + & & + ( 2)^-1[96(4 - 21^2 - 15^3 - 3 - 12 - ) _ 3]- further derivatives give @xmath75 , @xmath76 and @xmath77 .
the later two particularly involve extremely complicated expressions that can not be written down here . when constraining 6 derivatives of the hubble parameter from data , it is expected that the constraints will be poor .
the top two panels of fig 5 show the constraints on the initial parameters . for reasons discussed in section
ii , we will now linearize the relations for the modified gravity parameters about the mean likelihood values of the inital parameters . in this way
we illustrate a space of possible solutions .
( there would be a number of isolated solutions not represented in these results , but because these are isolated solutions they will also be suppressed in the full posterior .
better data should reduce the number of such solutions . )
gravity in the metric formalism . ]
[ fig : fgb ] linearizing the previous equations about the mean values of the the derivatives of @xmath32 and @xmath36 one finds & = & -4.52918 + 14.6151 - 3.28575
+ -0.200232 + & & - 0.006872
+ 0.0004706
+ 8.555810 ^ -6 + 18.3942 _ m + & = & -0.42128 + 4.91222 - 1.71918 + 0.1071 + & & - 0.003735
+ 0.000257 + 4.653510 ^ -6 + 5.39085 _ m
_ 1&= & -0.008276 - 0.05569 + 0.03140 - 0.002392 + & & + 0.00004811 - 6.276410 ^ -6 - 1.0319710 ^ -7 - 0.0705 _ m + _ 2&= & 0.0006553 + 0.007342 - 0.005405 + 0.000428 - 6.240310 ^ -6 + & & + 1.207310 ^ -6 + 1.907610 ^ -8 + 0.01237 _ m + _ 3&= & 0.0003048 - 0.000411 + 0.0009694 - 0.00008026 + 6.9910310 ^ -7 + & & - 2.536810 ^ -7 - 3.835410 ^ -9 - 0.001939 _ m these relations together with the mcmc chains result in the constraints shown in the bottom panel of fig 5 .
the parameters are all consistent with zero , the significance is in the order of the constraints . we have checked using the full equations that the @xmath23 s are in fact well constrained to be close to zero as indicated by the plot .
this is because the relations contain combinations of the derivatives of @xmath32 in their denominators .
hence this is a consequence of the equations of motion .
this result is interesting . again
the @xmath24s=0 is a singular point for the @xmath23 s ( as well as for @xmath33 and @xmath22 the relations for which contain a division by @xmath24 ) .
again there is a solution allowed by data that is a singular point in this theory , but overall ( ie .
for most allowed combinations of the derivatives of @xmath11 ) the parameters of the theory are well behaved , and the distributions shown encompass the main range of values for these parameters that are acceptable as per current data .
@xmath4 and @xmath5 models , by definition , introduce higher derivatives of the expansion rate into the equations of motion .
this automatically requires data which can allow for precise measurements of derivatives of the hubble parameter .
this severe demand on observations arises because currently we have no theory and no symmetry which can exactly predict the form of @xmath4 .
this problem is evidently similar to the task of finding the form for the potential of a quintessence - like field , or more generally to finding a dynamical alternative to a cosmological constant . on the other hand
, we should not feel so complacent as to accept a tiny cosmological constant as the solution to the dark energy problem without giving due consideration to the daunting dynamical alternatives in the gravity sector . in order
to get rid of the possibilities discussed in this paper of modifying gravity , one may introduce the axiom that only a linear combination of lovelock terms can enter into the lagrangian for gravity , although this axiom does not follow from any symmetry consideration
. however , this would still not be enough to solve the cosmological constant problem , namely the zeroth lovelock term , the cosmological constant that we would predict from qft is far too large compared to the one needed by observations .
@xmath4 models , or more generally @xmath5 theories , thus remain a relevant consideration .
we have used distance measurements from cmb , bao and sne ia to place preliminary bounds on the parameters of @xmath4 models .
@xmath4 was taylor expanded about today , keeping terms upto second order in @xmath6 .
equations of motion were found in both the metric and palatini formalisms .
these allow us to solve for the coefficients of the taylor expansion of @xmath4 .
the equations of motion are non - linear , and we study the main space of possible solutions .
the solutions are interesting in that they reflect the order of magnitude of the coefficients allowed by data . under the metric formalism
we find that @xmath78 is small and negative over a substantial part of the allowed region , in which case in order for these models to have a consistent gr - like evolution at early times(see @xcite ) , this quantity must have changed sign some time in the past .
we have also analyzed a more general class of theories , @xmath79 . in order to modify gravity , without introducing spurious degrees of freedom ( such as extra ghost - like spin-2 particles )
, one needs to use lovelock scalars @xcite . in four dimensions the only such terms which give a non - zero contribution are a constant , @xmath6 , and @xmath2 .
these models are more general and the constraints are weaker . however , as for the @xmath4 theories , we should make a serious attempt to shrink the allowed parameter space for such theories in different ways ( using data and theoretical considerations ) .
in addition to the analysis presented here , at the moment , other considerations ( convergence to gr - like evolution at early times , instabilities ) are still the most serious issues that these theories need to survive .
for example , it is known that some models for @xmath4 and @xmath5 do not have a frw background consistent with gr at early times @xcite .
furthermore at high - redshift gr - like models may lead to unstable behaviour in the evolution of perturbations , because of the existence of either a tachionic mode or a mode with imaginary speed of propagation @xcite .
such considerations will further constrain the parameter space of the models considered here . in this paper
we have also not considered solar system type constraints , for reasons discussed in the introduction . to conclude , we have used current data to place constraints on the first and second derivatives of general functions of the ricci and the gauss - bonnet scalars , under different formalisms .
these derivatives are important quantities for theoretical considerations relating to ghosts and instabilities .
next generation dark energy surveys @xcite can measure the cosmic expansion history much more precisely ; this would dramatically shrink the presently allowed parameter space of the modified gravity models considered here .
we thank andrew liddle and mark hindmarsh for helpful comments . adf and pm
are supported by stfc , uk .
yw is supported in part by nsf career grants ast-0094335
. 99 a. g. riess , et al .
j. , 116 , 1009 ( 1998 ) s. perlmutter , et al .
, apj , 517 , 565(1999 ) r. h. dicke , phys .
* 125 * ( 1962 ) 2163 .
s. m. carroll , a. de felice and m. trodden , phys .
d * 71 * ( 2005 ) 023525 [ arxiv : astro - ph/0408081 ] .
l. perivolaropoulos , phys .
d * 67 * ( 2003 ) 123516 [ arxiv : hep - ph/0301237 ] .
v. faraoni , int .
j. mod .
d * 11 * ( 2002 ) 471 [ arxiv : astro - ph/0110067 ] .
a. riazuelo and j. p. uzan , phys .
d * 66 * ( 2002 ) 023525 [ arxiv : astro - ph/0107386 ] .
j. p. uzan , phys .
d * 59 * ( 1999 ) 123510 [ arxiv : gr - qc/9903004 ] .
f. perrotta and c. baccigalupi , phys .
d * 65 * ( 2002 ) 123505 [ arxiv : astro - ph/0201335 ] .
v. pettorino , c. baccigalupi and g. mangano , jcap * 0501 * ( 2005 ) 014 [ arxiv : astro - ph/0412334 ] .
x. l. chen and m. kamionkowski , phys .
d * 60 * ( 1999 ) 104036 [ arxiv : astro - ph/9905368 ] .
d. f. torres , phys .
d * 66 * ( 2002 ) 043522 [ arxiv : astro - ph/0204504 ] . c. santos and r. gregory , annals phys .
* 258 * ( 1997 ) 111 [ arxiv : gr - qc/9611065 ] .
a. de felice and m. trodden , phys .
d * 72 * ( 2005 ) 043512 [ arxiv : hep - ph/0412020 ] .
a. de felice , g. mangano , p. d. serpico and m. trodden , phys .
d * 74 * , 103005 ( 2006 ) [ arxiv : astro - ph/0510359 ] .
s. m. carroll , v. duvvuri , m. trodden and m. s. turner , phys .
d * 70 * , 043528 ( 2004 ) [ arxiv : astro - ph/0306438 ] .
s. capozziello , s. carloni and a. troisi , arxiv : astro - ph/0303041 .
s. tsujikawa , arxiv:0705.1032 [ astro - ph ] .
s. fay , r. tavakol and s. tsujikawa , phys .
d * 75 * , 063509 ( 2007 ) [ arxiv : astro - ph/0701479 ] . l. amendola , r. gannouji , d. polarski and s. tsujikawa , phys .
d * 75 * , 083504 ( 2007 ) [ arxiv : gr - qc/0612180 ] . s. a. appleby and r. a. battye , arxiv:0705.3199 [ astro - ph ] .
m. s. movahed , s. baghram and s. rahvar , arxiv:0705.0889 [ astro - ph ] .
i. sawicki and w. hu , arxiv : astro - ph/0702278 .
r. bean , d. bernat , l. pogosian , a. silvestri and m. trodden , phys . rev .
d * 75 * , 064020 ( 2007 ) [ arxiv : astro - ph/0611321 ] . v. faraoni , phys .
d * 72 * , 124005 ( 2005 ) [ arxiv : gr - qc/0511094 ] .
y. s. song , w. hu and i. sawicki , phys .
d * 75 * , 044004 ( 2007 ) [ arxiv : astro - ph/0610532 ] .
l. randall and r. sundrum , phys .
lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] .
g. r. dvali , g. gabadadze and m. porrati , phys .
b * 485 * , 208 ( 2000 ) [ arxiv : hep - th/0005016 ] . c. deffayet , g. r. dvali , g. gabadadze and a. i. vainshtein , phys .
d * 65 * , 044026 ( 2002 ) [ arxiv : hep - th/0106001 ] . c. deffayet , g. r. dvali and g. gabadadze , phys .
d * 65 * , 044023 ( 2002 ) [ arxiv : astro - ph/0105068 ] .
k. koyama , phys .
d * 72 * , 123511 ( 2005 ) [ arxiv : hep - th/0503191 ] .
s. m. carroll , a. de felice , v. duvvuri , d. a. easson , m. trodden and m. s. turner , phys .
d * 71 * , 063513 ( 2005 ) [ arxiv : astro - ph/0410031 ] .
o. mena , j. santiago and j. weller , arxiv : astro - ph/0510453 .
i. navarro and k. van acoleyen , phys .
b * 622 * , 1 ( 2005 ) [ arxiv : gr - qc/0506096 ] .
a. de felice , m. hindmarsh and m. trodden , jcap * 0608 * , 005 ( 2006 ) [ arxiv : astro - ph/0604154 ] .
g. calcagni , b. de carlos and a. de felice , nucl .
b * 752 * , 404 ( 2006 ) [ arxiv : hep - th/0604201 ] .
m. sahlen , a. r. liddle and d. parkinson , phys .
d * 75 * , 023502 ( 2007 ) [ arxiv : astro - ph/0610812 ] .
w. hu and i. sawicki , arxiv:0705.1158 [ astro - ph ] .
t. chiba , t. l. smith and a. l. erickcek , arxiv : astro - ph/0611867 .
a. rajaraman , arxiv : astro - ph/0311160 .
t. multamaki and i. vilja , phys .
d * 74 * , 064022 ( 2006 ) [ arxiv : astro - ph/0606373 ] .
v. faraoni , phys .
d * 74 * , 023529 ( 2006 ) [ arxiv : gr - qc/0607016 ] .
t. faulkner , m. tegmark , e. f. bunn and y. mao , arxiv : astro - ph/0612569 .
g. allemandi , m. francaviglia , m. l. ruggiero and a. tartaglia , gen .
grav . * 37 * , 1891 ( 2005 ) [ arxiv : gr - qc/0506123 ] .
d. f. mota and j. d. barrow , mon . not .
soc . * 349 * , 291 ( 2004 ) [ arxiv : astro - ph/0309273 ] .
i. navarro and k. van acoleyen , jcap * 0702 * , 022 ( 2007 ) [ arxiv : gr - qc/0611127 ] .
a. d. dolgov and m. kawasaki , phys .
b * 573 * , 1 ( 2003 ) [ arxiv : astro - ph/0307285 ] .
a. nunez and s. solganik , phys .
b * 608 * , 189 ( 2005 ) [ arxiv : hep - th/0411102 ] .
t. chiba , jcap * 0503 * , 008 ( 2005 ) [ arxiv : gr - qc/0502070 ] .
a. g. riess _ et al .
_ , arxiv : astro - ph/0611572 .
y. wang and p. mukherjee , arxiv : astro - ph/0703780 .
d. n. spergel _ et al .
_ , arxiv : astro - ph/0603449 .
d. j. eisenstein _ et al .
_ [ sdss collaboration ] , astrophys . j. * 633 * , 560 ( 2005 ) [ arxiv : astro - ph/0501171 ] .
y. wang and p. mukherjee , astrophys .
j. * 606 * , 654 ( 2004 ) [ arxiv : astro - ph/0312192 ] .
y. wang and m. tegmark , phys .
. , 92 , 241302 ( 2004 ) [ arxiv : astro - ph/0403292 ] y. wang and m. tegmark , phys .
d * 71 * , 103513 ( 2005 ) [ arxiv : astro - ph/0501351 ] .
t. p. sotiriou , phys .
d * 73 * , 063515 ( 2006 ) [ arxiv : gr - qc/0509029 ] .
g. j. olmo and w. komp , arxiv : gr - qc/0403092 .
s. capozziello , v. f. cardone and m. francaviglia , gen .
grav . * 38 * , 711 ( 2006 ) [ arxiv : astro - ph/0410135 ] .
a. hindawi , b. a. ovrut and d. waldram , phys .
d * 53 * , 5583 ( 1996 ) [ arxiv : hep - th/9509142 ] .
a. de felice and m. hindmarsh , arxiv:0705.3375 [ astro - ph ] .
baojiu li , john d. barrow , david f. mota , arxiv:0705.3795 [ gr - qc ] a. albrecht ; g. bernstein ; r. cahn ; w. l. freedman ; j. hewitt ; w. hu ; j. huth ; m. kamionkowski ; e.w .
kolb ; l. knox ; j.c .
mather ; s. staggs ; n.b .
suntzeff , report of the dark energy task force [ arxiv : astro - ph/0609591 ] see for example , http://www.astro.ubc.ca/lmt/alpaca/ ; http://www.lsst.org/ ; http://www.as.utexas.edu / hetdex/. @xcite contains a more complete list of future dark energy experiments .
y. wang , et al .
, baas , v36 , n5 , 1560 ( 2004 ) ; a. crotts , et al .
( 2005 ) [ arxiv : astro - ph/0507043 ] ; e. cheng ; y. wang ; et al .
, proc . of spie ,
6265 , 626529 ( 2006 ) ; http://jedi.nhn.ou.edu/
consider the vector @xmath80 .
if the hubble parameter is measured in @xmath81 redshift bins , one can solve for @xmath81 of its ( present day ) derivatives by assuming a truncated taylor expansion for each of @xmath82 about today upto order @xmath81 , which gives @xmath83 this is a linear equation , @xmath84 , where @xmath85 , which can be inverted to find derivatives of @xmath32 ( today ) .
we then change variables from @xmath86 to @xmath53 , where @xmath87 .
one has @xmath88^{\!i}{\mathcal{h}}\,\ ] ] so that @xmath89 in fact it can be shown that @xmath90 with @xmath91 this then specifies the whole series of polynomials . writing the linear relation as @xmath92 we have @xmath93 or @xmath94 for two bins case ( @xmath95 , and @xmath96 for current data ) we get m= ( cc - 2.85714 & 0.714286 + - 1.22449 & 1.32653 ) for three bins ( @xmath97 , @xmath98 , and @xmath99 ) then we get m= ( ccc - 6.4286 & 3.2143 & - 0.71429 + - 16.531 & 15.153&- 3.8775 + 32.930&- 22.369 & 3.2216 ) for 4 bins equally spaced out to @xmath100 , we get m= ( cccc - 11.429 & 8.5714&- 3.8095 & 0.71428 + - 59.320 & 68.980&-34.286 & 6.7687 + - 9.0962 & 55.802&- 52.789&13.251 + 509.11&-745.18 & 450.18&- 91.607 ) for 5 bins , we have m= ( ccccc - 17.857 & 17.857&- 11.905 & 4.4647&- 0.71429 + - 145.83&209.61&- 153.91 & 60.374&-9.9150 + - 335.37 & 679.30&- 630.53 &
276.88&-48.546 + 1445.8&- 2258.6 & 1643.0&- 562.54 & 79.241 + 2094.7&- 6287.7 & 7809.3&- 4285.9 & 838.55 ) and , finally , for 6 bins , we have m= ( cccccc - 19.3 & 16.109&- 7.193 & 0.03758 & 1.271&- 0.3546 + - 179.35 & 218.78&- 108.04 & 0.67106 & 20.4867&- 5.823 + - 500.66 & 833.8&- 523.69 & 4.9750 & 118.16&- 35.142 + 1837.10&- 2403.5 & 1168.7 & 4.745&- 149.15 & 32.4457 + 4187.9&- 9101.2 & 7205.2&- 112.35&- 2310.5 & 732.79 + - 54953 & 91546&- 59415 & 191.52 & 14897&- 4190.4 ) with the present day derivatives of the hubble parameter thus derived from model independent measurements of the hubble parameter in linear redshift bins from cosmological data , we can then obtain constraints on parameters of modified gravity models , using appropriate equations of motion . this is done in sections ii , iii and iv of this paper .
we could find the following distinct solutions of the equations @xmath101 for @xmath33 , @xmath22 , and @xmath23 , about the mean values for @xmath102 : [ -0.857943 , -0.428971 , -0.14299 ] , [ -1.46637 , -0.422284 , -0.140728 ] , [ 0.399627 , 0.824305 , 0.852729 ] , [ 3.78266 , 0.748323 , -0.00792803 ] , [ 0.222102 , 0.00488155 , 4.12207e-05 ] , [ -0.739164 , -0.00165984 , 0.0000280094 ] , [ -0.866737 , -0.00527848 , 0.000315044 ] , [ 0.252698 , -0.00306054 , 0.0000152404 ] , [ -0.29676 , 0.00333869 , 0.0000390608 ] , [ -0.811148 , -0.00218394 , 0.0000614933 ] , [ -0.410838 , -0.0124848 , -0.0000849403 ] , [ -0.132685 , -0.00276524 , 0.0000205499 ] , and [ 0.00810481 , -0.00504213 , -0.00002751 ] . | modified gravity provides a possible explanation for the currently observed cosmic accelaration . in this paper
, we study general classes of modified gravity models .
the einstein - hilbert action is modified by using general functions of the ricci and the gauss - bonnet scalars , both in the metric and in the palatini formalisms .
we do not use an explicit form for the functions , but a general form with a valid taylor expansion up to second order about redshift zero in the riemann - scalars .
the coefficients of this expansion are then reconstructed via the cosmic expansion history measured using current cosmological observations .
these are the quantities of interest for theoretical considerations relating to ghosts and instabilities .
we find that current data provide interesting constraints on the coefficients .
the next - generation dark energy surveys should shrink the allowed parameter space for modifed gravity models quite dramatically . |
cavity optomechanics is an exciting , novel experimental platform that will allow us to explore fundamental questions of quantum mechanics and paves the way for applications in quantum - information processing , high - precision metrology , and gravitational - wave detection @xcite .
it was recognized early on that cavity optomechanical systems can squeeze light similar to a nonlinear kerr medium @xcite . experimentally , this ponderomotive squeezing has recently been demonstrated as well @xcite . as far as correlations between photons
is concerned , photon antibunching has been predicted @xcite to occur in the single - photon strong - coupling regime @xcite and two - mode optomechanical systems @xcite .
moreover , two - time photon correlation functions have been proposed as a means to observe the onset of parametric instability @xcite and unconventional photon blockade @xcite .
so far , most research in the field of optomechanics has focused on dispersive coupling , where mechanical motion modulates the resonance frequency of the cavity .
et al_. @xcite have proposed a novel kind of optomechanics , where mechanical motion modulates the linewidth of the cavity . in this case
, the radiation pressure force spectrum features a fano resonance modifying the interaction between light and mechanics dramatically .
cooling of the mechanical oscillator with dissipative coupling has been predicted @xcite and recently also been demonstrated @xcite .
however , to date most of the properties of dissipative optomechanics remain unexplored . in particular , the experimental progress in this direction
motivates us to investigate the photonic properties of this setup . in this work ,
we study theoretically the potential of dissipative optomechanics as a source for squeezed light and correlated photons .
our analysis reveals that the system can generate strong squeezing of light and we predict oscillations of the photon correlation function with a suppression below shot noise at finite times .
it turns out that both the squeezing spectrum and the photon correlation function can be understood in terms of the normal modes of the linearized system .
the remainder of this paper is organized as follows . in sec .
[ model ] we introduce the hamiltonian of an optomechanical system with dispersive as well as dissipative coupling and derive the linearized equations of motion describing this system . in sec . [ squeezing ] we present our results on the squeezing spectrum and in sec .
[ correlations ] we discuss the properties of the second - order photon correlation function .
we conclude in sec .
[ conclusion ] .
we study an optomechanical system in which a mechanical degree of freedom modulates the resonance frequency @xmath1 ( dispersive coupling ) and the linewidth @xmath2 ( dissipative coupling ) of a cavity mode .
the hamiltonian ( @xmath3 ) is given by @xcite @xmath4(\hat{b } + \hat{b}^\dagger ) \label{hamiltonian}\end{aligned}\ ] ] where the dispersive and dissipative coupling strengths are @xmath5 and @xmath6 , respectively .
the first and second term in ( [ hamiltonian ] ) describe the cavity mode ( frequency @xmath1 ) with bosonic field operator @xmath7 and the mechanical oscillator ( frequency @xmath8 ) with bosonic field operator @xmath9 .
the mechanical displacement @xmath10 has zero - point fluctuations @xmath11 with mass @xmath12 . @xmath13 and @xmath14 describe the damping of the cavity due to the optical and mechanical baths , respectively . here , @xmath15 are bosonic field operators of the optical bath , @xmath16 is the density of states of the optical bath , @xmath2 and @xmath17 are optical and mechanical damping rates , respectively .
fluctuations in the input from the optical ( mechanical ) bath are described by operators @xmath18 and @xmath19 , respectively . in our analysis
, we assume markovian baths , where the mechanical bath is characterized by a thermal phonon number @xmath20 with inverse temperature @xmath21 , whereas the optical bath is assumed to be at zero temperature .
the non - zero expectation values hence are @xmath22 and @xmath23 for the mechanical input , and @xmath24 for the optical input . writing @xmath25 with @xmath26 real , we decompose the optical field into mean amplitude @xmath26 and fluctuations @xmath27 and move to a frame rotating at the laser frequency @xmath28 .
then , we employ standard input - output theory @xcite and linearize the langevin equations of motion @xcite @xmath29 where @xmath30 @xmath31 and @xmath32 with the coefficients @xmath33 and @xmath34 and @xmath35 the laser detuning .
the linearized input - output relation for the optical field is @xcite @xmath36 from the exact solution @xcite to the linearized equations of motion ( [ eom ] ) we find that the following two steady - state correlation functions @xmath37 as well as @xmath38 are given by @xmath39 \ , \bar{a } ^2 \ , \alpha ( - \omega)\ , \alpha ^ * ( \omega)}{q(-\omega ) } \nonumber \\
n(\omega ) = & \frac{\kappa \bar{a}^2 |\alpha ( - \omega)|^2 } { | q ( \omega ) |^2 } \left[4 \kappa \bar{a}^2 \omega_m ^2 |\alpha ( \omega)|^2 \right .
\nonumber\\ & \left .
+ \gamma ( \bar{n}+1 ) |\chi_m ^{-1 } ( - \omega ) | ^2 + \gamma \bar{n } |\chi_m ^{-1 } ( \omega ) |
^2 \right ] \label{expecvalues}\end{aligned}\ ] ] with the cavity response function @xmath40^{-1 } $ ] , the mechanical response @xmath41^{-1 } $ ] , the optomechanical self - energy @xmath42 where @xmath43 $ ] , @xmath44 $ ] , @xmath45 $ ] , and @xmath46 with @xmath47 , @xmath48 $ ] and @xmath49 with @xmath50 .
in this section we will investigate the squeezing of the outgoing light by a dissipative optomechanical system ( oms ) .
squeezing is conveniently characterized by the spectrum of squeezing @xmath51 where @xmath52 and @xmath53 . in the following we concentrate on the _ optimal _ squeezing spectrum @xcite @xmath54 which can be expressed in terms of the two correlation functions @xmath55 and @xmath56 given in eq .
( [ expecvalues ] ) above . choosing the mechanical frequency @xmath57 to be unity
, the optimal squeezing spectrum @xmath58 is a function of detuning @xmath59 , cavity linewidth @xmath2 , coupling strength @xmath60 and @xmath61 , as well as thermal phonon number @xmath62 .
note that the single - photon coupling strengths , @xmath63 and @xmath64 , and the intracavity amplitude @xmath26 only appear as a product and not individually in eqs .
( [ expecvalues ] ) and ( [ optsqueezing ] ) .
[ spectrafig ] ( a ) shows the optimal squeezing spectrum @xmath58 for a dissipative oms as a function of detuning @xmath59 .
shaded areas correspond to regions where the solution of the linearized equations of motion ( [ eom ] ) is unstable , i.e. the eigenvalues of the matrix @xmath65 ( [ matrix ] ) do not all have negative real parts . focusing on the stable regions , we find two different types of behavior . on the red - detuned side @xmath66 we observe two dips below shot - noise level @xmath67 . in fig .
[ spectrafig ] we also plot the normal - mode frequencies of the system obtained from the eigenvalues of the matrix @xmath65 ( [ matrix ] ) .
we see that the squeezing dips coincide with the normal - mode frequencies of the system .
this is qualitatively similar for the dispersive oms shown in fig .
[ spectrafig ] ( b ) .
mathematically , this is a consequence of the fact that @xmath68 is the determinant of the matrix @xmath65 .
physically , this corresponds to the fact that the oms will respond most strongly close to its resonances , i.e. normal - mode frequencies , and the large response can in turn strongly affect the outgoing light field . in fig .
[ spectrafig ] ( b ) we see that in the limit @xmath69 squeezing is small for dispersive coupling at @xmath70 .
stronger squeezing can be achieved around resonance @xmath71 close to the mechanical frequency @xmath72 and in particular close to zero frequency @xmath73 @xcite . for dissipative coupling , and in contrast to the dispersive case , there is a window of stability on the blue - detuned side @xmath74 . for small positive detuning @xmath59 strong
squeezing occurs at frequencies close to the mechanical frequency @xmath72 .
a special point is @xmath75 where the squeezing spectrum has a single dip close to the frequency of the optical - like mode @xmath76 .
moreover , strong squeezing also occurs at the point where the two normal - mode frequencies merge close to the onset of the parametric instability .
[ spectrafig ] ( c ) shows the optimal squeezing spectrum @xmath58 as a function of the dissipative coupling strength @xmath61 for a system driven on the red sideband @xmath77 . at weak coupling
the two normal modes are degenerate leading to a single dip in the squeezing spectrum . at larger coupling the degeneracy
is broken and as normal - mode splitting ( nms ) develops two dips emerge in the squeezing spectrum @xmath58 .
nms in the squeezing spectrum is also present in the dispersive case , see fig .
[ spectrafig ] ( d ) .
note that squeezing on the sideband @xmath77 is much weaker in general for dispersive coupling .
[ spectrafig ] ( e ) shows the optimal squeezing spectrum @xmath58 of a dissipative oms as a function of cavity linewidth @xmath2 . for @xmath78
we find two narrow dips in the squeezing spectrum . for larger cavity linewidth @xmath2 the dips broaden and merge into a single broad region of squeezing .
this is in contrast to the case of a dispersive oms , see fig .
[ spectrafig ] ( f ) , where there is only weak squeezing in the resolved sideband limit @xmath78 , and a broad dip in the squeezing spectrum occurs in the bad - cavity limit @xmath79 .
we note that since ground - state cooling for dispersive coupling is possible only in the resolved - sideband limit , a dissipative oms offers the advantage that strong squeezing and ground - state cooling can be realized simultaneously .
we study next the photon correlations in the outgoing light of a dissipative oms . applying wick s theorem , which is exact for the gaussian states considered here , and using the decomposition of field operators @xmath25 , one obtains the second - order photon correlation function @xmath80 @xcite @xmath81 + 2 \mathrm{re}\left[(\bar{a}^*)^2 _ \text{out } s(\tau)\right]}{[|\bar{a}_\text{out}|^2 + n(0)]^2 } \nonumber
\\ & + \frac{|n(\tau)|^2 + |s(\tau)|^2 } { [ |\bar{a}_\text{out}|^2 + n(0)]^2 } \label{g2}\end{aligned}\ ] ] where the output field given by @xmath82 , @xmath83 , and @xmath84 .
the correlation functions can be found by fourier transforming the analytic expressions ( [ expecvalues ] ) or solving numerically their equation of motion that can be derived from eq .
( [ eom ] ) .
we find that @xmath85 shows damped oscillations with one or several frequencies depending on the choice of parameters , see fig .
[ corrfunc ] ( a ) , ( c ) , ( e ) . to identify the spectral contributions we calculate numerically @xmath86 , the fourier transform of @xmath87 , as shown in fig .
[ corrfunc ] ( b ) , ( d ) , ( f ) .
we see that in ( a ) the dynamics of @xmath85 is dominated by one spectral component close to the mechanical frequency @xmath57 . in fig .
[ corrfunc ] ( b ) we also plot a lorentzian whose linewidth and frequency are given by the mechanical - like normal mode , and in fig .
[ corrfunc ] ( a ) we plot the corresponding exponential envelope . from this analysis , we conclude that in this case @xmath85 is dominated by the mechanical - like normal mode , i.e. its frequency is the ( effective ) mechanical frequency and the linewidth is the ( effective ) mechanical linewidth .
in contrast , in fig .
[ corrfunc ] ( c ) and ( d ) , the frequency and linewidth are determined by the optical - like normal mode and its properties . in fig .
[ corrfunc ] ( e ) and ( f ) , both normal modes contribute significantly leading to a beating in @xmath85 , as it can be nicely seen in the spectral domain @xmath88 .
it is worth noting that we observe @xmath0 at finite times @xmath89 , i.e. the oms suppresses the probability of photon arrival in the outgoing light beam which may be useful for applications .
in addition to the response at the normal - mode frequencies , fig .
[ corrfunc ] ( b ) shows small features at zero and twice the normal - mode frequency .
these are due to the non - linear terms @xmath90 and @xmath91 in ( [ g2 ] ) . while they are important at the single - photon level @xmath92 , they become gradually less important in the limit of large intracavity field @xmath93 @xcite . to investigate their effect
further we plot in fig .
[ a - dep ] ( a ) , ( c ) , ( e ) the exact photon correlation function @xmath80 as well as @xmath94 + 2 \mathrm{re}\left[(\bar{a}^*)^2 _ \text{out } s(\tau)\right]}{[|\bar{a}_\text{out}|^2 + n(0)]^2}\ ] ] neglecting the nonlinear terms @xmath90 and @xmath91 in ( [ g2 ] ) . in fig .
[ a - dep ] ( b ) , ( d ) , ( f ) we plot the exact correlation spectra @xmath86 as well as @xmath95 , the fourier transform of @xmath96 , @xmath97 + \textrm{re}[(\bar{a}^*)^2 _ \text{out } s(\omega)]}{[|\bar{a}_\text{out}|^2 + n(0)]^2}. \label{g2spec}\ ] ] in fig .
[ a - dep ] ( b ) we clearly observe the non - linear contributions , @xmath90 and @xmath98 , at zero and twice the normal - mode frequency which modify @xmath80 dramatically , see fig .
[ a - dep ] ( a ) .
we note in passing that for single - photon coupling strength @xmath99 as shown in ( a ) one should compare our results to numerical solutions of the quantum master equation .
in contrast to the squeezing spectrum @xmath100 ( [ optsqueezing ] ) , the photon correlation function @xmath80 depends not only on the product @xmath61 of coupling strength and intracavity amplitude , but also on the intracavity amplitude @xmath26 itself . inspecting the scaling in ( [ g2 ] )
we find that photon correlations can be large for small cavity amplitudes @xmath92 , but approach those of coherent states @xmath101 for @xmath102 with @xmath61 fixed , see fig .
[ a - dep ] @xcite . for a large intracavity field @xmath93 , where @xmath95 is a good approximation , the photon correlation function @xmath103 only depends on detuning @xmath59 , cavity linewidth @xmath2 , coupling strength @xmath61 , and thermal phonon number @xmath62 . in fig
. [ corrdet ] ( a ) we show the correlation spectrum @xmath95 of a dissipative oms as a function of detuning @xmath59 .
the shaded areas correspond to regions where the solution to the linearised equations of motion is unstable .
similar to the squeezing spectrum @xmath100 the correlation spectrum @xmath104 can be understood in terms of the normal modes of the system whose frequencies we also plot in fig .
[ corrdet ] .
we identify regions where the ( strongly broadened ) mechanical - like normal mode dominates the dynamics of the photon correlation dynamics around @xmath71 , like in fig .
[ corrfunc ] ( a ) , the special point @xmath75 where the optical - like normal mode dominates , like in fig . [ corrfunc ] ( b ) , or other cases , like in fig .
[ corrfunc ] ( c ) , where both modes contribute significantly . before concluding we note that as expected with increasing thermal phonon number @xmath62 the quantum features in the second - order correlation function as well as the squeezing spectrum gradually vanish .
we investigated dissipative oms as a source of squeezed light and correlated photons .
strong squeezing occurs at the normal - mode frequencies of the system , and in contrast to dispersive oms , even in the sideband - resolved limit .
the photon correlation function @xmath80 exhibits damped oscillations whose frequency and linewidth are given by the properties of the normal modes .
depending on parameters this is the mechanical frequency and its ( effective ) decay time , or the optical detuning and cavity linewidth , or both modes contribute significantly leading to a beating in the correlation function .
an would like to thank talitha weiss and niels lrch for discussions .
an holds a university research fellowship from the royal society and acknowledges support from the winton programme for the physics of sustainability .
22ifxundefined [ 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.86.1391 [ * * , ( ) ] link:\doibase 10.1103/physreva.49.1337 [ * * , ( ) ] link:\doibase 10.1103/physreva.49.4055 [ * * , ( ) ] http://dx.doi.org/10.1038/nature11325 [ * * , ( ) ] http://dx.doi.org/10.1038/nature12307 [ * * , ( ) ] link:\doibase 10.1103/physrevx.3.031012 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.063601 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.013847 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.063602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.063601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.013603 [ * * , ( ) ] http://stacks.iop.org/0953-4075/46/i=3/a=035502 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.013818 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.253601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.207209 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.213604 [ * * , ( ) ] http://stacks.iop.org/1367-2630/15/i=4/a=045017 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.023850 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.043601 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physreva.90.063824 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.063815 [ * * , ( ) ] | we study theoretically the squeezing spectrum and second - order correlation function of the output light for an optomechanical system in which a mechanical oscillator modulates the cavity linewidth ( dissipative coupling ) .
we find strong squeezing coinciding with the normal - mode frequencies of the linearized system .
in contrast to dispersive coupling , squeezing is possible in the resolved - sideband limit simultaneously with sideband cooling .
the second - order correlation function shows damped oscillations , whose properties are given by the mechanical - like , the optical - like normal mode , or both , and can be below shot - noise level at finite times , @xmath0 . |
the nature of dark matter ( dm ) is one of the major mysteries in our current understanding of the universe .
a comprehensive strategy encompassing direct and indirect detection plus collider searches , as well as indirect information from cosmological observations , will certainly be necessary to determine the fundamental nature of the dm .
the results from the dama / libra collaboration @xcite and , more recently , from the cogent @xcite , cresst @xcite and cdms @xcite experiments , might point to the first direct signals of light weakly interacting massive particles ( wimps ) scattering off of nuclei @xcite .
the gamma - ray excess from the galactic center region @xcite and the tentative gamma - ray line observed with the fermi large area telescope @xcite have also been interpreted as resulting from wimp annihilation .
the excess cosmic - ray positrons , recently confirmed by ams-02 @xcite , is an additional puzzling result that some have associated with wimp annihilation @xcite ( see however @xcite ) . besides these putative direct and indirect detection signals ,
recent measurements of the angular power spectrum of the cosmic microwave background ( cmb ) by a variety of telescopes and satellites seem to indicate the existence of an extra component of radiation in the early universe . in particular
, the wmap collaboration has recently presented their 9-year data and concluded that , after combining with data from baryonic acoustic oscillations ( bao ) , from the atacama cosmology telescope ( act ) , and from the hubble space telescope ( hst ) ( @xmath0 , @xcite ) , the number of effective massless neutrinos is @xmath1 @xcite .
the situation is , however , quite complex and subtle when the results of other experiments are considered .
for example , the spt collaboration has also recently reported new results . combining their data with bao and hst
, they concluded that @xmath2 @xcite .
depending on the data set used , these conclusions can , however , qualitatively change .
for instance , combining data from wmap9 and spt , ref .
@xcite finds @xmath3 , while from a combination of wmap9 , spt , and the 3-year supernova legacy survey ( snls3 ) @xcite , @xmath4 @xcite .
act also reported recently a much lower value than what quoted above , namely @xmath5 @xcite , when utilizing wmap7 data , and @xmath6 when including wmap9 data @xcite .
this value can , however , be much larger if one includes bao and hst data , reaching @xmath7 @xcite .
it is clear that spt and act results are somewhat in tension , and this may be related to the different values used for the lensing amplitude parameter @xmath8 @xcite .
lastly , the eagerly awaited planck results have just been reported . at face value ,
planck data alone seem to disfavor an extra radiation component , pointing to @xmath9 @xcite .
it is important to note , however , that the planck collaboration adopted fairly low values ( at the @xmath10 level ) for @xmath11 compared to previous studies @xcite .
since @xmath12 and @xmath11 are positively correlated , increasing @xmath11 would naturally yield higher values for @xmath12 @xcite .
interestingly , adding the @xmath11 measurement and bao data , the planck collaboration finds , in fact , @xmath13 @xcite .
this latter value is @xmath14 above the @xmath15 value favored by big bang nucleosynthesis ( bbn ) @xcite . as the discussion above implies ,
the best fit value for @xmath12 varies significantly depending on the data set used .
regardless of that , the combined data reviewed above seem to be tentatively pointing towards the existence of an extra radiation component .
interestingly , it has been shown that this additional component can be interpreted as a thermally produced light dm species @xcite , or as _ non - thermally _ produced _ relativistic _ dm particles , as shown in a model - independent way in ref .
@xcite , and in the context of supersymmetric frameworks in ref .
@xcite . in this article , we present a few illustrative models , including a supersymmetric example , where the tentative extra radiation component is interpreted in terms of wimp dm primarily composed of a standard cold component plus a small fraction ( roughly @xmath16 or less ) produced non - thermally in a relativistic state . in section [ sec : derive ] we relate the energy density associated with non - thermally , relativistically produced dm with the number of effective neutrinos . in section
[ sec : cosmo ] we examine the cosmological bounds from bbn , structure formation and the cmb on this scenario . in section
[ sec : models ] we present four models that satify the cosmological bounds and produce @xmath17 .
we summarize and conclude in section [ sec : conclusion ] .
in order to relate the energy density associated with non - thermally , relativistically produced dm with the number of effective neutrinos , we start by calculating the ratio between their respective energy densities .
since the cold dark matter ( cdm ) and neutrino densities are given by @xmath18 and @xmath19 , at matter - radiation equality ( mre ) , the ratio between neutrino and dm energy density is @xmath20 where @xmath21 , @xmath22 , @xmath23 is the number of neutrinos , and @xmath24 is the scale factor at mre . for @xmath25 , we thus find that the energy density of one neutrino is @xmath26 of the cdm density . as a result ,
if dm particles had a kinetic energy equivalent to @xmath27 at mre , this fraction would produce the same effect as an extra neutrino species in the expansion of the universe at mre , as already pointed out in ref .
@xcite .
assuming that some fraction @xmath28 of the dm in the universe is produced via the decay @xmath29 , we can relate the boost factor of the dm particle produced in the decay process ( see appendix [ boostfactor ] ) with the number of additional effective neutrinos in the universe via @xmath30 f. \label{deltaneff}\end{aligned}\ ] ] eq .
( [ deltaneff ] ) illustrates that this scenario posseses three free parameters : * the @xmath31 lifetime for the decay process , * the mass ratio @xmath32 , and * the fraction @xmath28 of the dm density produced via the decay . with specific choices for the 3 parameters ,
one can easily reproduce @xmath33 and therefore mimic the effect of an extra effective relativistic degree of freedom @xcite .
it is important to point out that in the present setup most of the dm would still be produced thermally , for example via the standard thermal freeze - out picture , with only a small fraction being produced non - thermally and with a large kinetic energy .
we show in the next section that the fraction of dm produced relativistically must be small ( @xmath34 ) due to the fact this component directly affects the formation of structure and other cosmological bounds .
for example , the electromagnetic energy released in the decay could have catastrophic impacts on the abundance of light elements as synthesized in the big bang nucleosynthesis ( bbn ) framework , as well as on the spectrum of the cosmic microwave background radiation ( cmb ) .
we analyze in detail structure formation , bbn and cmb bounds on @xmath28 and on the @xmath31 lifetime in the following section .
it is well known that the primordial universe had small inhomogeneities in the density distribution of matter , which evolved linearly at first , while eventually growing non - linearly to form the structures observed today . in the standard picture of structure formation
, the dark matter played a crucial role in seeding the growth of structure at an acceptable rate .
a crucial ingredient to this picture is that the dm particles must have had a negligible kinetic energy at the time of structure formation . if not
, the formation of structures at large scales would be slowed down because dm particles with large kinetic energies would not cluster at sufficiently small scales due to free - streaming .
this slowing down the growth of structures would wash out matter density fluctuations .
it is therefore critical to investigate the suppression on the growth of structure caused by the fraction of dm which is non - thermally produced in a relativistic state via the aforementioned decay process , and to derive quantitative bounds on the fraction of dm particles produced in this manner .
it has been shown that , at scales above the free - streaming length , relativistic dm particles effectively behave like cold dm . at small scales ,
the matter fluctuations of cold dm particles is governed by a linear equation which , during the matter - dominated stage of the universe , follows the power - law behavior @xcite @xmath35 in the equation above , @xmath36 is the growth rate of the cold dm field , which during the matter - dominated epoch is given , to first order , by @xcite @xmath37 where @xmath28 is , as above , the fraction of the dm density produced non - thermally in a relativisitic state , and where @xmath34 .
it is then possible to determine the suppression on small - scales caused by the non - thermal production of relativistic dm particles by comparing the matter fluctuation given in eqs.([delta])-([alpha ] ) with the pure cold dm case at mre . to first order
, we find @xmath38 as mentioned above , it is important to notice that eq .
( [ g ] ) is applicable only in the matter - dominated regime , and a more accurate calculation considering the effects of , e.g. , a cosmological constant , is out of the scope of this work . after combining measurements of the amplitude of matter fluctuations in the universe on scales of @xmath39 mpc from the wmap9 results @xcite with lyman - alpha forest data @xcite ,
the resulting limit is @xmath40 .
this limit implies , from eq .
( [ g ] ) , that @xmath41 . in summary ,
structure formation constraints require that only a small fraction ( less than @xmath16 ) of the dm in the universe might have been produced in a relativistic state from a non - thermal process such as the decay we consider in the present framework .
as mentioned above , the lifetime and the energy released by the mother particle @xmath31 are also constrained by bbn bounds .
this is because the energy released at a given time in the history of the universe might induce electromagnetic showers that create and/or destroy light elements synthesized in the early universe @xcite . in this subsection
we analyze the possible effects of this scenario on bbn .
given the overall quantitative success of bbn in the `` standard cosmological model '' , the first requirement for new physics is that it not drastically alter any of the light elemental abundances such as those of d , @xmath42he , @xmath43li .
energy releases which occur long after bbn may , in principle , spoil the successful bbn predictions for light elemental abundances .
the energy of photons produced in late decays of @xmath44 would have been rapidly redistributed through scattering off background photons ( @xmath45 ) as well as through inverse compton scattering ( @xmath46 ) @xcite . as a result ,
the constraints we obtain from bbn are , to an excellent approximation , independent of the initial energy distribution of the injected photons and are only sensitive to the _ total energy
_ released in the decay process @xcite . in order to derive bbn limits on the fraction of relativistic , non - thermally produced dm via the decay @xmath29 , we therefore need to calculate the total electromagnetic energy released .
let @xmath47 , where @xmath48 is the number density of particles of a particular species and @xmath49 is the number density of cmb photons , given by @xmath50 the total electromagnetic energy released from the @xmath31 decay is thus @xmath51 .
if for each @xmath44 particle we have the production of a dm particle plus a photon , then @xmath52 , where @xmath53 determines the number density of particles at a time equal to the lifetime of @xmath44 , and @xmath54 is the number density of dm particles today .
we thus find that the normalized number density of dm particles is given by @xmath55 this expression evaluates to @xmath56 the factor @xmath28 showed up in eq .
( [ yx ] ) because we assume here that only a fraction of the dm of the universe is produced in the decay process , while the majority of it is produced in a non - relativistically by some other mechanism that does not contribute any significant energy release at bbn ( for example , via a standard thermal freeze - out process )
. since the photon energy produced in the decay is , @xmath57 the total electromagnetic energy released is given by @xmath58 in the limit @xmath59 we can directly relate the total energy release given in eq .
( [ zeta ] ) with the quantity @xmath60 as well as with @xmath61 , as given in eq .
( [ deltaneff ] ) . lifetime ( @xmath62 ) and
the fraction of relativistically produced dm ( @xmath28 ) times the mother - to - daughter mass ratio @xmath63 , and constraints from bbn and cmb .
the shaded regions show bbn bounds on the non - thermal production of dm via the decay @xmath29 , corresponding to an excess relativistic degrees of freedom @xmath64 . the green curve represents the cmb bound ( regions to the right of the curve are ruled out ) .
we assume @xmath59 . ] with this in mind , we can use the limits on the light elemental abundances derived in ref . @xcite . since their results were obtained as a function of the total energy released , we straightforwardly convert them in terms of the quantity @xmath65 .
we show our results in fig .
[ graph7 ] , where the shaded regions show portions of parameter space ruled out by bbn , as calculated in ref .
@xcite employing the `` baryometer '' parameter @xmath66 .
the straight lines indicate combinations of the quantity @xmath67 and @xmath62 producing the @xmath61 as in the labels .
[ graph7 ] shows that the bbn bounds are weak for early decays because at early times the universe is hot and the initial photon spectrum is rapidly thermalized , leaving just a few high - energy photons able to modify the light elemental abundances .
however , for lifetimes longer than @xmath68 s , bbn excludes most of the relevant parameter space .
similar to the bbn constraints , bounds from measurements of the cmb spectrum are largely independent of the injected photon energy spectrum .
these bounds also depend primarily only on the total energy release .
the key effect of the additional energy injection in the form of photons is related to spectral distortions to the cmb black - body spectrum @xcite . for times
@xmath69 s the processes of bremsstrahlung , i.e. @xmath70 ( where x is an ion ) , compton scattering and double compton scattering @xmath71 quickly thermalize the injected photon energy @xcite . for @xmath72
s , however , the bremsstrahlung and double compton processes become inefficient , and the photon spectrum relaxes to a bose - einstein distribution with a chemical potential ( @xmath73 ) different from zero .
limits on @xmath73 can then be used to constrain this additional injected energy component .
the change in the cmb spectrum caused by the energy injection is constrained by precise measurements from current satellites and telescopes . following the procedure used in ref .
@xcite , we calculated the limit ( shown as a solid green curve in fig .
[ graph7 ] ) imposed by cmb data on the non - thermal production of dm at a given late time . in particular , measurements of the cmb spectrum constrain any perturbation which might cause it to depart from an almost perfect blackbody spectrum .
these limits are expressed in terms of deviations of the chemical potential @xmath73 from 0 .
current bounds imply that @xmath74 @xcite .
this upper limit can then be translated into a bound on the amount of energy released during the decay , providing a constrain on the quantity @xmath75 for a given lifetime .
[ graph7 ] clearly shows that the cmb bounds become more restrictive than those from bbn for lifetimes @xmath62 longer than @xmath76 s , ruling out almost all the relevant parameter space to explain @xmath61 with relativistic non - thermally produced wimps .
to summarize , we found thus far that constraints from structure formation , bbn and cmb require ( i ) @xmath77 and ( ii ) a lifetime for the mother particle @xmath31 shorter than @xmath68 s. for concreteness , we present here four example dark matter models , including a supersymmetric case , where dm particles could be partly produced non - thermally in association with a photon .
we also show that these models have the potential to reproduce @xmath17 while obeying the cosmological bounds described in the previous section .
unless otherwise noted , we shall use the symbol @xmath78 for the mass of the mother particle and @xmath79 for the mass of the dm particle , while we use different symbols to indicate the associated fields . before discussing specific models , we note that , in general , we can recast eq .
( [ deltaneff ] ) as @xmath80 this relation , in view of the constraints obtained above , namely that @xmath81 s and that @xmath82 , implies @xmath83 we thus find that the ratio between the mother and the dm particle , for appreciable values of @xmath61 , must be at least of the order of @xmath84 .
we then conclude that to produce a large enough number of effective relativistic degrees of freedom , the mother particle and the dm particle _ can not be nearly degenerate_. in particular , this rules out scenarios where the two relevant particles are expected to be close in mass , such as , for example , in any universal extra dimensional setup ( see e.g. ref .
@xcite ) . in this first case ,
the mother particle @xmath31 is a heavy scalar ( @xmath85 , with mass @xmath78 ) which decays into a spin-1 , stable dm particle ( @xmath86 , with mass @xmath79 ) plus a photon .
the relevant lagrangian interaction term is given by the effective dimension-5 operator @xmath87 where @xmath88 is the @xmath89 electromagnetic field tensor .
the resulting lifetime of the mother particle in this case is found to be ) ; we thus neglect @xmath32 factors in the expression for the lifetime . ]
@xmath90 with @xmath91 the cutoff scale of the effective theory , as in eq .
( [ eq : leffspin0 ] ) . for a model in which a heavy scalar particle decays into a spin-1 dm particle plus a photon , for different @xmath61 .
the relevant effective operator for this model is given in eq .
( [ eq : spin-0 ] ) .
we assume in this figure @xmath92 and a lifetime @xmath93s . ]
the upper limit on the lifetime we derived above , @xmath94 , can then be recast as a lower limit on the mother particle mass , for a given cutoff scale @xmath91 .
this lower limit on the mother particle mass can then also be , in turn , converted to a lower limit on the dm particle mass @xmath79 , once values for the parameters @xmath28 and @xmath61 are fixed . using the expression for the lifetime , along with eq .
( [ deltaneff ] ) , we obtain the requirement @xmath95 in fig .
[ fig : spin-0 ] we plot the lower limit for the dm mass as a function of the cutoff scale @xmath91 , for @xmath92 and @xmath93s . to obtain dm masses in the @xmath96 gev mass range in this effective model would require @xmath97 gev .
had we included a coupling constant @xmath98 in eq .
( [ eq : spin-0 ] ) , we would have concluded that for @xmath99 gev , if @xmath91 is of the order of the gut scale ( @xmath100 gev ) the required coupling would have been @xmath101 . in this case , a heavy spin-1 boson ( @xmath102 , again with mass @xmath78 ) decays into a spin-1 dm particle ( @xmath86 , of mass @xmath79 ) plus a photon .
a lagrangian term which mediates this decay through a renormalizable operator is @xmath103 where , again , @xmath88 is the standard model electromagnetism field tensor .
the lifetime of the mother particle in this case is @xmath104 we can again use the upper limit on the lifetime @xmath105 and eq .
( [ deltaneff ] ) to set a lower limit on the dm mass , @xmath106 we can thus calculate the values for @xmath98 and for the dm particle mass producing a given number of effective relativistic degrees of freedom @xmath61 , once the fraction of dm produced non - thermally ( @xmath107 ) and the lifetime ( @xmath108 s ) are set to values compatible with bbn and structure formation bounds .
we show our results in fig .
[ fig : spin-1 ] . .
the relevant operator for this effective model is given in eq .
( [ eq : spin-1 ] ) .
we employed @xmath92 and a lifetime @xmath93s .
note that the x - axis is multiplied by a factor @xmath109 . ]
the figure implies that in this particular scenario a dm particle with a mass in the @xmath110 gev range might account for a @xmath111 extra radiation component for @xmath112 .
in this case a heavy fermion ( @xmath113 , with mass @xmath78 ) decays into a spin-1/2 dm particle ( @xmath114 , mass @xmath79 ) plus a photon according to the effective dimension-5 operator @xmath115 the lifetime of the mother particle in this scenario is @xmath116 we can again use the upper limit on the lifetime to set a lower limit on the dm particle mass , @xmath117 this case closely mirrors the spin-0 example of sec .
[ sec : spin0 ] , the only difference being a numerical factor of 32 in the lifetime .
this translates to moving the dm mass upper limit down by a factor of @xmath118 .
therefore we conclude that , similarly to the spin-0 case , the non - thermal production of a @xmath110 gev dm particle can mimic one extra relativistic neutrino in the early universe for @xmath97 gev , as shown in fig .
[ fig : spin-12 ] . .
the relevant operator for this effective model is given in eq .
( [ lfermion ] ) .
we employed @xmath92 and a lifetime @xmath93s . ]
neutralinos are the mass eigenstates resulting from a mixture of neutral b - ino , w - ino , and higgs - inos . in many supersymmetric models ,
including certain supergravity scenarios with universal soft supersymmetry breaking masses at the grand unification scale , the lightest neutralino is an almost pure bino .
the lightest neutralino is also the next - to - lightest supersymemtric particle with the lightest supersymmetric particle being the gravitino .
binos decay into a gravitino - photon final state via the interaction lagrangian term @xcite @xmath119\gamma^{\mu}\tilde{b^{\mu}}f_{\nu \rho } , \label{eqgravitino}\]]where @xmath120 gev is the reduced planck mass . in models with low - scale supersymmetry breaking @xmath121 , which is exactly the limit needed to realize the scenario we are interested in . in this case , from eq .
( [ eqgravitino ] ) we find a neutralino lifetime of @xmath122 similarly to previous sections , we can convert the bound on the lifetime into a lower limit on the dm mass , of gravitinos are non - thermally produced by the decay of pure binos .
the relevant lifetime is derived from eq .
( [ eqgravitino ] ) . as in the previous figures ,
we have used @xmath92 and . ]
@xmath123 setting , as above , @xmath92 and @xmath108 s , we show in fig .
[ figravitino ] that a gravitino with mass @xmath124 in the 2 to 20 kev range mimics the effect of an extra neutrino while still obeying cosmological bounds .
we note that such light gravitinos are marginally consistent with bounds from small scale structure ( for example lyman - alpha forest data , see e.g. ref .
@xcite ) . also , assuming a gluino mass in the tev range , gravitinos in the mass range needed here would require a reheating temperature @xmath125 close to the electro - weak scale to avoid over - closing the universe , since the thermal production of gravitinos in the early universe @xcite implies @xmath126 such a low reheating temperature would rule out certain scenarios for the production of the baryon asymmetry in the universe , such as leptogenesis , but is in general not phenomenologically implausible .
finally , we note that the constraints on the lifetime and eq .
( [ eq : binolifetime ] ) imply a lower limit on the bino mass , which must be larger than about 1 gev .
this is , of course , perfectly compatible with the supersymmetric models relevant here .
recent data show tentative evidence for excess effective radiation degrees of freedom in the early universe . in this study
, we analyzed scenarios where such extra radiation stems from a sub - dominant population of wimp dm particles which is produced relativistically from the decay of a heavier particle .
such decay could affect structure formation , the cosmic microwave background , and the production of light elements in the early universe .
this framework is therefore constrained in a model - independent way by various bounds from cosmological observations .
we showed that structure formation limits the _
fraction _ of the dm energy density that can be produced in the aforementioned relativistic state to less than a percent , while cmb and bbn force the decay when the dm is produced to occur with a lifetime shorter than about @xmath68 s. given these two key constraints , we showed that a large enough contribution to the effective relativistic degrees of freedom @xmath61 is achieved only for a relatively large mass hierarchy ( on the order of @xmath84 ) between the mother ( @xmath31 ) and daughter ( @xmath127 ) particle masses in the decay ( for example , @xmath128 ) .
we presented a set of illustrative particle dm models that implement this scenario , including effective theories as well as a supersymmetric example .
for each example , we derived the relevant combination of parameters leading to a significant @xmath61 , while being consistent with all cosmological bounds from structure formation , bbn and cmb . as a final remark
, we note that in the context of the present framework it is in principle easy to accommodate both : * @xmath129 gev , i.e. in the mass region of particular interest today from both direct detection @xcite and indirect detection @xcite , and which has been extensively investigated in many particle physics models @xcite , or * assume that the dm particle mass is in the classic wimp @xmath130 gev range , which is also particularly interesting from the standpoint of possible indirect detection signals @xcite , and which has also been exhaustively explored in particle physics models as well , see e.g. ref .
@xcite . for any dm particle mass , the key requirements that the present setup imposes are to have : 1 . a large enough mother particle mass @xmath131 compared to the dark matter mass , to produce a sizable @xmath61 according to eq . ( [ deltaneff ] ) , 2 . a short enough lifetime for decay into the dm particle ( @xmath132 s ) to evade bbn and cmb constraints , and 3 . a small enough fraction of relativistically produced particles ( @xmath82 ) to evade structure formation bounds .
this work is partly supported by the department of energy under contract de - fg02 - 04er41286 ( sp ) , and by the brazilian national counsel for technological and scientific development ( cnpq ) ( fq ) .
the boost factor of a dm particle produced in the decay @xmath29 can be found using energy and momentum conservation , @xmath133 assuming that @xmath44 decays at rest , we find @xmath134 substituting eq .
( [ daneq3 ] ) into eq .
( [ daneq1 ] ) and taking @xmath135 we get @xmath136 at the time of decay , the universe was dominated by radiation with @xmath139 , thus @xmath140 .
\label{gammax}\ ] ] therefore , at matter - radiation equality ( @xmath141 ) we obtain @xmath142 .
\label{gammax2}\ ] ] eq .
( [ gammax2 ] ) gives us the boost factor of the dm particle as a function of time and of the mother - to - daughter mass ratio , for the case where all decays happen when the universe was radiation dominated .
since the additional radiation density reads @xmath143where @xmath28 is the fraction of dm non - thermally produced in the decays , we conclude that the number of effective relativistic degrees of freedom at matter - radiation equality associated with the non - thermal production of dm is @xmath144 , where @xmath145 is the number density of one neutrino species at the same epoch . using eq .
( [ rhodmnu ] ) , we find @xmath146 thus , substituting eq .
( [ gammax ] ) into eq .
( [ neffeq ] ) , we obtain @xmath147\times f. \label{deltaneff2}\end{aligned}\ ] ] this expression indicates that if some fraction of the dm of the universe is produced non - thermally via the decay @xmath29 , for reasonable values of the lifetime and mass ratio ( @xmath148 ) , this production mechanism produces an effective excess of relativistic degrees of freedom .
r. bernabei , p. belli , a. di marco , f. cappella , a. dangelo , a. incicchitti , v. caracciolo and r. cerulli _ et al .
_ , arxiv:1301.6243 [ astro-ph.ga ] . c. e. aalseth
_ et al . _
[ cogent collaboration ] , phys .
lett . * 106 * , 131301 ( 2011 ) .
[ arxiv:1002.4703 [ astro-ph.co ] ] .
j. kopp , arxiv:1304.1184 [ hep - ph ] .
a. de simone , a. riotto and w. xue , arxiv:1304.1336 [ hep - ph ] .
m. ibe , s. iwamoto , s. matsumoto , t. moroi and n. yokozaki , arxiv:1304.1483 [ hep - ph ] . h. -b .
jin , y. -l .
wu and y. -f .
zhou , arxiv:1304.1997 [ hep - ph ] .
t. linden and s. profumo , arxiv:1304.1791 [ astro-ph.he ] .
j. l. feng , a. rajaraman and f. takayama , phys .
lett . * 91 * , 011302 ( 2003 ) [ hep - ph/0302215 ] .
jonathan l. feng , arvind rajaraman , and fumihiro takayama , phys . rev .
d 68 , 063504 ( 2003 ) , [ hep - ph/0307375 ] . | the planck and wmap9 satellites , as well as the atacama and south pole telescopes , have recently presented results on the angular power spectrum of the comic microwave background .
data tentatively point to the existence of an extra radiation component in the early universe . here
, we show that this extra component can be mimicked by ordinary wimp dark matter particles whose majority is cold , but with a small fraction being non - thermally produced in a relativistic state .
we present a few example theories where this scenario is explicitly realized , and explore the relevant parameter space consistent with bbn , cmb and structure formation bounds . |
spontaneous parametric down - conversion ( spdc ) has proven to be an excellent technology for quantum communication , with spdc photons functioning as `` flying qubits . ''
the discovery by knill _
et al . _
@xcite that linear optics and single photon detectors are sufficient for scalable quantum computation has opened the possibility that spdc may also be useful for quantum computation .
as the proposals for quantum information processing with spdc become more sophisticated , the technical demands placed on spdc sources become more stringent . for example , a quantum key distribution experiment based on polarization entanglement requires that two two - photon amplitudes ( @xmath0 and @xmath1 , for example ) be made indistinguishable .
the spectral properties of the two photons are only important if they are correlated to the polarization degree of freedom . a more stringent form of indistinguishability is typically required for quantum computation with linear optics : it must be impossible to determine which source produced a certain photon after that photon emerges from a beamsplitter .
this in turn requires that all of the photons properties be controlled such that it is impossible to learn any information about the identity of a given photon s source .
photon pairs produced by spdc are often correlated in one or more of their properties ( frequency , direction , etc . ) .
these correlations can destroy the requisite indistinguishability by enabling one to learn about one photon by performing measurements on its twin @xcite . in this paper
, we describe an spdc source that produces photon pairs that have arbitrary correlation in frequency .
the source we propose enables an unusual flexibility in the control of the marginal spectra of the spdc photons .
specifically , our source can produce frequency - uncorrelated photon pair in which the center frequency and the bandwidth of each photon is controlled independently , regardless of the nonlinear material s dispersion curves .
this makes the source well - suited for applications that span quantum communication and computation , such as teleportation @xcite and entanglement swapping @xcite . in these applications ,
one photon of a pair ( the `` communication '' photon ) is often sent to another party through a long span of optical fiber , while the other photon ( the `` computation '' photon ) is analyzed and detected in a localized interferometer .
it is desirable that the communication photon be narrow - band ( such that effects like polarization mode dispersion are mitigated ) and have a wavelength in the infrared ( where optical fiber is least lossy ) .
contrariwise , the computation photon should be broad - band ( such that it can be used in interferometers with small path - length differences ) and have a wavelength suited for high - efficiency single - photon counters .
the paper is organized as follows . in section [ gapmspdc ]
, we introduce the new technique and demonstrate that it permits the generation of photon pairs with arbitrary joint spectrum . in section [ example ] , we show how this spectral control can be combined with polarization entanglement by considering a specific example involving a bbo waveguide . in section [ int ]
, we discuss the possibility of using this source as part of a distributed quantum information processor based on integrated optics .
finally , we summarize our results in section [ conclusions ] .
our source is a generalization of the design we previously introduced under the name _ auto - phase - matched _
spdc @xcite ; thus , we name the new scheme _ generalized _ auto - phase - matched spdc . like the original scheme , the new scheme features counter - propagating spdc created in a single - mode nonlinear waveguide by a transverse pump pulse . in the original scheme ( fig .
[ setup]a ) , the pump pulse is cross - spectrally pure ( _ i.e. _ the complex envelope factors into separate functions of space and time ) , and impinges on the waveguide at normal incidence . in the new scheme ( fig . [ setup]b ) , the pump pulse may have cross - spectral correlations , and may approach the waveguide at non - normal incidence . with these constraints on the pump pulse
relaxed , the center frequency and bandwidth of each spdc photon may be controlled independently . in typical spdc experiments ,
a monochromatic pump beam is used . in this situation ,
the sum of the frequencies of the signal and idler photons is fixed , thus the photons frequencies are anticorrelated .
spdc with generalized joint spectral properties ( _ i.e. _ correlated , uncorrelated , anticorrelated ) was studied theoretically by campos _
et al . _
@xcite . here
we review the experimental proposals for generating frequency - uncorrelated spdc .
a number of techniques have been proposed for creating frequency - uncorrelated spdc ; however , they all impose certain constraints on the center frequencies and/or bandwidths of the spdc photons .
proposed a method for creating frequency - uncorrelated spdc based on a group - velocity matching condition introduced by keller and rubin @xcite .
their method can be used to create degenerate , frequency - uncorrelated photons ; however , the center frequency of down - conversion is fixed by the nonlinear material , and the bandwidth of two spdc photons must be equal .
they also demonstrate that degenerate , frequency - uncorrelated photons with different bandwidths may be generated ; however , in this case the bandwidths are fixed and can not be independently controlled .
proposed extending the approach of grice _ el al .
_ by using a periodically poled nonlinear crystal @xcite .
this allows one to to satisfy the zeroth - order term in the phase - matching relation at an arbitrary pump wavelength , making the group - velocity matching relation easier to satisfy .
even with such an enhancement , this approach does not have sufficient flexibility to allow independent control of the marginal spectra .
a distinct approach for creating frequency - uncorrelated spdc was independently discovered by uren _
et al . _
@xcite and our group @xcite . instead of relying on the satisfaction of a group - velocity matching condition , these approaches rely on the geometrical symmetry of degenerate , non - collinear type - i spdc ( the previously mentioned techniques only worked with type - ii spdc ) .
the essential difference between the two proposals is that for uren _ et al .
_ , the phase - matching relation in the pump propagation direction is a constraint that must be satisfied , while for our auto - phase - matched technique , the single - mode waveguide ensures that this relation is satisfied , regardless of the system parameters .
the relative lack of constraints for these techniques makes them attractive , since it suggests that they remain viable options even if the center frequency of spdc is constrained by some other factor ( optical fiber loss , detector efficiency , etc . ) .
nonetheless , both techniques suffer a lack of flexibility that is reminiscent of the previously mentioned schemes .
the spdc photons must be degenerate , and the bandwidths of the two photons must be equal . in the next section ,
we show that by generalizing our auto - phase - matched technique , we can obtain independent spectral control of the spdc photons .
to begin , we review the relationship between the pump pulse and the spdc in the geometry of fig .
[ setup ] . following the derivation in ref .
@xcite , a classical pump pulse described on the free - space side of the waveguide - air interface by @xmath2 stimulates the creation of a pair of photons described by the two - photon wavefunction @xmath3 where @xmath4\ ] ] @xmath5 is the refractive index for the pump polarization . here , and for the rest of the paper , we use the variable @xmath6 to refer to the component of the pump wavevector along the @xmath7-axis .
the ket @xmath8 represents a signal photon in the frequency mode @xmath9 and an idler photon in the frequency mode @xmath10 with corresponding propagation constants @xmath11 and @xmath12 , respectively . equation ( [ wavefunction ] ) conveys the main result of this paper : assuming the dispersion properties of the medium are known , it is possible to generate a down - converted photon pair with arbitrary joint spectrum by appropriately engineering the spatial and temporal characteristics of the pump pulse . in fig .
[ setup]a , the pump pulse is parameterized by three numbers : the center frequency , the temporal coherence length @xmath13 , and the spatial coherence length @xmath14 .
these parameters may be chosen to produce degenerate spdc with controllable entanglement , as described in ref .
@xcite ; however , in order to obtain independent control of the center frequency and bandwidth of each spdc photon , one must relax the constraints on the pump pulse , as in fig .
[ setup]b . using eqs .
( [ pump ] ) and ( [ wavefunction ] ) , it is straightforward to show that a pump pulse described by @xmath15\ ] ] will yield the following frequency - uncorrelated two - photon state @xmath16\,|\omega_s\rangle_s\,|\omega_i\rangle_i,\ ] ] where @xmath17 and @xmath18 are the center frequencies of the signal and idler beams , respectively , and we have used the following definitions and approximations : @xmath19 these approximations are valid in typical situations ; however , if required , more terms may be used at the expense of a more complicated expression for the pump pulse .
equations ( [ genpump ] ) and ( [ genwavefunction ] ) summarize the central result of this work . taken together , these relations can be thought of as an algorithm for producing frequency - uncorrelated spdc with arbitrary marginal spectra .
the wavefunction in eq .
( [ genwavefunction ] ) describes a frequency - uncorrelated two - photon state in which the signal photon is centered on @xmath17 with a bandwidth @xmath20 , and the idler photon is centered on @xmath18 with a bandwidth @xmath21 .
note that these two photons are not themselves indistinguishable ( unless @xmath22 and @xmath23 ) .
as previously mentioned , the indistinguishability arises in a multi - photon experiment when one photon of the pair enters an interferometer with one or more photons that have identical spectra . in this case
, the lack of frequency correlations between the signal and idler prevents a loss of interferometric visibility by ensuring that spectral measurements on one photon of the pair wo nt reveal any spectral information about the other .
equations ( [ genpump]-[end ] ) demonstrate that the four numbers @xmath17 , @xmath18 , @xmath20 , and @xmath21 , along with the dispersion properties of the waveguide , are sufficient to determine the form of the pump pulse required to generate the desired wavefunction .
we can simplify the description of the pump pulse by rewriting eq .
( [ genpump ] ) as @xmath24,\ ] ] where @xmath25 the algorithm for creating the appropriate pump pulse to produce the state in eq .
( [ genwavefunction ] ) is as follows .
a pulse is created with center frequency @xmath26 , spectral bandwidth @xmath27 and spatial bandwidth @xmath28 .
next , a dispersive element such as a wedge of quartz or a diffraction grating is used to correlate @xmath6 and @xmath29 by effecting the substitution @xmath30 finally , the pulse it directed towards the nonlinear waveguide at incidence angle @xmath31 where @xmath32 is measured outside the waveguide ( see fig . [ setup]b ) , and @xmath33 is the speed of light in vacuum . in fig .
[ jointspectrum ] , we present a graphical depiction of the relationship between the pump pulse and the resulting two - photon state , for both auto - phase - matched spdc ( fig .
[ jointspectrum]a ) and generalized auto - phase - matched spdc ( fig .
[ jointspectrum]b ) . in both cases ,
a plot of @xmath34 is superposed over the joint spectrum of the signal and idler photons . by plotting @xmath34 at the correct location and on the correct inner axes
, one can immediately infer the joint spectrum of the down - converted photons , simply by interpreting the plot using the outer axes . in fig .
[ jointspectrum]a , the non - zero portion of @xmath34 is centered on the @xmath35 axis , and one of the inner axes is scaled by the factor @xmath36 , which is the first derivative of @xmath37 evaluated at @xmath38 . in fig .
[ jointspectrum]b , the non - zero portion of @xmath34 is located at a general position , and the inner axes are no longer orthogonal ( unless @xmath39 ) . using this figure ,
the two desirable features of the two - photon joint spectrum ( non - degeneracy and independently controlled bandwidths ) are easily interpreted in terms of the pump pulse .
that is , the non - degeneracy of the photon pair derives from the condition @xmath40 , which in turn derives from the non - normal incidence of the pump pulse .
similarly , the independent control of the two photons respective bandwidths derives from the cross - spectral correlation in the pump pulse ( @xmath34 does not factor into a function of @xmath6 times a function of @xmath29 ) .
[ table ] .parameters that describe the pump pulse required to produce non - degenerate , frequency - uncorrelated , polarization - entangled spdc in a single - mode bbo waveguide , using the technique depicted in fig .
[ setup]b .
the signal photon is at @xmath41 @xmath42 with coherence length @xmath43 mm , and the idler photon is at @xmath44 @xmath42 with coherence length @xmath43 cm .
the pump pulse is comprised of coherently superposed , independently controlled pulses @xmath45 and @xmath46 in the two polarization modes @xmath7 and @xmath47 , respectively .
the parameters @xmath27 , @xmath28 , @xmath48 , and @xmath32 are defined in sec .
[ gapmspdc ] .
the negative values of @xmath32 indicate that the projection of the pump wavevector along the waveguide is oriented in the negative @xmath7 direction ( see fig .
[ setup]b ) . [ cols="<,<,<",options="header " , ] in the generalized auto - phase - matched technique , one can obtain polarization entanglement by adjusting the polarization state of the pump pulse , without sacrificing the spectral control described above . to illustrate this feature , we present an example involving non - degenerate
, polarization - entangled spdc produced in a single - mode bbo waveguide .
the general idea is to use two of the nonlinear medium s @xmath49 tensor elements at the same time , by preparing a coherent superposition of two polarization modes of the pump pulse . when producing polarization - entangled photon pairs , it is typically desirable that a given photon have the same spectral properties for both two - photon polarization amplitudes
. therefore , in creating the pump pulse , we use the same four numbers @xmath17 , @xmath18 , @xmath20 , and @xmath21 in calculating the desired pulse shape for both pump polarization modes .
however , since the two two - photon amplitudes relate to spdc processes taking place in distinct polarization modes , the dispersion properties of the waveguide will in general be different .
thus , using the notation of fig .
[ setup ] , the pump pulse will be characterized by two functions : @xmath46 , which describes the @xmath47-polarized component of the pump - pulse , and @xmath45 , which describes the @xmath7-polarized component of the pump - pulse in the case of bbo , the relevant tensor elements are @xmath50 pm / v and @xmath51 pm / v and @xmath52 are not equal , we adjust the ratio of the optical powers in each pump polarization mode in order to obtain the maximally entangled state @xmath53 , where the relative phase @xmath54 is determined by the relative phase between the two pump polarization modes .
it is straightforward to produce non - maximally entangled and/or mixed polarization states with this technique .
non - maximally entangled states may be produced by adjusting the ratio of the optical powers in each of the pump s polarization modes .
mixed states may be produced by allowing the pump pulse to become partially depolarized .
although in this paper we are concerned with frequency - uncorrelated ( and thus , frequency - unentangled ) photon pairs , analogous generalizations for partial entanglement and mixedness in the frequency degrees of freedom are possible , given the appropriate manipulations of the pump pulse . ] .
therefore , using the notation of fig .
[ setup ] , the pump beam will approach the waveguide in the @xmath55-@xmath7 plane , and will be composed of a @xmath47-polarized pulse and a @xmath7-polarized pulse . in table
i , we list the calculated values of @xmath27 , @xmath28 , @xmath48 , and @xmath32 ( defined in sec .
[ gapmspdc ] ) that correspond to a frequency - uncorrelated polarization - entangled pair of photons with the signal photon at @xmath41 @xmath42 with coherence length @xmath43 mm , and the idler photon at @xmath44 @xmath42 with coherence length @xmath43 cm .
these values for center wavelength and coherence length were chosen in order to make the signal photon suitable for long - distance optical fiber transmission , and the idler photon suitable for local processing in an integrated optical circuit .
in calculating the values in table i , we have ignored waveguide dispersion , using instead the selmeier curves to describe the material dispersion in the bbo single - mode waveguide .
generalized auto - phase - matched spdc is particularly well - suited for quantum information processing on an integrated optical circuit ( see fig . [ integrated ] ) . among the advantages of replacing an array of discrete optical elements with an integrated optical circuit
are the following : reduced size , reduced loss due to fewer connectors , and `` common mode '' noise processes because of the close proximity of optical elements .
however , there are substantial experimental challenges associated with constructing an integrated optical quantum information processor .
perhaps the most obvious challenge is finding a material that can perform as many of the required functions ( photon source , modulation , detection ) as possible . a significant advantage of generalized auto - phase - matched spdc in that the choice of material places essentially no limitation on the spectral and polarization properties of the photon pairs that will be produced .
all that is required is that the material s @xmath49 tensor have the appropriate non - zero elements such that , for a given orientation of the optic axis with respect to the waveguide , the desired spdc process will occur .
figure [ integrated ] depicts a conceptual schematic of an integrated optical quantum information processor which employs generalized auto - phase - matched spdc for generating photons .
the figure highlights several of the practical advantages associated with this technology .
first , the sources may be placed at the edge of the circuit and combined with single - photon counters to implement conditional single - photon sources .
second , due to the transverse - pump configuration , the photon pairs may be created within the interior of the optical circuit . finally , since there is no group - velocity matching relation associated with generalized auto - phase - matched spdc , poling of the nonlinear waveguide at each source is not required .
we have described a scheme for generating polarization - entangled photons pairs with arbitrary spectrum . by controlling the spatial , temporal , and polarization properties of the pump pulse , it is possible to generate the desired two - photon state , regardless of the dispersion properties of the nonlinear medium .
we provided a calculation of the parameters describing the pump pulse required to generate a photon - pair with a particular joint spectrum in a single - mode bbo waveguide .
finally , we discussed the role this source technology might play in a distributed quantum information processor based on integrated optics .
this work was supported by the national science foundation ; the center for subsurface sensing and imaging systems ( censsis ) , an nsf engineering research center ; the defense advanced research projects agency ( darpa ) ; and the david and lucile packard foundation . | we present a scheme for generating polarization - entangled photons pairs with arbitrary joint spectrum .
specifically , we describe a technique for spontaneous parametric down - conversion in which both the center frequencies and the bandwidths of the down - converted photons may be controlled by appropriate manipulation of the pump pulse .
the spectral control offered by this technique permits one to choose the operating wavelengths for each photon of a pair based on optimizations of other system parameters ( loss in optical fiber , photon counter performance , etc . ) .
the combination of spectral control , polarization control , and lack of group - velocity matching conditions makes this technique particularly well - suited for a distributed quantum information processing architecture in which integrated optical circuits are connected by spans of optical fiber . |
we denote the complex plane by @xmath8 , and the unit disk @xmath9 by @xmath10 . in 1989 , schwick @xcite proved a normality criterion which states that : _ for positive integers @xmath11 , @xmath2 , let @xmath3 be a family of functions meromorphic in @xmath4 .
if each @xmath5 satisfies @xmath6 for @xmath7 , then @xmath3 is a normal family .
_ this result holds good for holomorphic functions with the case @xmath12 .
the following theorem is a result of wang and fang @xcite .
the proof was omitted in that article , here we give a proof of this result and extend this theorem .
[ thm ] let @xmath1 be positive integers and @xmath12 and @xmath0 be a domain in @xmath8 .
let @xmath3 be a family of functions meromorphic on @xmath4 .
if each @xmath5 satisfies @xmath6 for @xmath7 , then @xmath3 is a normal family .
+ it is natural to ask what can happen if we have a solution of @xmath13 . for this question we can extend theorem [ thm ] for the case @xmath14 in the following manner .
[ thm1 ] let @xmath1 be positive integers and @xmath15 and @xmath0 be a domain in @xmath8 .
let @xmath3 be a family of functions meromorphic on @xmath4 .
if for each function @xmath16 has at most one zero ignoring multiplicity @xmath17 in @xmath4 , then @xmath3 is a normal family .
+ in this paper , we use the following standard notations of value distribution theory , @xmath18 .
we denote @xmath19 any function satisfying @xmath20 , as @xmath21 possibly outside of a set with finite measure .
in order to prove our results we need the following lemmas .
+ \ { @xcite , p. 216 ; @xcite , p. 814}[lem1](zalcman s lemma ) + let @xmath3 be a family of meromorphic functions in the unit disk @xmath10 , with the property that for every function @xmath22 the zeros of @xmath23 are of multiplicity at least @xmath24 and the poles of @xmath23 are of multiplicity at least @xmath11 .
if @xmath3 is not normal at @xmath25 in @xmath10 , then for @xmath26 , there exist 1 . a sequence of complex numbers @xmath27 , @xmath28 , 2 . a sequence of functions @xmath29 , 3 .
a sequence of positive numbers @xmath30 , such that @xmath31 converges to a non - constant meromorphic function @xmath32 on @xmath8 with @xmath33 . moreover @xmath32 is of order at most two .
+ [ lem2]\ { @xcite , lemma 2.5 , lemma 2.6 ; @xcite , lemma 2.2 } let @xmath34 be a rational function and @xmath35 be non - constant .
then @xmath36 where @xmath11 is a positive integer , @xmath37 @xmath38 and @xmath39 denotes the degree of p. + [ lem3 ] @xcite let @xmath40 where @xmath41 are constants , @xmath42 is a positive integer and @xmath43 are polynomials with deg@xmath44 deg@xmath45 . if @xmath46 , then @xmath47 + [ lem4 ] let @xmath48 is two positive integer and @xmath49 let @xmath23 be a non - constant rational function then @xmath50 has a root for all nonzero complex numbers @xmath51 .
+ suppose @xmath52 has no root .
first we suppose @xmath23 is a non - constant polynomial of degree @xmath53 , then @xmath50 is a polynomial of degree @xmath54 .
thus by fundamental theorem of algebra @xmath50 has a solution .
+ again , let @xmath23 is a non - polynomial rational function .
we set @xmath55 where @xmath56 is a nonzero constant and @xmath57 are positive integers .
we denote @xmath58 @xmath59 where @xmath60 is a polynomial and deg@xmath61 suppose @xmath62 , then @xmath63 from and @xmath64 deg@xmath65 deg@xmath66 deg@xmath67 .
this gives @xmath68 @xmath69 .
this implies @xmath70 .
so @xmath71 hence @xmath72 .
therefore we can express @xmath73 as follows @xmath74 where @xmath41 are constants , @xmath75 is an integer , @xmath76 and @xmath77 are polynomials with deg@xmath44 deg@xmath45 .
since @xmath78 , then by [ lem3 ] we get @xmath79 which contradicts the fact that deg@xmath80 deg@xmath81 .
thus @xmath82 has a solution in @xmath83 + [ lemmali ] @xcite let @xmath1 be positive integers such that @xmath15 and @xmath84 be a finite complex number , and @xmath23 be a non - constant rational meromorphic function , then @xmath85 has at least two distinct zeros .
[ lemm1]\ { @xcite p.38 } let @xmath86 be a transcendental meromorphic function on @xmath8 , then @xmath87 for every positive integer @xmath11 .
+ [ lemm2 ] let @xmath23 be a non - constant meromorphic function on @xmath8 .
then @xmath88 @xmath89 [ lem5 ] @xcite let @xmath86 be a transcendental meromorphic function .
then for each positive number @xmath90 and each positive integer @xmath11 , we have @xmath91 [ lem6]\ { @xcite corollary 3 .
} if a meromorphic function of finite order @xmath92 has only finitely many critical values , then it has at most @xmath93 asymptotic values .
+ [ lem7 ] @xcite let @xmath60 be a transcendental meromorphic function and suppose that @xmath94 and the set of finite critical and asymptotic values of @xmath60 is bounded
. then there exists @xmath95 such that @xmath96 which are not poles of @xmath60 .
+ [ lem8 ] @xcite if @xmath23 is a trancendental meromorphic function and @xmath11 be a positive integer , then , for every positive number @xmath90 , @xmath97 the following lemma was proved by bergweiler @xcite and wang @xcite independently . here
we are giving another proof of this lemma .
[ lemma ] let @xmath86 be a transcendental meromorphic function with finite order .
let @xmath98 be two positive integers such that @xmath12 , then @xmath99 has infinitely many zeros for all @xmath100 .
suppose on the contrary that @xmath101 assumes the value @xmath51 only finitely many times
. then @xmath102 by nevanlinna s first fundamental theorem and lemma [ lemm1 ] and lemma [ lemm2 ] @xmath103 together with nevanlinna s first fundamental theorem this yields @xmath104 first , we consider the case when @xmath105 , then by lemma [ lem8 ] , for every @xmath106 , we have @xmath107 from and , and using the fact that zeros of @xmath108 has multiplicity at least 3 in this case , we get @xmath109 now , taking @xmath110 , from and , we obtain @xmath111 which contradicts the fact that @xmath23 is a transcendental meromorphic function .
thus , lemma [ lem8 ] is proved for the case @xmath112 .
+ now , for the case @xmath113 , we use the method of fang @xcite .
we first consider that @xmath86 has only finitely many zeros so is @xmath114 has only finitely many zeros @xmath115 . and
invoke the lemma [ lem5 ] and combine it with , we have @xmath116 thus , we obtain @xmath117 which is a contradiction , therefore the theorem is valid in this case .
now , consider the case when @xmath86 has infinitely many zeros @xmath118 .
define @xmath119 then @xmath120 if we show that @xmath121 has infinitely many zeros then we have done .
suppose @xmath121 has only finitely many zeros , so @xmath60 has only finitely many critical values and hence @xmath60 has only finitely many asymptotic values . without any loss of generality we may assume that @xmath122 thus by lemma [ lem7 ] , we get @xmath123 since @xmath124 as @xmath125 , @xmath126 as @xmath127 but @xmath128 as @xmath129 a contradiction . hence we deduce that @xmath130 has infinitely many zeros .
this completes the proof of theorem .
+ [ lem9 ] @xcite let @xmath23 be an entire function . if the spherical derivative @xmath131 is bounded in @xmath8 , then the order of @xmath23 is at most 1 .
since normality is a local property , we assume that @xmath132 .
suppose @xmath133 is not normal in @xmath134 . without loss of generality
we assume that @xmath133 is not normal at the point @xmath25 in @xmath10
. then by lemma [ lem1 ] , there exist such that @xmath139 converges locally uniformly to a non - constant meromorphic function @xmath140 in @xmath8 with @xmath141
. moreover @xmath32 is of order at most two .
we see that @xmath142 converges locally uniformly with respect to the spherical metric . by hurwitz
s theorem , @xmath143 or @xmath144 + let @xmath143 , then @xmath32 has no pole this implies that @xmath32 is an entire function having no zero . since @xmath145 , we may put @xmath146 , where @xmath147 and @xmath148 are constants
. therefore we get @xmath149 which is not possible .
+ thus @xmath150 , which contradicts lemma [ lem4 ] and lemma [ lemma ] .
thus @xmath133 is normal in @xmath4 .
this completes the proof of theorem . + since normality is a local property , we assume that @xmath132 .
suppose @xmath133 is not normal in @xmath134 .
without loss of generality we assume that @xmath133 is not normal at the point @xmath25 in @xmath10
. then by lemma [ lem1 ] , there exist such that @xmath139 converges locally uniformly to a non - constant meromorphic function @xmath140 in @xmath8 with @xmath141
. moreover @xmath32 is of order at most two .
we see that @xmath151 converges locally uniformly with respect to the spherical metric .
+ now we claim @xmath152 has at most one zero i m .
suppose @xmath152 has two distinct zeros @xmath153 and @xmath154 and choose @xmath155 small enough so that @xmath156 and @xmath152 has no other zeros in @xmath157 , where @xmath158 and @xmath159 . by hurwitz s theorem ,
there exist two sequences @xmath160 converging to @xmath161 respectively and from , for sufficiently large j , we have + @xmath162 since @xmath163 and @xmath164 , we have @xmath165 and @xmath166 for sufficiently large @xmath167 , so @xmath168 has two distinct zeros , which contradicts the fact that @xmath169 has at most one zero .
but lemma [ lemmali ] and lemma [ lemma ] confirms the non existence of such non - constant meromorphic function .
this contradiction shows that @xmath133 is normal in @xmath10 and this proves the theorem .
+ l. v. ahlfors , _ complex analysis , _ third edition , mcgraw - hill , 1979 .
w. bergweiler , a. eremenko , _ on the singularities of the inverse to a meromorphic function of finite order , _ rev . mat .
2 , ( 1995 ) , 355373 .
w. bergweiler , _ on the zeros of certain differential polynomials , _ arch . math . 64 ( 1995 ) , 199202 .
j. clunie , w. k. hayman , _ the spherical derivative of integral and meromorphic function , _ comment .
40 ( 1966 ) , 117148 .
m. l. fang , _ picard values and normality criterion , _ bull .
korean math .
2 , ( 2001 ) , 379387
. g. frank , g. weissenborn , _ rational deficient functions of meromorphic functions , _ bull . londan math .
( 1986 ) , 2933 .
w. k. hayman , meromorphic functions , _ clarendon press , oxford _ , 1964 .
y. li , y. gu , _ on normal families of meromorphic functions _ , j. math .
appl . , 354 ( 2009 ) , 421425 . j. schiff , normal families , _ springer - verlag , berlin , _ 1993
. w. schwick , _ normality criteria for families of meromophic functions , _ j. anal .
math 52 ( 1989 ) , 241289 .
y. f. wang , _ on mues conjecture and picard values , _ sci .
china 36 ( 1993 ) , 2835 .
y. f. wang , m. l. fang , _ picard values and normal families of meromorphic functions with multiple zeros , _ acta math .
sinica , new series 14 , no .
1 , ( 1998 ) , 1726 . l. yang , value distribution theory , _
springer - verlag , berlin , _ 1993 .
l. zalcman , _ a heuristic principle in complex function theory , _ the amer .
monthly , 82 ( 1975 ) , 813817 .
l. zalcman , _ normal families : new perspectives , _ bull .
soc . , 35 , no . 3 ( 1998 ) , 215230 . c. zeng , _ normality and
shared values with multiple zeros , _
( 2012 ) , 683686 .
s. zeng , i. lahiri , _ a normality criterion for meromorphic functions , _ kodai math .
j. 35 ( 2012 ) , 105114 .
s. zeng , i. lahiri , _ a normality criterion for meromorphic functions having multiple zeros , _ ann .
math.(to appear ) . | let @xmath0 be a domain , @xmath1 be positive integers and @xmath2 .
let @xmath3 be a family of functions meromorphic in @xmath4 .
if each @xmath5 satisfies @xmath6 for @xmath7 , then @xmath3 is a normal family .
this result was proved by schwick @xcite , in this paper we extend this theorem . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.